feat(disk-cache): rewrite diskcache backend to be k8s and NFS safe.

- no longer depend on diskcache
- use simpler version with Redis lock, and filesystem storage
- fixes  RCCE-78

rhodecode/lib/rc_cache/archive_cache/__init__.py

@@ -0,0 +1,29 b''
+# Copyright (C) 2015-2024 RhodeCode GmbH
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License, version 3
+# (only), as published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+# This program is dual-licensed. If you wish to learn more about the
+# RhodeCode Enterprise Edition, including its added features, Support services,
+# and proprietary license terms, please see https://rhodecode.com/licenses/
+
+from .fanout_cache import get_archival_cache_store
+from .fanout_cache import get_archival_config
+
+from .utils import archive_iterator
+from .utils import ArchiveCacheLock
+
+
+def includeme(config):
+    # init our cache at start
+    settings = config.get_settings()
+    get_archival_cache_store(settings)

rhodecode/lib/rc_cache/archive_cache/lock.py

@@ -0,0 +1,60 b''
+# Copyright (C) 2015-2024 RhodeCode GmbH
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License, version 3
+# (only), as published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+# This program is dual-licensed. If you wish to learn more about the
+# RhodeCode Enterprise Edition, including its added features, Support services,
+# and proprietary license terms, please see https://rhodecode.com/licenses/
+
+import redis
+from rhodecode.lib._vendor import redis_lock
+
+from .utils import ArchiveCacheLock
+
+
+class GenerationLock:
+    """
+    Locking mechanism that detects if a lock is acquired
+
+    with GenerationLock(lock_key):
+        compute_archive()
+    """
+    lock_timeout = 7200
+
+    def __init__(self, lock_key, url):
+        self.lock_key = lock_key
+        self._create_client(url)
+        self.lock = self.get_lock()
+
+    def _create_client(self, url):
+        connection_pool = redis.ConnectionPool.from_url(url)
+        self.writer_client = redis.StrictRedis(
+            connection_pool=connection_pool
+        )
+        self.reader_client = self.writer_client
+
+    def get_lock(self):
+        return redis_lock.Lock(
+            redis_client=self.writer_client,
+            name=self.lock_key,
+            expire=self.lock_timeout,
+            strict=True
+        )
+
+    def __enter__(self):
+        acquired = self.lock.acquire(blocking=False)
+        if not acquired:
+            raise ArchiveCacheLock('Failed to create a lock')
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        self.lock.release()

rhodecode/lib/rc_cache/archive_cache/utils.py

@@ -0,0 +1,30 b''
+# Copyright (C) 2015-2024 RhodeCode GmbH
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Affero General Public License, version 3
+# (only), as published by the Free Software Foundation.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU Affero General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+# This program is dual-licensed. If you wish to learn more about the
+# RhodeCode Enterprise Edition, including its added features, Support services,
+# and proprietary license terms, please see https://rhodecode.com/licenses/
+
+
+class ArchiveCacheLock(Exception):
+    pass
+
+
+def archive_iterator(_reader, block_size: int = 4096 * 512):
+    # 4096 * 64 = 64KB
+    while 1:
+        data = _reader.read(block_size)
+        if not data:
+            break
+        yield data

configs/development.ini

@@ -1,803 +1,813 b''
 
 ; #########################################
 ; RHODECODE COMMUNITY EDITION CONFIGURATION
 ; #########################################
 
 [DEFAULT]
 ; Debug flag sets all loggers to debug, and enables request tracking
 debug = true
 
 ; ########################################################################
 ; EMAIL CONFIGURATION
 ; These settings will be used by the RhodeCode mailing system
 ; ########################################################################
 
 ; prefix all emails subjects with given prefix, helps filtering out emails
 #email_prefix = [RhodeCode]
 
 ; email FROM address all mails will be sent
 #app_email_from = rhodecode-noreply@localhost
 
 #smtp_server = mail.server.com
 #smtp_username =
 #smtp_password =
 #smtp_port =
 #smtp_use_tls = false
 #smtp_use_ssl = true
 
 [server:main]
 ; COMMON HOST/IP CONFIG, This applies mostly to develop setup,
 ; Host port for gunicorn are controlled by gunicorn_conf.py
 host = 127.0.0.1
 port = 10020
 
 
 ; ###########################
 ; GUNICORN APPLICATION SERVER
 ; ###########################
 
 ; run with gunicorn --paste rhodecode.ini --config gunicorn_conf.py
 
 ; Module to use, this setting shouldn't be changed
 use = egg:gunicorn#main
 
 ; Prefix middleware for RhodeCode.
 ; recommended when using proxy setup.
 ; allows to set RhodeCode under a prefix in server.
 ; eg https://server.com/custom_prefix. Enable `filter-with =` option below as well.
 ; And set your prefix like: `prefix = /custom_prefix`
 ; be sure to also set beaker.session.cookie_path = /custom_prefix if you need
 ; to make your cookies only work on prefix url
 [filter:proxy-prefix]
 use = egg:PasteDeploy#prefix
 prefix = /
 
 [app:main]
 ; The %(here)s variable will be replaced with the absolute path of parent directory
 ; of this file
 ; Each option in the app:main can be override by an environmental variable
 ;
 ;To override an option:
 ;
 ;RC_<KeyName>
 ;Everything should be uppercase, . and - should be replaced by _.
 ;For example, if you have these configuration settings:
 ;rc_cache.repo_object.backend = foo
 ;can be overridden by
 ;export RC_CACHE_REPO_OBJECT_BACKEND=foo
 
 use = egg:rhodecode-enterprise-ce
 
 ; enable proxy prefix middleware, defined above
 #filter-with = proxy-prefix
 
 ; #############
 ; DEBUG OPTIONS
 ; #############
 
 pyramid.reload_templates = true
 
 # During development the we want to have the debug toolbar enabled
 pyramid.includes =
     pyramid_debugtoolbar
 
 debugtoolbar.hosts = 0.0.0.0/0
 debugtoolbar.exclude_prefixes =
     /css
     /fonts
     /images
     /js
 
 ## RHODECODE PLUGINS ##
 rhodecode.includes =
     rhodecode.api
 
 
 # api prefix url
 rhodecode.api.url = /_admin/api
 
 ; enable debug style page
 debug_style = true
 
 ; #################
 ; END DEBUG OPTIONS
 ; #################
 
 ; encryption key used to encrypt social plugin tokens,
 ; remote_urls with credentials etc, if not set it defaults to
 ; `beaker.session.secret`
 #rhodecode.encrypted_values.secret =
 
 ; decryption strict mode (enabled by default). It controls if decryption raises
 ; `SignatureVerificationError` in case of wrong key, or damaged encryption data.
 #rhodecode.encrypted_values.strict = false
 
 ; Pick algorithm for encryption. Either fernet (more secure) or aes (default)
 ; fernet is safer, and we strongly recommend switching to it.
 ; Due to backward compatibility aes is used as default.
 #rhodecode.encrypted_values.algorithm = fernet
 
 ; Return gzipped responses from RhodeCode (static files/application)
 gzip_responses = false
 
 ; Auto-generate javascript routes file on startup
 generate_js_files = false
 
 ; System global default language.
 ; All available languages: en (default), be, de, es, fr, it, ja, pl, pt, ru, zh
 lang = en
 
 ; Perform a full repository scan and import on each server start.
 ; Settings this to true could lead to very long startup time.
 startup.import_repos = false
 
 ; URL at which the application is running. This is used for Bootstrapping
 ; requests in context when no web request is available. Used in ishell, or
 ; SSH calls. Set this for events to receive proper url for SSH calls.
 app.base_url = http://rhodecode.local
 
 ; Host at which the Service API is running.
 app.service_api.host = http://rhodecode.local:10020
 
 ; Secret for Service API authentication.
 app.service_api.token =
 
 ; Unique application ID. Should be a random unique string for security.
 app_instance_uuid = rc-production
 
 ; Cut off limit for large diffs (size in bytes). If overall diff size on
 ; commit, or pull request exceeds this limit this diff will be displayed
 ; partially. E.g 512000 == 512Kb
 cut_off_limit_diff = 512000
 
 ; Cut off limit for large files inside diffs (size in bytes). Each individual
 ; file inside diff which exceeds this limit will be displayed partially.
 ;  E.g 128000 == 128Kb
 cut_off_limit_file = 128000
 
 ; Use cached version of vcs repositories everywhere. Recommended to be `true`
 vcs_full_cache = true
 
 ; Force https in RhodeCode, fixes https redirects, assumes it's always https.
 ; Normally this is controlled by proper flags sent from http server such as Nginx or Apache
 force_https = false
 
 ; use Strict-Transport-Security headers
 use_htsts = false
 
 ; Set to true if your repos are exposed using the dumb protocol
 git_update_server_info = false
 
 ; RSS/ATOM feed options
 rss_cut_off_limit = 256000
 rss_items_per_page = 10
 rss_include_diff = false
 
 ; gist URL alias, used to create nicer urls for gist. This should be an
 ; url that does rewrites to _admin/gists/{gistid}.
 ; example: http://gist.rhodecode.org/{gistid}. Empty means use the internal
 ; RhodeCode url, ie. http[s]://rhodecode.server/_admin/gists/{gistid}
 gist_alias_url =
 
 ; List of views (using glob pattern syntax) that AUTH TOKENS could be
 ; used for access.
 ; Adding ?auth_token=TOKEN_HASH to the url authenticates this request as if it
 ; came from the the logged in user who own this authentication token.
 ; Additionally @TOKEN syntax can be used to bound the view to specific
 ; authentication token. Such view would be only accessible when used together
 ; with this authentication token
 ; list of all views can be found under `/_admin/permissions/auth_token_access`
 ; The list should be "," separated and on a single line.
 ; Most common views to enable:
 
 #    RepoCommitsView:repo_commit_download
 #    RepoCommitsView:repo_commit_patch
 #    RepoCommitsView:repo_commit_raw
 #    RepoCommitsView:repo_commit_raw@TOKEN
 #    RepoFilesView:repo_files_diff
 #    RepoFilesView:repo_archivefile
 #    RepoFilesView:repo_file_raw
 #    GistView:*
 api_access_controllers_whitelist =
 
 ; Default encoding used to convert from and to unicode
 ; can be also a comma separated list of encoding in case of mixed encodings
 default_encoding = UTF-8
 
 ; instance-id prefix
 ; a prefix key for this instance used for cache invalidation when running
 ; multiple instances of RhodeCode, make sure it's globally unique for
 ; all running RhodeCode instances. Leave empty if you don't use it
 instance_id =
 
 ; Fallback authentication plugin. Set this to a plugin ID to force the usage
 ; of an authentication plugin also if it is disabled by it's settings.
 ; This could be useful if you are unable to log in to the system due to broken
 ; authentication settings. Then you can enable e.g. the internal RhodeCode auth
 ; module to log in again and fix the settings.
 ; Available builtin plugin IDs (hash is part of the ID):
 ; egg:rhodecode-enterprise-ce#rhodecode
 ; egg:rhodecode-enterprise-ce#pam
 ; egg:rhodecode-enterprise-ce#ldap
 ; egg:rhodecode-enterprise-ce#jasig_cas
 ; egg:rhodecode-enterprise-ce#headers
 ; egg:rhodecode-enterprise-ce#crowd
 
 #rhodecode.auth_plugin_fallback = egg:rhodecode-enterprise-ce#rhodecode
 
 ; Flag to control loading of legacy plugins in py:/path format
 auth_plugin.import_legacy_plugins = true
 
 ; alternative return HTTP header for failed authentication. Default HTTP
 ; response is 401 HTTPUnauthorized. Currently HG clients have troubles with
 ; handling that causing a series of failed authentication calls.
 ; Set this variable to 403 to return HTTPForbidden, or any other HTTP code
 ; This will be served instead of default 401 on bad authentication
 auth_ret_code =
 
 ; use special detection method when serving auth_ret_code, instead of serving
 ; ret_code directly, use 401 initially (Which triggers credentials prompt)
 ; and then serve auth_ret_code to clients
 auth_ret_code_detection = false
 
 ; locking return code. When repository is locked return this HTTP code. 2XX
 ; codes don't break the transactions while 4XX codes do
 lock_ret_code = 423
 
 ; Filesystem location were repositories should be stored
 repo_store.path = /var/opt/rhodecode_repo_store
 
 ; allows to setup custom hooks in settings page
 allow_custom_hooks_settings = true
 
 ; Generated license token required for EE edition license.
 ; New generated token value can be found in Admin > settings > license page.
 license_token =
 
 ; This flag hides sensitive information on the license page such as token, and license data
 license.hide_license_info = false
 
 ; supervisor connection uri, for managing supervisor and logs.
 supervisor.uri =
 
 ; supervisord group name/id we only want this RC instance to handle
 supervisor.group_id = dev
 
 ; Display extended labs settings
 labs_settings_active = true
 
 ; Custom exception store path, defaults to TMPDIR
 ; This is used to store exception from RhodeCode in shared directory
 #exception_tracker.store_path =
 
 ; Send email with exception details when it happens
 #exception_tracker.send_email = false
 
 ; Comma separated list of recipients for exception emails,
 ; e.g [e-mail],[e-mail]
 ; Can be left empty, then emails will be sent to ALL super-admins
 #exception_tracker.send_email_recipients =
 
 ; optional prefix to Add to email Subject
 #exception_tracker.email_prefix = [RHODECODE ERROR]
 
 ; File store configuration. This is used to store and serve uploaded files
 file_store.enabled = true
 
 ; Storage backend, available options are: local
 file_store.backend = local
 
 ; path to store the uploaded binaries and artifacts
 file_store.storage_path = /var/opt/rhodecode_data/file_store
 
 ; Uncomment and set this path to control settings for archive download cache.
 ; Generated repo archives will be cached at this location
 ; and served from the cache during subsequent requests for the same archive of
 ; the repository. This path is important to be shared across filesystems and with
 ; RhodeCode and vcsserver
 
+; Redis url to acquire/check generation of archives locks
+archive_cache.locking.url = redis://redis:6379/1
+
+; Storage backend, only 'filesystem' is available now
+archive_cache.backend.type = filesystem
+
 ; Default is $cache_dir/archive_cache if not set
-archive_cache.store_dir = /var/opt/rhodecode_data/tarballcache
+archive_cache.filesystem.store_dir = /var/opt/rhodecode_data/archive_cache
 
 ; The limit in GB sets how much data we cache before recycling last used, defaults to 10 gb
-archive_cache.cache_size_gb = 10
+archive_cache.filesystem.cache_size_gb = 1
+
+; Eviction policy used to clear out after cache_size_gb limit is reached
+archive_cache.filesystem.eviction_policy = least-recently-stored
 
 ; By default cache uses sharding technique, this specifies how many shards are there
-archive_cache.cache_shards = 4
+archive_cache.filesystem.cache_shards = 8
+
 
 ; #############
 ; CELERY CONFIG
 ; #############
 
 ; manually run celery: /path/to/celery worker --task-events --beat --app rhodecode.lib.celerylib.loader --scheduler rhodecode.lib.celerylib.scheduler.RcScheduler --loglevel DEBUG --ini /path/to/rhodecode.ini
 
 use_celery = true
 
 ; path to store schedule database
 #celerybeat-schedule.path =
 
 ; connection url to the message broker (default redis)
 celery.broker_url = redis://redis:6379/8
 
 ; results backend to get results for (default redis)
 celery.result_backend = redis://redis:6379/8
 
 ; rabbitmq example
 #celery.broker_url = amqp://rabbitmq:qweqwe@localhost:5672/rabbitmqhost
 
 ; maximum tasks to execute before worker restart
 celery.max_tasks_per_child = 20
 
 ; tasks will never be sent to the queue, but executed locally instead.
 celery.task_always_eager = false
 
 ; #############
 ; DOGPILE CACHE
 ; #############
 
 ; Default cache dir for caches. Putting this into a ramdisk can boost performance.
 ; eg. /tmpfs/data_ramdisk, however this directory might require large amount of space
 cache_dir = /var/opt/rhodecode_data
 
 ; *********************************************
 ; `sql_cache_short` cache for heavy SQL queries
 ; Only supported backend is `memory_lru`
 ; *********************************************
 rc_cache.sql_cache_short.backend = dogpile.cache.rc.memory_lru
 rc_cache.sql_cache_short.expiration_time = 30
 
 
 ; *****************************************************
 ; `cache_repo_longterm` cache for repo object instances
 ; Only supported backend is `memory_lru`
 ; *****************************************************
 rc_cache.cache_repo_longterm.backend = dogpile.cache.rc.memory_lru
 ; by default we use 30 Days, cache is still invalidated on push
 rc_cache.cache_repo_longterm.expiration_time = 2592000
 ; max items in LRU cache, set to smaller number to save memory, and expire last used caches
 rc_cache.cache_repo_longterm.max_size = 10000
 
 
 ; *********************************************
 ; `cache_general` cache for general purpose use
 ; for simplicity use rc.file_namespace backend,
 ; for performance and scale use rc.redis
 ; *********************************************
 rc_cache.cache_general.backend = dogpile.cache.rc.file_namespace
 rc_cache.cache_general.expiration_time = 43200
 ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set
 #rc_cache.cache_general.arguments.filename = /tmp/cache_general_db
 
 ; alternative `cache_general` redis backend with distributed lock
 #rc_cache.cache_general.backend = dogpile.cache.rc.redis
 #rc_cache.cache_general.expiration_time = 300
 
 ; redis_expiration_time needs to be greater then expiration_time
 #rc_cache.cache_general.arguments.redis_expiration_time = 7200
 
 #rc_cache.cache_general.arguments.host = localhost
 #rc_cache.cache_general.arguments.port = 6379
 #rc_cache.cache_general.arguments.db = 0
 #rc_cache.cache_general.arguments.socket_timeout = 30
 ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends
 #rc_cache.cache_general.arguments.distributed_lock = true
 
 ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen
 #rc_cache.cache_general.arguments.lock_auto_renewal = true
 
 ; *************************************************
 ; `cache_perms` cache for permission tree, auth TTL
 ; for simplicity use rc.file_namespace backend,
 ; for performance and scale use rc.redis
 ; *************************************************
 rc_cache.cache_perms.backend = dogpile.cache.rc.file_namespace
 rc_cache.cache_perms.expiration_time = 3600
 ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set
 #rc_cache.cache_perms.arguments.filename = /tmp/cache_perms_db
 
 ; alternative `cache_perms` redis backend with distributed lock
 #rc_cache.cache_perms.backend = dogpile.cache.rc.redis
 #rc_cache.cache_perms.expiration_time = 300
 
 ; redis_expiration_time needs to be greater then expiration_time
 #rc_cache.cache_perms.arguments.redis_expiration_time = 7200
 
 #rc_cache.cache_perms.arguments.host = localhost
 #rc_cache.cache_perms.arguments.port = 6379
 #rc_cache.cache_perms.arguments.db = 0
 #rc_cache.cache_perms.arguments.socket_timeout = 30
 ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends
 #rc_cache.cache_perms.arguments.distributed_lock = true
 
 ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen
 #rc_cache.cache_perms.arguments.lock_auto_renewal = true
 
 ; ***************************************************
 ; `cache_repo` cache for file tree, Readme, RSS FEEDS
 ; for simplicity use rc.file_namespace backend,
 ; for performance and scale use rc.redis
 ; ***************************************************
 rc_cache.cache_repo.backend = dogpile.cache.rc.file_namespace
 rc_cache.cache_repo.expiration_time = 2592000
 ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set
 #rc_cache.cache_repo.arguments.filename = /tmp/cache_repo_db
 
 ; alternative `cache_repo` redis backend with distributed lock
 #rc_cache.cache_repo.backend = dogpile.cache.rc.redis
 #rc_cache.cache_repo.expiration_time = 2592000
 
 ; redis_expiration_time needs to be greater then expiration_time
 #rc_cache.cache_repo.arguments.redis_expiration_time = 2678400
 
 #rc_cache.cache_repo.arguments.host = localhost
 #rc_cache.cache_repo.arguments.port = 6379
 #rc_cache.cache_repo.arguments.db = 1
 #rc_cache.cache_repo.arguments.socket_timeout = 30
 ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends
 #rc_cache.cache_repo.arguments.distributed_lock = true
 
 ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen
 #rc_cache.cache_repo.arguments.lock_auto_renewal = true
 
 ; ##############
 ; BEAKER SESSION
 ; ##############
 
 ; beaker.session.type is type of storage options for the logged users sessions. Current allowed
 ; types are file, ext:redis, ext:database, ext:memcached
 ; Fastest ones are ext:redis and ext:database, DO NOT use memory type for session
 #beaker.session.type = file
 #beaker.session.data_dir = %(here)s/data/sessions
 
 ; Redis based sessions
 beaker.session.type = ext:redis
 beaker.session.url = redis://redis:6379/2
 
 ; DB based session, fast, and allows easy management over logged in users
 #beaker.session.type = ext:database
 #beaker.session.table_name = db_session
 #beaker.session.sa.url = postgresql://postgres:secret@localhost/rhodecode
 #beaker.session.sa.url = mysql://root:secret@127.0.0.1/rhodecode
 #beaker.session.sa.pool_recycle = 3600
 #beaker.session.sa.echo = false
 
 beaker.session.key = rhodecode
 beaker.session.secret = develop-rc-uytcxaz
 beaker.session.lock_dir = /data_ramdisk/lock
 
 ; Secure encrypted cookie. Requires AES and AES python libraries
 ; you must disable beaker.session.secret to use this
 #beaker.session.encrypt_key = key_for_encryption
 #beaker.session.validate_key = validation_key
 
 ; Sets session as invalid (also logging out user) if it haven not been
 ; accessed for given amount of time in seconds
 beaker.session.timeout = 2592000
 beaker.session.httponly = true
 
 ; Path to use for the cookie. Set to prefix if you use prefix middleware
 #beaker.session.cookie_path = /custom_prefix
 
 ; Set https secure cookie
 beaker.session.secure = false
 
 ; default cookie expiration time in seconds, set to `true` to set expire
 ; at browser close
 #beaker.session.cookie_expires = 3600
 
 ; #############################
 ; SEARCH INDEXING CONFIGURATION
 ; #############################
 
 ; Full text search indexer is available in rhodecode-tools under
 ; `rhodecode-tools index` command
 
 ; WHOOSH Backend, doesn't require additional services to run
 ; it works good with few dozen repos
 search.module = rhodecode.lib.index.whoosh
 search.location = %(here)s/data/index
 
 ; ####################
 ; CHANNELSTREAM CONFIG
 ; ####################
 
 ; channelstream enables persistent connections and live notification
 ; in the system. It's also used by the chat system
 
 channelstream.enabled = true
 
 ; server address for channelstream server on the backend
 channelstream.server = channelstream:9800
 
 ; location of the channelstream server from outside world
 ; use ws:// for http or wss:// for https. This address needs to be handled
 ; by external HTTP server such as Nginx or Apache
 ; see Nginx/Apache configuration examples in our docs
 channelstream.ws_url = ws://rhodecode.yourserver.com/_channelstream
 channelstream.secret = ENV_GENERATED
 channelstream.history.location = /var/opt/rhodecode_data/channelstream_history
 
 ; Internal application path that Javascript uses to connect into.
 ; If you use proxy-prefix the prefix should be added before /_channelstream
 channelstream.proxy_path = /_channelstream
 
 
 ; ##############################
 ; MAIN RHODECODE DATABASE CONFIG
 ; ##############################
 
 #sqlalchemy.db1.url = sqlite:///%(here)s/rhodecode.db?timeout=30
 #sqlalchemy.db1.url = postgresql://postgres:qweqwe@localhost/rhodecode
 #sqlalchemy.db1.url = mysql://root:qweqwe@localhost/rhodecode?charset=utf8
 ; pymysql is an alternative driver for MySQL, use in case of problems with default one
 #sqlalchemy.db1.url = mysql+pymysql://root:qweqwe@localhost/rhodecode
 
 sqlalchemy.db1.url = sqlite:///%(here)s/rhodecode.db?timeout=30
 
 ; see sqlalchemy docs for other advanced settings
 ; print the sql statements to output
 sqlalchemy.db1.echo = false
 
 ; recycle the connections after this amount of seconds
 sqlalchemy.db1.pool_recycle = 3600
 
 ; the number of connections to keep open inside the connection pool.
 ; 0 indicates no limit
 ; the general calculus with gevent is:
 ; if your system allows 500 concurrent greenlets (max_connections) that all do database access,
 ; then increase pool size + max overflow so that they add up to 500.
 #sqlalchemy.db1.pool_size = 5
 
 ; The number of connections to allow in connection pool "overflow", that is
 ; connections that can be opened above and beyond the pool_size setting,
 ; which defaults to five.
 #sqlalchemy.db1.max_overflow = 10
 
 ; Connection check ping, used to detect broken database connections
 ; could be enabled to better handle cases if MySQL has gone away errors
 #sqlalchemy.db1.ping_connection = true
 
 ; ##########
 ; VCS CONFIG
 ; ##########
 vcs.server.enable = true
 vcs.server = vcsserver:10010
 
 ; Web server connectivity protocol, responsible for web based VCS operations
 ; Available protocols are:
 ; `http` - use http-rpc backend (default)
 vcs.server.protocol = http
 
 ; Push/Pull operations protocol, available options are:
 ; `http` - use http-rpc backend (default)
 vcs.scm_app_implementation = http
 
 ; Push/Pull operations hooks protocol, available options are:
 ; `http` - use http-rpc backend (default)
 ; `celery` - use celery based hooks
 vcs.hooks.protocol = http
 
 ; Host on which this instance is listening for hooks. vcsserver will call this host to pull/push hooks so it should be
 ; accessible via network.
 ; Use vcs.hooks.host = "*" to bind to current hostname (for Docker)
 vcs.hooks.host = *
 
 ; Start VCSServer with this instance as a subprocess, useful for development
 vcs.start_server = false
 
 ; List of enabled VCS backends, available options are:
 ; `hg`  - mercurial
 ; `git` - git
 ; `svn` - subversion
 vcs.backends = hg, git, svn
 
 ; Wait this number of seconds before killing connection to the vcsserver
 vcs.connection_timeout = 3600
 
 ; Cache flag to cache vcsserver remote calls locally
 ; It uses cache_region `cache_repo`
 vcs.methods.cache = true
 
 ; ####################################################
 ; Subversion proxy support (mod_dav_svn)
 ; Maps RhodeCode repo groups into SVN paths for Apache
 ; ####################################################
 
 ; Compatibility version when creating SVN repositories. Defaults to newest version when commented out.
 ; Set a numeric version for your current SVN e.g 1.8, or 1.12
 ; Legacy available options are: pre-1.4-compatible, pre-1.5-compatible, pre-1.6-compatible, pre-1.8-compatible, pre-1.9-compatible
 #vcs.svn.compatible_version = 1.8
 
 ; Enable SVN proxy of requests over HTTP
 vcs.svn.proxy.enabled = true
 
 ; host to connect to running SVN subsystem
 vcs.svn.proxy.host = http://svn:8090
 
 ; Enable or disable the config file generation.
 svn.proxy.generate_config = true
 
 ; Generate config file with `SVNListParentPath` set to `On`.
 svn.proxy.list_parent_path = true
 
 ; Set location and file name of generated config file.
 svn.proxy.config_file_path = /etc/rhodecode/conf/svn/mod_dav_svn.conf
 
 ; alternative mod_dav config template. This needs to be a valid mako template
 ; Example template can be found in the source code:
 ; rhodecode/apps/svn_support/templates/mod-dav-svn.conf.mako
 #svn.proxy.config_template = ~/.rccontrol/enterprise-1/custom_svn_conf.mako
 
 ; Used as a prefix to the `Location` block in the generated config file.
 ; In most cases it should be set to `/`.
 svn.proxy.location_root = /
 
 ; Command to reload the mod dav svn configuration on change.
 ; Example: `/etc/init.d/apache2 reload` or /home/USER/apache_reload.sh
 ; Make sure user who runs RhodeCode process is allowed to reload Apache
 #svn.proxy.reload_cmd = /etc/init.d/apache2 reload
 
 ; If the timeout expires before the reload command finishes, the command will
 ; be killed. Setting it to zero means no timeout. Defaults to 10 seconds.
 #svn.proxy.reload_timeout = 10
 
 ; ####################
 ; SSH Support Settings
 ; ####################
 
 ; Defines if a custom authorized_keys file should be created and written on
 ; any change user ssh keys. Setting this to false also disables possibility
 ; of adding SSH keys by users from web interface. Super admins can still
 ; manage SSH Keys.
 ssh.generate_authorized_keyfile = true
 
 ; Options for ssh, default is `no-pty,no-port-forwarding,no-X11-forwarding,no-agent-forwarding`
 # ssh.authorized_keys_ssh_opts =
 
 ; Path to the authorized_keys file where the generate entries are placed.
 ; It is possible to have multiple key files specified in `sshd_config` e.g.
 ; AuthorizedKeysFile %h/.ssh/authorized_keys %h/.ssh/authorized_keys_rhodecode
 ssh.authorized_keys_file_path = /etc/rhodecode/conf/ssh/authorized_keys_rhodecode
 
 ; Command to execute the SSH wrapper. The binary is available in the
 ; RhodeCode installation directory.
 ; legacy: /usr/local/bin/rhodecode_bin/bin/rc-ssh-wrapper
 ; new rewrite: /usr/local/bin/rhodecode_bin/bin/rc-ssh-wrapper-v2
 ssh.wrapper_cmd = /usr/local/bin/rhodecode_bin/bin/rc-ssh-wrapper
 
 ; Allow shell when executing the ssh-wrapper command
 ssh.wrapper_cmd_allow_shell = false
 
 ; Enables logging, and detailed output send back to the client during SSH
 ; operations. Useful for debugging, shouldn't be used in production.
 ssh.enable_debug_logging = true
 
 ; Paths to binary executable, by default they are the names, but we can
 ; override them if we want to use a custom one
 ssh.executable.hg = /usr/local/bin/rhodecode_bin/vcs_bin/hg
 ssh.executable.git = /usr/local/bin/rhodecode_bin/vcs_bin/git
 ssh.executable.svn = /usr/local/bin/rhodecode_bin/vcs_bin/svnserve
 
 ; Enables SSH key generator web interface. Disabling this still allows users
 ; to add their own keys.
 ssh.enable_ui_key_generator = true
 
 ; Statsd client config, this is used to send metrics to statsd
 ; We recommend setting statsd_exported and scrape them using Prometheus
 #statsd.enabled = false
 #statsd.statsd_host = 0.0.0.0
 #statsd.statsd_port = 8125
 #statsd.statsd_prefix =
 #statsd.statsd_ipv6 = false
 
 ; configure logging automatically at server startup set to false
 ; to use the below custom logging config.
 ; RC_LOGGING_FORMATTER
 ; RC_LOGGING_LEVEL
 ; env variables can control the settings for logging in case of autoconfigure
 
 #logging.autoconfigure = true
 
 ; specify your own custom logging config file to configure logging
 #logging.logging_conf_file = /path/to/custom_logging.ini
 
 ; Dummy marker to add new entries after.
 ; Add any custom entries below. Please don't remove this marker.
 custom.conf = 1
 
 
 ; #####################
 ; LOGGING CONFIGURATION
 ; #####################
 
 [loggers]
 keys = root, sqlalchemy, beaker, celery, rhodecode, ssh_wrapper
 
 [handlers]
 keys = console, console_sql
 
 [formatters]
 keys = generic, json, color_formatter, color_formatter_sql
 
 ; #######
 ; LOGGERS
 ; #######
 [logger_root]
 level = NOTSET
 handlers = console
 
 [logger_sqlalchemy]
 level = INFO
 handlers = console_sql
 qualname = sqlalchemy.engine
 propagate = 0
 
 [logger_beaker]
 level = DEBUG
 handlers =
 qualname = beaker.container
 propagate = 1
 
 [logger_rhodecode]
 level = DEBUG
 handlers =
 qualname = rhodecode
 propagate = 1
 
 [logger_ssh_wrapper]
 level = DEBUG
 handlers =
 qualname = ssh_wrapper
 propagate = 1
 
 [logger_celery]
 level = DEBUG
 handlers =
 qualname = celery
 
 
 ; ########
 ; HANDLERS
 ; ########
 
 [handler_console]
 class = StreamHandler
 args = (sys.stderr, )
 level = DEBUG
 ; To enable JSON formatted logs replace 'generic/color_formatter' with 'json'
 ; This allows sending properly formatted logs to grafana loki or elasticsearch
 formatter = color_formatter
 
 [handler_console_sql]
 ; "level = DEBUG" logs SQL queries and results.
 ; "level = INFO" logs SQL queries.
 ; "level = WARN" logs neither.  (Recommended for production systems.)
 class = StreamHandler
 args = (sys.stderr, )
 level = WARN
 ; To enable JSON formatted logs replace 'generic/color_formatter_sql' with 'json'
 ; This allows sending properly formatted logs to grafana loki or elasticsearch
 formatter = color_formatter_sql
 
 ; ##########
 ; FORMATTERS
 ; ##########
 
 [formatter_generic]
 class = rhodecode.lib.logging_formatter.ExceptionAwareFormatter
 format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s
 datefmt = %Y-%m-%d %H:%M:%S
 
 [formatter_color_formatter]
 class = rhodecode.lib.logging_formatter.ColorFormatter
 format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s
 datefmt = %Y-%m-%d %H:%M:%S
 
 [formatter_color_formatter_sql]
 class = rhodecode.lib.logging_formatter.ColorFormatterSql
 format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s
 datefmt = %Y-%m-%d %H:%M:%S
 
 [formatter_json]
 format = %(timestamp)s %(levelname)s %(name)s %(message)s %(req_id)s
 class = rhodecode.lib._vendor.jsonlogger.JsonFormatter

configs/production.ini

@@ -1,771 +1,781 b''
 
 ; #########################################
 ; RHODECODE COMMUNITY EDITION CONFIGURATION
 ; #########################################
 
 [DEFAULT]
 ; Debug flag sets all loggers to debug, and enables request tracking
 debug = false
 
 ; ########################################################################
 ; EMAIL CONFIGURATION
 ; These settings will be used by the RhodeCode mailing system
 ; ########################################################################
 
 ; prefix all emails subjects with given prefix, helps filtering out emails
 #email_prefix = [RhodeCode]
 
 ; email FROM address all mails will be sent
 #app_email_from = rhodecode-noreply@localhost
 
 #smtp_server = mail.server.com
 #smtp_username =
 #smtp_password =
 #smtp_port =
 #smtp_use_tls = false
 #smtp_use_ssl = true
 
 [server:main]
 ; COMMON HOST/IP CONFIG, This applies mostly to develop setup,
 ; Host port for gunicorn are controlled by gunicorn_conf.py
 host = 127.0.0.1
 port = 10020
 
 
 ; ###########################
 ; GUNICORN APPLICATION SERVER
 ; ###########################
 
 ; run with gunicorn --paste rhodecode.ini --config gunicorn_conf.py
 
 ; Module to use, this setting shouldn't be changed
 use = egg:gunicorn#main
 
 ; Prefix middleware for RhodeCode.
 ; recommended when using proxy setup.
 ; allows to set RhodeCode under a prefix in server.
 ; eg https://server.com/custom_prefix. Enable `filter-with =` option below as well.
 ; And set your prefix like: `prefix = /custom_prefix`
 ; be sure to also set beaker.session.cookie_path = /custom_prefix if you need
 ; to make your cookies only work on prefix url
 [filter:proxy-prefix]
 use = egg:PasteDeploy#prefix
 prefix = /
 
 [app:main]
 ; The %(here)s variable will be replaced with the absolute path of parent directory
 ; of this file
 ; Each option in the app:main can be override by an environmental variable
 ;
 ;To override an option:
 ;
 ;RC_<KeyName>
 ;Everything should be uppercase, . and - should be replaced by _.
 ;For example, if you have these configuration settings:
 ;rc_cache.repo_object.backend = foo
 ;can be overridden by
 ;export RC_CACHE_REPO_OBJECT_BACKEND=foo
 
 use = egg:rhodecode-enterprise-ce
 
 ; enable proxy prefix middleware, defined above
 #filter-with = proxy-prefix
 
 ; encryption key used to encrypt social plugin tokens,
 ; remote_urls with credentials etc, if not set it defaults to
 ; `beaker.session.secret`
 #rhodecode.encrypted_values.secret =
 
 ; decryption strict mode (enabled by default). It controls if decryption raises
 ; `SignatureVerificationError` in case of wrong key, or damaged encryption data.
 #rhodecode.encrypted_values.strict = false
 
 ; Pick algorithm for encryption. Either fernet (more secure) or aes (default)
 ; fernet is safer, and we strongly recommend switching to it.
 ; Due to backward compatibility aes is used as default.
 #rhodecode.encrypted_values.algorithm = fernet
 
 ; Return gzipped responses from RhodeCode (static files/application)
 gzip_responses = false
 
 ; Auto-generate javascript routes file on startup
 generate_js_files = false
 
 ; System global default language.
 ; All available languages: en (default), be, de, es, fr, it, ja, pl, pt, ru, zh
 lang = en
 
 ; Perform a full repository scan and import on each server start.
 ; Settings this to true could lead to very long startup time.
 startup.import_repos = false
 
 ; URL at which the application is running. This is used for Bootstrapping
 ; requests in context when no web request is available. Used in ishell, or
 ; SSH calls. Set this for events to receive proper url for SSH calls.
 app.base_url = http://rhodecode.local
 
 ; Host at which the Service API is running.
 app.service_api.host = http://rhodecode.local:10020
 
 ; Secret for Service API authentication.
 app.service_api.token =
 
 ; Unique application ID. Should be a random unique string for security.
 app_instance_uuid = rc-production
 
 ; Cut off limit for large diffs (size in bytes). If overall diff size on
 ; commit, or pull request exceeds this limit this diff will be displayed
 ; partially. E.g 512000 == 512Kb
 cut_off_limit_diff = 512000
 
 ; Cut off limit for large files inside diffs (size in bytes). Each individual
 ; file inside diff which exceeds this limit will be displayed partially.
 ;  E.g 128000 == 128Kb
 cut_off_limit_file = 128000
 
 ; Use cached version of vcs repositories everywhere. Recommended to be `true`
 vcs_full_cache = true
 
 ; Force https in RhodeCode, fixes https redirects, assumes it's always https.
 ; Normally this is controlled by proper flags sent from http server such as Nginx or Apache
 force_https = false
 
 ; use Strict-Transport-Security headers
 use_htsts = false
 
 ; Set to true if your repos are exposed using the dumb protocol
 git_update_server_info = false
 
 ; RSS/ATOM feed options
 rss_cut_off_limit = 256000
 rss_items_per_page = 10
 rss_include_diff = false
 
 ; gist URL alias, used to create nicer urls for gist. This should be an
 ; url that does rewrites to _admin/gists/{gistid}.
 ; example: http://gist.rhodecode.org/{gistid}. Empty means use the internal
 ; RhodeCode url, ie. http[s]://rhodecode.server/_admin/gists/{gistid}
 gist_alias_url =
 
 ; List of views (using glob pattern syntax) that AUTH TOKENS could be
 ; used for access.
 ; Adding ?auth_token=TOKEN_HASH to the url authenticates this request as if it
 ; came from the the logged in user who own this authentication token.
 ; Additionally @TOKEN syntax can be used to bound the view to specific
 ; authentication token. Such view would be only accessible when used together
 ; with this authentication token
 ; list of all views can be found under `/_admin/permissions/auth_token_access`
 ; The list should be "," separated and on a single line.
 ; Most common views to enable:
 
 #    RepoCommitsView:repo_commit_download
 #    RepoCommitsView:repo_commit_patch
 #    RepoCommitsView:repo_commit_raw
 #    RepoCommitsView:repo_commit_raw@TOKEN
 #    RepoFilesView:repo_files_diff
 #    RepoFilesView:repo_archivefile
 #    RepoFilesView:repo_file_raw
 #    GistView:*
 api_access_controllers_whitelist =
 
 ; Default encoding used to convert from and to unicode
 ; can be also a comma separated list of encoding in case of mixed encodings
 default_encoding = UTF-8
 
 ; instance-id prefix
 ; a prefix key for this instance used for cache invalidation when running
 ; multiple instances of RhodeCode, make sure it's globally unique for
 ; all running RhodeCode instances. Leave empty if you don't use it
 instance_id =
 
 ; Fallback authentication plugin. Set this to a plugin ID to force the usage
 ; of an authentication plugin also if it is disabled by it's settings.
 ; This could be useful if you are unable to log in to the system due to broken
 ; authentication settings. Then you can enable e.g. the internal RhodeCode auth
 ; module to log in again and fix the settings.
 ; Available builtin plugin IDs (hash is part of the ID):
 ; egg:rhodecode-enterprise-ce#rhodecode
 ; egg:rhodecode-enterprise-ce#pam
 ; egg:rhodecode-enterprise-ce#ldap
 ; egg:rhodecode-enterprise-ce#jasig_cas
 ; egg:rhodecode-enterprise-ce#headers
 ; egg:rhodecode-enterprise-ce#crowd
 
 #rhodecode.auth_plugin_fallback = egg:rhodecode-enterprise-ce#rhodecode
 
 ; Flag to control loading of legacy plugins in py:/path format
 auth_plugin.import_legacy_plugins = true
 
 ; alternative return HTTP header for failed authentication. Default HTTP
 ; response is 401 HTTPUnauthorized. Currently HG clients have troubles with
 ; handling that causing a series of failed authentication calls.
 ; Set this variable to 403 to return HTTPForbidden, or any other HTTP code
 ; This will be served instead of default 401 on bad authentication
 auth_ret_code =
 
 ; use special detection method when serving auth_ret_code, instead of serving
 ; ret_code directly, use 401 initially (Which triggers credentials prompt)
 ; and then serve auth_ret_code to clients
 auth_ret_code_detection = false
 
 ; locking return code. When repository is locked return this HTTP code. 2XX
 ; codes don't break the transactions while 4XX codes do
 lock_ret_code = 423
 
 ; Filesystem location were repositories should be stored
 repo_store.path = /var/opt/rhodecode_repo_store
 
 ; allows to setup custom hooks in settings page
 allow_custom_hooks_settings = true
 
 ; Generated license token required for EE edition license.
 ; New generated token value can be found in Admin > settings > license page.
 license_token =
 
 ; This flag hides sensitive information on the license page such as token, and license data
 license.hide_license_info = false
 
 ; supervisor connection uri, for managing supervisor and logs.
 supervisor.uri =
 
 ; supervisord group name/id we only want this RC instance to handle
 supervisor.group_id = prod
 
 ; Display extended labs settings
 labs_settings_active = true
 
 ; Custom exception store path, defaults to TMPDIR
 ; This is used to store exception from RhodeCode in shared directory
 #exception_tracker.store_path =
 
 ; Send email with exception details when it happens
 #exception_tracker.send_email = false
 
 ; Comma separated list of recipients for exception emails,
 ; e.g [e-mail],[e-mail]
 ; Can be left empty, then emails will be sent to ALL super-admins
 #exception_tracker.send_email_recipients =
 
 ; optional prefix to Add to email Subject
 #exception_tracker.email_prefix = [RHODECODE ERROR]
 
 ; File store configuration. This is used to store and serve uploaded files
 file_store.enabled = true
 
 ; Storage backend, available options are: local
 file_store.backend = local
 
 ; path to store the uploaded binaries and artifacts
 file_store.storage_path = /var/opt/rhodecode_data/file_store
 
 ; Uncomment and set this path to control settings for archive download cache.
 ; Generated repo archives will be cached at this location
 ; and served from the cache during subsequent requests for the same archive of
 ; the repository. This path is important to be shared across filesystems and with
 ; RhodeCode and vcsserver
 
+; Redis url to acquire/check generation of archives locks
+archive_cache.locking.url = redis://redis:6379/1
+
+; Storage backend, only 'filesystem' is available now
+archive_cache.backend.type = filesystem
+
 ; Default is $cache_dir/archive_cache if not set
-archive_cache.store_dir = /var/opt/rhodecode_data/tarballcache
+archive_cache.filesystem.store_dir = /var/opt/rhodecode_data/archive_cache
 
 ; The limit in GB sets how much data we cache before recycling last used, defaults to 10 gb
-archive_cache.cache_size_gb = 40
+archive_cache.filesystem.cache_size_gb = 40
+
+; Eviction policy used to clear out after cache_size_gb limit is reached
+archive_cache.filesystem.eviction_policy = least-recently-stored
 
 ; By default cache uses sharding technique, this specifies how many shards are there
-archive_cache.cache_shards = 4
+archive_cache.filesystem.cache_shards = 8
+
 
 ; #############
 ; CELERY CONFIG
 ; #############
 
 ; manually run celery: /path/to/celery worker --task-events --beat --app rhodecode.lib.celerylib.loader --scheduler rhodecode.lib.celerylib.scheduler.RcScheduler --loglevel DEBUG --ini /path/to/rhodecode.ini
 
 use_celery = true
 
 ; path to store schedule database
 #celerybeat-schedule.path =
 
 ; connection url to the message broker (default redis)
 celery.broker_url = redis://redis:6379/8
 
 ; results backend to get results for (default redis)
 celery.result_backend = redis://redis:6379/8
 
 ; rabbitmq example
 #celery.broker_url = amqp://rabbitmq:qweqwe@localhost:5672/rabbitmqhost
 
 ; maximum tasks to execute before worker restart
 celery.max_tasks_per_child = 20
 
 ; tasks will never be sent to the queue, but executed locally instead.
 celery.task_always_eager = false
 
 ; #############
 ; DOGPILE CACHE
 ; #############
 
 ; Default cache dir for caches. Putting this into a ramdisk can boost performance.
 ; eg. /tmpfs/data_ramdisk, however this directory might require large amount of space
 cache_dir = /var/opt/rhodecode_data
 
 ; *********************************************
 ; `sql_cache_short` cache for heavy SQL queries
 ; Only supported backend is `memory_lru`
 ; *********************************************
 rc_cache.sql_cache_short.backend = dogpile.cache.rc.memory_lru
 rc_cache.sql_cache_short.expiration_time = 30
 
 
 ; *****************************************************
 ; `cache_repo_longterm` cache for repo object instances
 ; Only supported backend is `memory_lru`
 ; *****************************************************
 rc_cache.cache_repo_longterm.backend = dogpile.cache.rc.memory_lru
 ; by default we use 30 Days, cache is still invalidated on push
 rc_cache.cache_repo_longterm.expiration_time = 2592000
 ; max items in LRU cache, set to smaller number to save memory, and expire last used caches
 rc_cache.cache_repo_longterm.max_size = 10000
 
 
 ; *********************************************
 ; `cache_general` cache for general purpose use
 ; for simplicity use rc.file_namespace backend,
 ; for performance and scale use rc.redis
 ; *********************************************
 rc_cache.cache_general.backend = dogpile.cache.rc.file_namespace
 rc_cache.cache_general.expiration_time = 43200
 ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set
 #rc_cache.cache_general.arguments.filename = /tmp/cache_general_db
 
 ; alternative `cache_general` redis backend with distributed lock
 #rc_cache.cache_general.backend = dogpile.cache.rc.redis
 #rc_cache.cache_general.expiration_time = 300
 
 ; redis_expiration_time needs to be greater then expiration_time
 #rc_cache.cache_general.arguments.redis_expiration_time = 7200
 
 #rc_cache.cache_general.arguments.host = localhost
 #rc_cache.cache_general.arguments.port = 6379
 #rc_cache.cache_general.arguments.db = 0
 #rc_cache.cache_general.arguments.socket_timeout = 30
 ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends
 #rc_cache.cache_general.arguments.distributed_lock = true
 
 ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen
 #rc_cache.cache_general.arguments.lock_auto_renewal = true
 
 ; *************************************************
 ; `cache_perms` cache for permission tree, auth TTL
 ; for simplicity use rc.file_namespace backend,
 ; for performance and scale use rc.redis
 ; *************************************************
 rc_cache.cache_perms.backend = dogpile.cache.rc.file_namespace
 rc_cache.cache_perms.expiration_time = 3600
 ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set
 #rc_cache.cache_perms.arguments.filename = /tmp/cache_perms_db
 
 ; alternative `cache_perms` redis backend with distributed lock
 #rc_cache.cache_perms.backend = dogpile.cache.rc.redis
 #rc_cache.cache_perms.expiration_time = 300
 
 ; redis_expiration_time needs to be greater then expiration_time
 #rc_cache.cache_perms.arguments.redis_expiration_time = 7200
 
 #rc_cache.cache_perms.arguments.host = localhost
 #rc_cache.cache_perms.arguments.port = 6379
 #rc_cache.cache_perms.arguments.db = 0
 #rc_cache.cache_perms.arguments.socket_timeout = 30
 ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends
 #rc_cache.cache_perms.arguments.distributed_lock = true
 
 ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen
 #rc_cache.cache_perms.arguments.lock_auto_renewal = true
 
 ; ***************************************************
 ; `cache_repo` cache for file tree, Readme, RSS FEEDS
 ; for simplicity use rc.file_namespace backend,
 ; for performance and scale use rc.redis
 ; ***************************************************
 rc_cache.cache_repo.backend = dogpile.cache.rc.file_namespace
 rc_cache.cache_repo.expiration_time = 2592000
 ; file cache store path. Defaults to `cache_dir =` value or tempdir if both values are not set
 #rc_cache.cache_repo.arguments.filename = /tmp/cache_repo_db
 
 ; alternative `cache_repo` redis backend with distributed lock
 #rc_cache.cache_repo.backend = dogpile.cache.rc.redis
 #rc_cache.cache_repo.expiration_time = 2592000
 
 ; redis_expiration_time needs to be greater then expiration_time
 #rc_cache.cache_repo.arguments.redis_expiration_time = 2678400
 
 #rc_cache.cache_repo.arguments.host = localhost
 #rc_cache.cache_repo.arguments.port = 6379
 #rc_cache.cache_repo.arguments.db = 1
 #rc_cache.cache_repo.arguments.socket_timeout = 30
 ; more Redis options: https://dogpilecache.sqlalchemy.org/en/latest/api.html#redis-backends
 #rc_cache.cache_repo.arguments.distributed_lock = true
 
 ; auto-renew lock to prevent stale locks, slower but safer. Use only if problems happen
 #rc_cache.cache_repo.arguments.lock_auto_renewal = true
 
 ; ##############
 ; BEAKER SESSION
 ; ##############
 
 ; beaker.session.type is type of storage options for the logged users sessions. Current allowed
 ; types are file, ext:redis, ext:database, ext:memcached
 ; Fastest ones are ext:redis and ext:database, DO NOT use memory type for session
 #beaker.session.type = file
 #beaker.session.data_dir = %(here)s/data/sessions
 
 ; Redis based sessions
 beaker.session.type = ext:redis
 beaker.session.url = redis://redis:6379/2
 
 ; DB based session, fast, and allows easy management over logged in users
 #beaker.session.type = ext:database
 #beaker.session.table_name = db_session
 #beaker.session.sa.url = postgresql://postgres:secret@localhost/rhodecode
 #beaker.session.sa.url = mysql://root:secret@127.0.0.1/rhodecode
 #beaker.session.sa.pool_recycle = 3600
 #beaker.session.sa.echo = false
 
 beaker.session.key = rhodecode
 beaker.session.secret = production-rc-uytcxaz
 beaker.session.lock_dir = /data_ramdisk/lock
 
 ; Secure encrypted cookie. Requires AES and AES python libraries
 ; you must disable beaker.session.secret to use this
 #beaker.session.encrypt_key = key_for_encryption
 #beaker.session.validate_key = validation_key
 
 ; Sets session as invalid (also logging out user) if it haven not been
 ; accessed for given amount of time in seconds
 beaker.session.timeout = 2592000
 beaker.session.httponly = true
 
 ; Path to use for the cookie. Set to prefix if you use prefix middleware
 #beaker.session.cookie_path = /custom_prefix
 
 ; Set https secure cookie
 beaker.session.secure = false
 
 ; default cookie expiration time in seconds, set to `true` to set expire
 ; at browser close
 #beaker.session.cookie_expires = 3600
 
 ; #############################
 ; SEARCH INDEXING CONFIGURATION
 ; #############################
 
 ; Full text search indexer is available in rhodecode-tools under
 ; `rhodecode-tools index` command
 
 ; WHOOSH Backend, doesn't require additional services to run
 ; it works good with few dozen repos
 search.module = rhodecode.lib.index.whoosh
 search.location = %(here)s/data/index
 
 ; ####################
 ; CHANNELSTREAM CONFIG
 ; ####################
 
 ; channelstream enables persistent connections and live notification
 ; in the system. It's also used by the chat system
 
 channelstream.enabled = true
 
 ; server address for channelstream server on the backend
 channelstream.server = channelstream:9800
 
 ; location of the channelstream server from outside world
 ; use ws:// for http or wss:// for https. This address needs to be handled
 ; by external HTTP server such as Nginx or Apache
 ; see Nginx/Apache configuration examples in our docs
 channelstream.ws_url = ws://rhodecode.yourserver.com/_channelstream
 channelstream.secret = ENV_GENERATED
 channelstream.history.location = /var/opt/rhodecode_data/channelstream_history
 
 ; Internal application path that Javascript uses to connect into.
 ; If you use proxy-prefix the prefix should be added before /_channelstream
 channelstream.proxy_path = /_channelstream
 
 
 ; ##############################
 ; MAIN RHODECODE DATABASE CONFIG
 ; ##############################
 
 #sqlalchemy.db1.url = sqlite:///%(here)s/rhodecode.db?timeout=30
 #sqlalchemy.db1.url = postgresql://postgres:qweqwe@localhost/rhodecode
 #sqlalchemy.db1.url = mysql://root:qweqwe@localhost/rhodecode?charset=utf8
 ; pymysql is an alternative driver for MySQL, use in case of problems with default one
 #sqlalchemy.db1.url = mysql+pymysql://root:qweqwe@localhost/rhodecode
 
 sqlalchemy.db1.url = postgresql://postgres:qweqwe@localhost/rhodecode
 
 ; see sqlalchemy docs for other advanced settings
 ; print the sql statements to output
 sqlalchemy.db1.echo = false
 
 ; recycle the connections after this amount of seconds
 sqlalchemy.db1.pool_recycle = 3600
 
 ; the number of connections to keep open inside the connection pool.
 ; 0 indicates no limit
 ; the general calculus with gevent is:
 ; if your system allows 500 concurrent greenlets (max_connections) that all do database access,
 ; then increase pool size + max overflow so that they add up to 500.
 #sqlalchemy.db1.pool_size = 5
 
 ; The number of connections to allow in connection pool "overflow", that is
 ; connections that can be opened above and beyond the pool_size setting,
 ; which defaults to five.
 #sqlalchemy.db1.max_overflow = 10
 
 ; Connection check ping, used to detect broken database connections
 ; could be enabled to better handle cases if MySQL has gone away errors
 #sqlalchemy.db1.ping_connection = true
 
 ; ##########
 ; VCS CONFIG
 ; ##########
 vcs.server.enable = true
 vcs.server = vcsserver:10010
 
 ; Web server connectivity protocol, responsible for web based VCS operations
 ; Available protocols are:
 ; `http` - use http-rpc backend (default)
 vcs.server.protocol = http
 
 ; Push/Pull operations protocol, available options are:
 ; `http` - use http-rpc backend (default)
 vcs.scm_app_implementation = http
 
 ; Push/Pull operations hooks protocol, available options are:
 ; `http` - use http-rpc backend (default)
 ; `celery` - use celery based hooks
 vcs.hooks.protocol = http
 
 ; Host on which this instance is listening for hooks. vcsserver will call this host to pull/push hooks so it should be
 ; accessible via network.
 ; Use vcs.hooks.host = "*" to bind to current hostname (for Docker)
 vcs.hooks.host = *
 
 ; Start VCSServer with this instance as a subprocess, useful for development
 vcs.start_server = false
 
 ; List of enabled VCS backends, available options are:
 ; `hg`  - mercurial
 ; `git` - git
 ; `svn` - subversion
 vcs.backends = hg, git, svn
 
 ; Wait this number of seconds before killing connection to the vcsserver
 vcs.connection_timeout = 3600
 
 ; Cache flag to cache vcsserver remote calls locally
 ; It uses cache_region `cache_repo`
 vcs.methods.cache = true
 
 ; ####################################################
 ; Subversion proxy support (mod_dav_svn)
 ; Maps RhodeCode repo groups into SVN paths for Apache
 ; ####################################################
 
 ; Compatibility version when creating SVN repositories. Defaults to newest version when commented out.
 ; Set a numeric version for your current SVN e.g 1.8, or 1.12
 ; Legacy available options are: pre-1.4-compatible, pre-1.5-compatible, pre-1.6-compatible, pre-1.8-compatible, pre-1.9-compatible
 #vcs.svn.compatible_version = 1.8
 
 ; Enable SVN proxy of requests over HTTP
 vcs.svn.proxy.enabled = true
 
 ; host to connect to running SVN subsystem
 vcs.svn.proxy.host = http://svn:8090
 
 ; Enable or disable the config file generation.
 svn.proxy.generate_config = true
 
 ; Generate config file with `SVNListParentPath` set to `On`.
 svn.proxy.list_parent_path = true
 
 ; Set location and file name of generated config file.
 svn.proxy.config_file_path = /etc/rhodecode/conf/svn/mod_dav_svn.conf
 
 ; alternative mod_dav config template. This needs to be a valid mako template
 ; Example template can be found in the source code:
 ; rhodecode/apps/svn_support/templates/mod-dav-svn.conf.mako
 #svn.proxy.config_template = ~/.rccontrol/enterprise-1/custom_svn_conf.mako
 
 ; Used as a prefix to the `Location` block in the generated config file.
 ; In most cases it should be set to `/`.
 svn.proxy.location_root = /
 
 ; Command to reload the mod dav svn configuration on change.
 ; Example: `/etc/init.d/apache2 reload` or /home/USER/apache_reload.sh
 ; Make sure user who runs RhodeCode process is allowed to reload Apache
 #svn.proxy.reload_cmd = /etc/init.d/apache2 reload
 
 ; If the timeout expires before the reload command finishes, the command will
 ; be killed. Setting it to zero means no timeout. Defaults to 10 seconds.
 #svn.proxy.reload_timeout = 10
 
 ; ####################
 ; SSH Support Settings
 ; ####################
 
 ; Defines if a custom authorized_keys file should be created and written on
 ; any change user ssh keys. Setting this to false also disables possibility
 ; of adding SSH keys by users from web interface. Super admins can still
 ; manage SSH Keys.
 ssh.generate_authorized_keyfile = true
 
 ; Options for ssh, default is `no-pty,no-port-forwarding,no-X11-forwarding,no-agent-forwarding`
 # ssh.authorized_keys_ssh_opts =
 
 ; Path to the authorized_keys file where the generate entries are placed.
 ; It is possible to have multiple key files specified in `sshd_config` e.g.
 ; AuthorizedKeysFile %h/.ssh/authorized_keys %h/.ssh/authorized_keys_rhodecode
 ssh.authorized_keys_file_path = /etc/rhodecode/conf/ssh/authorized_keys_rhodecode
 
 ; Command to execute the SSH wrapper. The binary is available in the
 ; RhodeCode installation directory.
 ; legacy: /usr/local/bin/rhodecode_bin/bin/rc-ssh-wrapper
 ; new rewrite: /usr/local/bin/rhodecode_bin/bin/rc-ssh-wrapper-v2
 ssh.wrapper_cmd = /usr/local/bin/rhodecode_bin/bin/rc-ssh-wrapper
 
 ; Allow shell when executing the ssh-wrapper command
 ssh.wrapper_cmd_allow_shell = false
 
 ; Enables logging, and detailed output send back to the client during SSH
 ; operations. Useful for debugging, shouldn't be used in production.
 ssh.enable_debug_logging = false
 
 ; Paths to binary executable, by default they are the names, but we can
 ; override them if we want to use a custom one
 ssh.executable.hg = /usr/local/bin/rhodecode_bin/vcs_bin/hg
 ssh.executable.git = /usr/local/bin/rhodecode_bin/vcs_bin/git
 ssh.executable.svn = /usr/local/bin/rhodecode_bin/vcs_bin/svnserve
 
 ; Enables SSH key generator web interface. Disabling this still allows users
 ; to add their own keys.
 ssh.enable_ui_key_generator = true
 
 ; Statsd client config, this is used to send metrics to statsd
 ; We recommend setting statsd_exported and scrape them using Prometheus
 #statsd.enabled = false
 #statsd.statsd_host = 0.0.0.0
 #statsd.statsd_port = 8125
 #statsd.statsd_prefix =
 #statsd.statsd_ipv6 = false
 
 ; configure logging automatically at server startup set to false
 ; to use the below custom logging config.
 ; RC_LOGGING_FORMATTER
 ; RC_LOGGING_LEVEL
 ; env variables can control the settings for logging in case of autoconfigure
 
 #logging.autoconfigure = true
 
 ; specify your own custom logging config file to configure logging
 #logging.logging_conf_file = /path/to/custom_logging.ini
 
 ; Dummy marker to add new entries after.
 ; Add any custom entries below. Please don't remove this marker.
 custom.conf = 1
 
 
 ; #####################
 ; LOGGING CONFIGURATION
 ; #####################
 
 [loggers]
 keys = root, sqlalchemy, beaker, celery, rhodecode, ssh_wrapper
 
 [handlers]
 keys = console, console_sql
 
 [formatters]
 keys = generic, json, color_formatter, color_formatter_sql
 
 ; #######
 ; LOGGERS
 ; #######
 [logger_root]
 level = NOTSET
 handlers = console
 
 [logger_sqlalchemy]
 level = INFO
 handlers = console_sql
 qualname = sqlalchemy.engine
 propagate = 0
 
 [logger_beaker]
 level = DEBUG
 handlers =
 qualname = beaker.container
 propagate = 1
 
 [logger_rhodecode]
 level = DEBUG
 handlers =
 qualname = rhodecode
 propagate = 1
 
 [logger_ssh_wrapper]
 level = DEBUG
 handlers =
 qualname = ssh_wrapper
 propagate = 1
 
 [logger_celery]
 level = DEBUG
 handlers =
 qualname = celery
 
 
 ; ########
 ; HANDLERS
 ; ########
 
 [handler_console]
 class = StreamHandler
 args = (sys.stderr, )
 level = INFO
 ; To enable JSON formatted logs replace 'generic/color_formatter' with 'json'
 ; This allows sending properly formatted logs to grafana loki or elasticsearch
 formatter = generic
 
 [handler_console_sql]
 ; "level = DEBUG" logs SQL queries and results.
 ; "level = INFO" logs SQL queries.
 ; "level = WARN" logs neither.  (Recommended for production systems.)
 class = StreamHandler
 args = (sys.stderr, )
 level = WARN
 ; To enable JSON formatted logs replace 'generic/color_formatter_sql' with 'json'
 ; This allows sending properly formatted logs to grafana loki or elasticsearch
 formatter = generic
 
 ; ##########
 ; FORMATTERS
 ; ##########
 
 [formatter_generic]
 class = rhodecode.lib.logging_formatter.ExceptionAwareFormatter
 format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s
 datefmt = %Y-%m-%d %H:%M:%S
 
 [formatter_color_formatter]
 class = rhodecode.lib.logging_formatter.ColorFormatter
 format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s
 datefmt = %Y-%m-%d %H:%M:%S
 
 [formatter_color_formatter_sql]
 class = rhodecode.lib.logging_formatter.ColorFormatterSql
 format = %(asctime)s.%(msecs)03d [%(process)d] %(levelname)-5.5s [%(name)s] %(message)s
 datefmt = %Y-%m-%d %H:%M:%S
 
 [formatter_json]
 format = %(timestamp)s %(levelname)s %(name)s %(message)s %(req_id)s
 class = rhodecode.lib._vendor.jsonlogger.JsonFormatter

requirements.txt

@@ -1,287 +1,286 b''
 # deps, generated via pipdeptree --exclude setuptools,wheel,pipdeptree,pip -f | tr '[:upper:]' '[:lower:]'
 
 alembic==1.13.1
   mako==1.2.4
     markupsafe==2.1.2
   sqlalchemy==1.4.52
     greenlet==3.0.3
   typing_extensions==4.9.0
 async-timeout==4.0.3
 babel==2.12.1
 beaker==1.12.1
 celery==5.3.6
   billiard==4.2.0
   click==8.1.3
   click-didyoumean==0.3.0
     click==8.1.3
   click-plugins==1.1.1
     click==8.1.3
   click-repl==0.2.0
     click==8.1.3
     prompt-toolkit==3.0.38
       wcwidth==0.2.6
     six==1.16.0
   kombu==5.3.5
     amqp==5.2.0
       vine==5.1.0
     vine==5.1.0
   python-dateutil==2.8.2
     six==1.16.0
   tzdata==2024.1
   vine==5.1.0
 channelstream==0.7.1
   gevent==24.2.1
     greenlet==3.0.3
     zope.event==5.0.0
     zope.interface==6.3.0
   itsdangerous==1.1.0
   marshmallow==2.18.0
   pyramid==2.0.2
     hupper==1.12
     plaster==1.1.2
     plaster-pastedeploy==1.0.1
       pastedeploy==3.1.0
       plaster==1.1.2
     translationstring==1.4
     venusian==3.0.0
     webob==1.8.7
     zope.deprecation==5.0.0
     zope.interface==6.3.0
   pyramid-apispec==0.3.3
     apispec==1.3.3
   pyramid-jinja2==2.10
     jinja2==3.1.2
       markupsafe==2.1.2
     markupsafe==2.1.2
     pyramid==2.0.2
       hupper==1.12
       plaster==1.1.2
       plaster-pastedeploy==1.0.1
         pastedeploy==3.1.0
         plaster==1.1.2
       translationstring==1.4
       venusian==3.0.0
       webob==1.8.7
       zope.deprecation==5.0.0
       zope.interface==6.3.0
     zope.deprecation==5.0.0
   python-dateutil==2.8.2
     six==1.16.0
   requests==2.28.2
     certifi==2022.12.7
     charset-normalizer==3.1.0
     idna==3.4
     urllib3==1.26.14
   ws4py==0.5.1
 deform==2.0.15
   chameleon==3.10.2
   colander==2.0
     iso8601==1.1.0
     translationstring==1.4
   iso8601==1.1.0
   peppercorn==0.6
   translationstring==1.4
   zope.deprecation==5.0.0
-diskcache==5.6.3
 docutils==0.19
 dogpile.cache==1.3.3
   decorator==5.1.1
   stevedore==5.1.0
     pbr==5.11.1
 formencode==2.1.0
   six==1.16.0
 gunicorn==21.2.0
   packaging==24.0
 gevent==24.2.1
   greenlet==3.0.3
   zope.event==5.0.0
   zope.interface==6.3.0
 ipython==8.14.0
   backcall==0.2.0
   decorator==5.1.1
   jedi==0.19.0
     parso==0.8.3
   matplotlib-inline==0.1.6
     traitlets==5.9.0
   pexpect==4.8.0
     ptyprocess==0.7.0
   pickleshare==0.7.5
   prompt-toolkit==3.0.38
     wcwidth==0.2.6
   pygments==2.15.1
   stack-data==0.6.2
     asttokens==2.2.1
       six==1.16.0
     executing==1.2.0
     pure-eval==0.2.2
   traitlets==5.9.0
 markdown==3.4.3
 msgpack==1.0.8
 mysqlclient==2.1.1
 nbconvert==7.7.3
   beautifulsoup4==4.12.3
     soupsieve==2.5
   bleach==6.1.0
     six==1.16.0
     webencodings==0.5.1
   defusedxml==0.7.1
   jinja2==3.1.2
     markupsafe==2.1.2
   jupyter_core==5.3.1
     platformdirs==3.10.0
     traitlets==5.9.0
   jupyterlab-pygments==0.2.2
   markupsafe==2.1.2
   mistune==2.0.5
   nbclient==0.8.0
     jupyter_client==8.3.0
       jupyter_core==5.3.1
         platformdirs==3.10.0
         traitlets==5.9.0
       python-dateutil==2.8.2
         six==1.16.0
       pyzmq==25.0.0
       tornado==6.2
       traitlets==5.9.0
     jupyter_core==5.3.1
       platformdirs==3.10.0
       traitlets==5.9.0
     nbformat==5.9.2
       fastjsonschema==2.18.0
       jsonschema==4.18.6
         attrs==22.2.0
         pyrsistent==0.19.3
       jupyter_core==5.3.1
         platformdirs==3.10.0
         traitlets==5.9.0
       traitlets==5.9.0
     traitlets==5.9.0
   nbformat==5.9.2
     fastjsonschema==2.18.0
     jsonschema==4.18.6
       attrs==22.2.0
       pyrsistent==0.19.3
     jupyter_core==5.3.1
       platformdirs==3.10.0
       traitlets==5.9.0
     traitlets==5.9.0
   pandocfilters==1.5.0
   pygments==2.15.1
   tinycss2==1.2.1
     webencodings==0.5.1
   traitlets==5.9.0
 orjson==3.10.3
 paste==3.10.1
 premailer==3.10.0
   cachetools==5.3.3
   cssselect==1.2.0
   cssutils==2.6.0
   lxml==4.9.3
   requests==2.28.2
     certifi==2022.12.7
     charset-normalizer==3.1.0
     idna==3.4
     urllib3==1.26.14
 psutil==5.9.8
 psycopg2==2.9.9
 py-bcrypt==0.4
 pycmarkgfm==1.2.0
   cffi==1.16.0
     pycparser==2.21
 pycryptodome==3.17
 pycurl==7.45.3
 pymysql==1.0.3
 pyotp==2.8.0
 pyparsing==3.1.1
 pyramid-debugtoolbar==4.12.1
   pygments==2.15.1
   pyramid==2.0.2
     hupper==1.12
     plaster==1.1.2
     plaster-pastedeploy==1.0.1
       pastedeploy==3.1.0
       plaster==1.1.2
     translationstring==1.4
     venusian==3.0.0
     webob==1.8.7
     zope.deprecation==5.0.0
     zope.interface==6.3.0
   pyramid-mako==1.1.0
     mako==1.2.4
       markupsafe==2.1.2
     pyramid==2.0.2
       hupper==1.12
       plaster==1.1.2
       plaster-pastedeploy==1.0.1
         pastedeploy==3.1.0
         plaster==1.1.2
       translationstring==1.4
       venusian==3.0.0
       webob==1.8.7
       zope.deprecation==5.0.0
       zope.interface==6.3.0
 pyramid-mailer==0.15.1
   pyramid==2.0.2
     hupper==1.12
     plaster==1.1.2
     plaster-pastedeploy==1.0.1
       pastedeploy==3.1.0
       plaster==1.1.2
     translationstring==1.4
     venusian==3.0.0
     webob==1.8.7
     zope.deprecation==5.0.0
     zope.interface==6.3.0
   repoze.sendmail==4.4.1
     transaction==3.1.0
       zope.interface==6.3.0
     zope.interface==6.3.0
   transaction==3.1.0
     zope.interface==6.3.0
 python-ldap==3.4.3
   pyasn1==0.4.8
   pyasn1-modules==0.2.8
     pyasn1==0.4.8
 python-memcached==1.59
   six==1.16.0
 python-pam==2.0.2
 python3-saml==1.15.0
   isodate==0.6.1
     six==1.16.0
   lxml==4.9.3
   xmlsec==1.3.13
     lxml==4.9.3
 pyyaml==6.0.1
 redis==5.0.4
   async-timeout==4.0.3
 regex==2022.10.31
 routes==2.5.1
   repoze.lru==0.7
   six==1.16.0
 simplejson==3.19.2
 sshpubkeys==3.3.1
   cryptography==40.0.2
     cffi==1.16.0
       pycparser==2.21
   ecdsa==0.18.0
     six==1.16.0
 sqlalchemy==1.4.52
   greenlet==3.0.3
   typing_extensions==4.9.0
 supervisor==4.2.5
 tzlocal==4.3
   pytz-deprecation-shim==0.1.0.post0
     tzdata==2024.1
 tempita==0.5.2
 unidecode==1.3.6
 urlobject==2.4.3
 waitress==3.0.0
 webhelpers2==2.1
   markupsafe==2.1.2
   six==1.16.0
 whoosh==2.7.4
 zope.cachedescriptors==5.0.0
 qrcode==7.4.2
 
 ## uncomment to add the debug libraries
 #-r requirements_debug.txt

rhodecode/apps/repository/views/repo_files.py

@@ -1,1708 +1,1715 b''
 # Copyright (C) 2011-2023 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 import itertools
 import logging
 import os
 import collections
 import urllib.request
 import urllib.parse
 import urllib.error
 import pathlib
+import time
+import random
 
 from pyramid.httpexceptions import HTTPNotFound, HTTPBadRequest, HTTPFound
 
 from pyramid.renderers import render
 from pyramid.response import Response
 
 import rhodecode
 from rhodecode.apps._base import RepoAppView
 
 
 from rhodecode.lib import diffs, helpers as h, rc_cache
 from rhodecode.lib import audit_logger
 from rhodecode.lib.hash_utils import sha1_safe
-from rhodecode.lib.rc_cache.archive_cache import get_archival_cache_store, get_archival_config, ReentrantLock
+from rhodecode.lib.rc_cache.archive_cache import (
+    get_archival_cache_store, get_archival_config, ArchiveCacheLock, archive_iterator)
 from rhodecode.lib.str_utils import safe_bytes, convert_special_chars
 from rhodecode.lib.view_utils import parse_path_ref
 from rhodecode.lib.exceptions import NonRelativePathError
 from rhodecode.lib.codeblocks import (
     filenode_as_lines_tokens, filenode_as_annotated_lines_tokens)
 from rhodecode.lib.utils2 import convert_line_endings, detect_mode
 from rhodecode.lib.type_utils import str2bool
 from rhodecode.lib.str_utils import safe_str, safe_int
 from rhodecode.lib.auth import (
     LoginRequired, HasRepoPermissionAnyDecorator, CSRFRequired)
 from rhodecode.lib.vcs import path as vcspath
 from rhodecode.lib.vcs.backends.base import EmptyCommit
 from rhodecode.lib.vcs.conf import settings
 from rhodecode.lib.vcs.nodes import FileNode
 from rhodecode.lib.vcs.exceptions import (
     RepositoryError, CommitDoesNotExistError, EmptyRepositoryError,
     ImproperArchiveTypeError, VCSError, NodeAlreadyExistsError,
     NodeDoesNotExistError, CommitError, NodeError)
 
 from rhodecode.model.scm import ScmModel
 from rhodecode.model.db import Repository
 
 log = logging.getLogger(__name__)
 
 
 def get_archive_name(db_repo_id, db_repo_name, commit_sha, ext, subrepos=False, path_sha='', with_hash=True):
     # original backward compat name of archive
     clean_name = safe_str(convert_special_chars(db_repo_name).replace('/', '_'))
 
     # e.g vcsserver-id-abcd-sub-1-abcfdef-archive-all.zip
     # vcsserver-id-abcd-sub-0-abcfdef-COMMIT_SHA-PATH_SHA.zip
     id_sha = sha1_safe(str(db_repo_id))[:4]
     sub_repo = 'sub-1' if subrepos else 'sub-0'
     commit = commit_sha if with_hash else 'archive'
     path_marker = (path_sha if with_hash else '') or 'all'
     archive_name = f'{clean_name}-id-{id_sha}-{sub_repo}-{commit}-{path_marker}{ext}'
 
     return archive_name
 
 
 def get_path_sha(at_path):
     return safe_str(sha1_safe(at_path)[:8])
 
 
 def _get_archive_spec(fname):
     log.debug('Detecting archive spec for: `%s`', fname)
 
     fileformat = None
     ext = None
     content_type = None
     for a_type, content_type, extension in settings.ARCHIVE_SPECS:
 
         if fname.endswith(extension):
             fileformat = a_type
             log.debug('archive is of type: %s', fileformat)
             ext = extension
             break
 
     if not fileformat:
         raise ValueError()
 
     # left over part of whole fname is the commit
     commit_id = fname[:-len(ext)]
 
     return commit_id, ext, fileformat, content_type
 
 
 class RepoFilesView(RepoAppView):
 
     @staticmethod
     def adjust_file_path_for_svn(f_path, repo):
         """
         Computes the relative path of `f_path`.
 
         This is mainly based on prefix matching of the recognized tags and
         branches in the underlying repository.
         """
         tags_and_branches = itertools.chain(
             repo.branches.keys(),
             repo.tags.keys())
         tags_and_branches = sorted(tags_and_branches, key=len, reverse=True)
 
         for name in tags_and_branches:
             if f_path.startswith(f'{name}/'):
                 f_path = vcspath.relpath(f_path, name)
                 break
         return f_path
 
     def load_default_context(self):
         c = self._get_local_tmpl_context(include_app_defaults=True)
         c.rhodecode_repo = self.rhodecode_vcs_repo
         c.enable_downloads = self.db_repo.enable_downloads
         return c
 
     def _ensure_not_locked(self, commit_id='tip'):
         _ = self.request.translate
 
         repo = self.db_repo
         if repo.enable_locking and repo.locked[0]:
             h.flash(_('This repository has been locked by %s on %s')
                     % (h.person_by_id(repo.locked[0]),
                     h.format_date(h.time_to_datetime(repo.locked[1]))),
                     'warning')
             files_url = h.route_path(
                 'repo_files:default_path',
                 repo_name=self.db_repo_name, commit_id=commit_id)
             raise HTTPFound(files_url)
 
     def forbid_non_head(self, is_head, f_path, commit_id='tip', json_mode=False):
         _ = self.request.translate
 
         if not is_head:
             message = _('Cannot modify file. '
                         'Given commit `{}` is not head of a branch.').format(commit_id)
             h.flash(message, category='warning')
 
             if json_mode:
                 return message
 
             files_url = h.route_path(
                 'repo_files', repo_name=self.db_repo_name, commit_id=commit_id,
                 f_path=f_path)
             raise HTTPFound(files_url)
 
     def check_branch_permission(self, branch_name, commit_id='tip', json_mode=False):
         _ = self.request.translate
 
         rule, branch_perm = self._rhodecode_user.get_rule_and_branch_permission(
             self.db_repo_name, branch_name)
         if branch_perm and branch_perm not in ['branch.push', 'branch.push_force']:
             message = _('Branch `{}` changes forbidden by rule {}.').format(
                 h.escape(branch_name), h.escape(rule))
             h.flash(message, 'warning')
 
             if json_mode:
                 return message
 
             files_url = h.route_path(
                 'repo_files:default_path', repo_name=self.db_repo_name, commit_id=commit_id)
 
             raise HTTPFound(files_url)
 
     def _get_commit_and_path(self):
         default_commit_id = self.db_repo.landing_ref_name
         default_f_path = '/'
 
         commit_id = self.request.matchdict.get(
             'commit_id', default_commit_id)
         f_path = self._get_f_path(self.request.matchdict, default_f_path)
         return commit_id, f_path
 
     def _get_default_encoding(self, c):
         enc_list = getattr(c, 'default_encodings', [])
         return enc_list[0] if enc_list else 'UTF-8'
 
     def _get_commit_or_redirect(self, commit_id, redirect_after=True):
         """
         This is a safe way to get commit. If an error occurs it redirects to
         tip with proper message
 
         :param commit_id: id of commit to fetch
         :param redirect_after: toggle redirection
         """
         _ = self.request.translate
 
         try:
             return self.rhodecode_vcs_repo.get_commit(commit_id)
         except EmptyRepositoryError:
             if not redirect_after:
                 return None
 
             add_new = upload_new = ""
             if h.HasRepoPermissionAny(
                     'repository.write', 'repository.admin')(self.db_repo_name):
                 _url = h.route_path(
                     'repo_files_add_file',
                     repo_name=self.db_repo_name, commit_id=0, f_path='')
                 add_new = h.link_to(
                     _('add a new file'), _url, class_="alert-link")
 
                 _url_upld = h.route_path(
                     'repo_files_upload_file',
                     repo_name=self.db_repo_name, commit_id=0, f_path='')
                 upload_new = h.link_to(
                     _('upload a new file'), _url_upld, class_="alert-link")
 
             h.flash(h.literal(
                 _('There are no files yet. Click here to %s or %s.') % (add_new, upload_new)), category='warning')
             raise HTTPFound(
                 h.route_path('repo_summary', repo_name=self.db_repo_name))
 
         except (CommitDoesNotExistError, LookupError) as e:
             msg = _('No such commit exists for this repository. Commit: {}').format(commit_id)
             h.flash(msg, category='error')
             raise HTTPNotFound()
         except RepositoryError as e:
             h.flash(h.escape(safe_str(e)), category='error')
             raise HTTPNotFound()
 
     def _get_filenode_or_redirect(self, commit_obj, path, pre_load=None):
         """
         Returns file_node, if error occurs or given path is directory,
         it'll redirect to top level path
         """
         _ = self.request.translate
 
         try:
             file_node = commit_obj.get_node(path, pre_load=pre_load)
             if file_node.is_dir():
                 raise RepositoryError('The given path is a directory')
         except CommitDoesNotExistError:
             log.exception('No such commit exists for this repository')
             h.flash(_('No such commit exists for this repository'), category='error')
             raise HTTPNotFound()
         except RepositoryError as e:
             log.warning('Repository error while fetching filenode `%s`. Err:%s', path, e)
             h.flash(h.escape(safe_str(e)), category='error')
             raise HTTPNotFound()
 
         return file_node
 
     def _is_valid_head(self, commit_id, repo, landing_ref):
         branch_name = sha_commit_id = ''
         is_head = False
         log.debug('Checking if commit_id `%s` is a head for %s.', commit_id, repo)
 
         for _branch_name, branch_commit_id in repo.branches.items():
             # simple case we pass in branch name, it's a HEAD
             if commit_id == _branch_name:
                 is_head = True
                 branch_name = _branch_name
                 sha_commit_id = branch_commit_id
                 break
             # case when we pass in full sha commit_id, which is a head
             elif commit_id == branch_commit_id:
                 is_head = True
                 branch_name = _branch_name
                 sha_commit_id = branch_commit_id
                 break
 
         if h.is_svn(repo) and not repo.is_empty():
             # Note: Subversion only has one head.
             if commit_id == repo.get_commit(commit_idx=-1).raw_id:
                 is_head = True
                 return branch_name, sha_commit_id, is_head
 
         # checked branches, means we only need to try to get the branch/commit_sha
         if repo.is_empty():
             is_head = True
             branch_name = landing_ref
             sha_commit_id = EmptyCommit().raw_id
         else:
             commit = repo.get_commit(commit_id=commit_id)
             if commit:
                 branch_name = commit.branch
                 sha_commit_id = commit.raw_id
 
         return branch_name, sha_commit_id, is_head
 
     def _get_tree_at_commit(self, c, commit_id, f_path, full_load=False, at_rev=None):
 
         repo_id = self.db_repo.repo_id
         force_recache = self.get_recache_flag()
 
         cache_seconds = safe_int(
             rhodecode.CONFIG.get('rc_cache.cache_repo.expiration_time'))
         cache_on = not force_recache and cache_seconds > 0
         log.debug(
             'Computing FILE TREE for repo_id %s commit_id `%s` and path `%s`'
             'with caching: %s[TTL: %ss]' % (
                 repo_id, commit_id, f_path, cache_on, cache_seconds or 0))
 
         cache_namespace_uid = f'repo.{rc_cache.FILE_TREE_CACHE_VER}.{repo_id}'
         region = rc_cache.get_or_create_region('cache_repo', cache_namespace_uid)
 
         @region.conditional_cache_on_arguments(namespace=cache_namespace_uid, condition=cache_on)
         def compute_file_tree(_name_hash, _repo_id, _commit_id, _f_path, _full_load, _at_rev):
             log.debug('Generating cached file tree at for repo_id: %s, %s, %s',
                       _repo_id, _commit_id, _f_path)
 
             c.full_load = _full_load
             return render(
                 'rhodecode:templates/files/files_browser_tree.mako',
                 self._get_template_context(c), self.request, _at_rev)
 
         return compute_file_tree(
             self.db_repo.repo_name_hash, self.db_repo.repo_id, commit_id, f_path, full_load, at_rev)
 
     def create_pure_path(self, *parts):
         # Split paths and sanitize them, removing any ../ etc
         sanitized_path = [
             x for x in pathlib.PurePath(*parts).parts
             if x not in ['.', '..']]
 
         pure_path = pathlib.PurePath(*sanitized_path)
         return pure_path
 
     def _is_lf_enabled(self, target_repo):
         lf_enabled = False
 
         lf_key_for_vcs_map = {
             'hg': 'extensions_largefiles',
             'git': 'vcs_git_lfs_enabled'
         }
 
         lf_key_for_vcs = lf_key_for_vcs_map.get(target_repo.repo_type)
 
         if lf_key_for_vcs:
             lf_enabled = self._get_repo_setting(target_repo, lf_key_for_vcs)
 
         return lf_enabled
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_archivefile(self):
         # archive cache config
         from rhodecode import CONFIG
         _ = self.request.translate
         self.load_default_context()
         default_at_path = '/'
         fname = self.request.matchdict['fname']
         subrepos = self.request.GET.get('subrepos') == 'true'
         with_hash = str2bool(self.request.GET.get('with_hash', '1'))
         at_path = self.request.GET.get('at_path') or default_at_path
 
         if not self.db_repo.enable_downloads:
             return Response(_('Downloads disabled'))
 
         try:
             commit_id, ext, fileformat, content_type = \
                 _get_archive_spec(fname)
         except ValueError:
             return Response(_('Unknown archive type for: `{}`').format(
                 h.escape(fname)))
 
         try:
             commit = self.rhodecode_vcs_repo.get_commit(commit_id)
         except CommitDoesNotExistError:
             return Response(_('Unknown commit_id {}').format(
                 h.escape(commit_id)))
         except EmptyRepositoryError:
             return Response(_('Empty repository'))
 
         # we used a ref, or a shorter version, lets redirect client ot use explicit hash
         if commit_id != commit.raw_id:
             fname=f'{commit.raw_id}{ext}'
             raise HTTPFound(self.request.current_route_path(fname=fname))
 
         try:
             at_path = commit.get_node(at_path).path or default_at_path
         except Exception:
             return Response(_('No node at path {} for this repository').format(h.escape(at_path)))
 
         path_sha = get_path_sha(at_path)
 
         # used for cache etc, consistent unique archive name
         archive_name_key = get_archive_name(
             self.db_repo.repo_id, self.db_repo_name, commit_sha=commit.short_id, ext=ext, subrepos=subrepos,
             path_sha=path_sha, with_hash=True)
 
         if not with_hash:
             path_sha = ''
 
         # what end client gets served
         response_archive_name = get_archive_name(
             self.db_repo.repo_id, self.db_repo_name, commit_sha=commit.short_id, ext=ext, subrepos=subrepos,
             path_sha=path_sha, with_hash=with_hash)
 
         # remove extension from our archive directory name
         archive_dir_name = response_archive_name[:-len(ext)]
 
         archive_cache_disable = self.request.GET.get('no_cache')
 
         d_cache = get_archival_cache_store(config=CONFIG)
 
         # NOTE: we get the config to pass to a call to lazy-init the SAME type of cache on vcsserver
         d_cache_conf = get_archival_config(config=CONFIG)
 
+        # This is also a cache key, and lock key
         reentrant_lock_key = archive_name_key + '.lock'
-        with ReentrantLock(d_cache, reentrant_lock_key):
-            # This is also a cache key
-            use_cached_archive = False
-            if not archive_cache_disable and archive_name_key in d_cache:
-                reader, tag = d_cache.get(archive_name_key, read=True, tag=True, retry=True)
-                use_cached_archive = True
-                log.debug('Found cached archive as key=%s tag=%s, serving archive from cache reader=%s',
-                          archive_name_key, tag, reader.name)
-            else:
-                reader = None
-                log.debug('Archive with key=%s is not yet cached, creating one now...', archive_name_key)
+
+        use_cached_archive = False
+        if not archive_cache_disable and archive_name_key in d_cache:
+            reader, metadata = d_cache.fetch(archive_name_key)
 
-            # generate new archive, as previous was not found in the cache
-            if not reader:
-
-                try:
-                    commit.archive_repo(archive_name_key, archive_dir_name=archive_dir_name,
-                                        kind=fileformat, subrepos=subrepos,
-                                        archive_at_path=at_path, cache_config=d_cache_conf)
-                except ImproperArchiveTypeError:
-                    return _('Unknown archive type')
-
-            reader, tag = d_cache.get(archive_name_key, read=True, tag=True, retry=True)
+            use_cached_archive = True
+            log.debug('Found cached archive as key=%s tag=%s, serving archive from cache reader=%s',
+                      archive_name_key, metadata, reader.name)
+        else:
+            reader = None
+            log.debug('Archive with key=%s is not yet cached, creating one now...', archive_name_key)
 
         if not reader:
-            raise ValueError('archive cache reader is empty, failed to fetch file from distributed archive cache')
+            # generate new archive, as previous was not found in the cache
+            try:
+                with d_cache.get_lock(reentrant_lock_key):
+                    try:
+                        commit.archive_repo(archive_name_key, archive_dir_name=archive_dir_name,
+                                            kind=fileformat, subrepos=subrepos,
+                                            archive_at_path=at_path, cache_config=d_cache_conf)
+                    except ImproperArchiveTypeError:
+                        return _('Unknown archive type')
+            except ArchiveCacheLock:
+                retry_after = round(random.uniform(0.3, 3.0), 1)
+                time.sleep(retry_after)
 
-        def archive_iterator(_reader, block_size: int = 4096*512):
-            # 4096 * 64 = 64KB
-            while 1:
-                data = _reader.read(block_size)
-                if not data:
-                    break
-                yield data
+                location = self.request.url
+                response = Response(
+                    f"archive {archive_name_key} generation in progress, Retry-After={retry_after}, Location={location}"
+                )
+                response.headers["Retry-After"] = str(retry_after)
+                response.status_code = 307 # temporary redirect
+
+                response.location = location
+                return response
+
+            reader, metadata = d_cache.fetch(archive_name_key)
 
         response = Response(app_iter=archive_iterator(reader))
         response.content_disposition = f'attachment; filename={response_archive_name}'
         response.content_type = str(content_type)
 
         try:
             return response
         finally:
             # store download action
             audit_logger.store_web(
                 'repo.archive.download', action_data={
                     'user_agent': self.request.user_agent,
                     'archive_name': archive_name_key,
                     'archive_spec': fname,
                     'archive_cached': use_cached_archive},
                 user=self._rhodecode_user,
                 repo=self.db_repo,
                 commit=True
             )
 
     def _get_file_node(self, commit_id, f_path):
         if commit_id not in ['', None, 'None', '0' * 12, '0' * 40]:
             commit = self.rhodecode_vcs_repo.get_commit(commit_id=commit_id)
             try:
                 node = commit.get_node(f_path)
                 if node.is_dir():
                     raise NodeError(f'{node} path is a {type(node)} not a file')
             except NodeDoesNotExistError:
                 commit = EmptyCommit(
                     commit_id=commit_id,
                     idx=commit.idx,
                     repo=commit.repository,
                     alias=commit.repository.alias,
                     message=commit.message,
                     author=commit.author,
                     date=commit.date)
                 node = FileNode(safe_bytes(f_path), b'', commit=commit)
         else:
             commit = EmptyCommit(
                 repo=self.rhodecode_vcs_repo,
                 alias=self.rhodecode_vcs_repo.alias)
             node = FileNode(safe_bytes(f_path), b'', commit=commit)
         return node
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_files_diff(self):
         c = self.load_default_context()
         f_path = self._get_f_path(self.request.matchdict)
         diff1 = self.request.GET.get('diff1', '')
         diff2 = self.request.GET.get('diff2', '')
 
         path1, diff1 = parse_path_ref(diff1, default_path=f_path)
 
         ignore_whitespace = str2bool(self.request.GET.get('ignorews'))
         line_context = self.request.GET.get('context', 3)
 
         if not any((diff1, diff2)):
             h.flash(
                 'Need query parameter "diff1" or "diff2" to generate a diff.',
                 category='error')
             raise HTTPBadRequest()
 
         c.action = self.request.GET.get('diff')
         if c.action not in ['download', 'raw']:
             compare_url = h.route_path(
                 'repo_compare',
                 repo_name=self.db_repo_name,
                 source_ref_type='rev',
                 source_ref=diff1,
                 target_repo=self.db_repo_name,
                 target_ref_type='rev',
                 target_ref=diff2,
                 _query=dict(f_path=f_path))
             # redirect to new view if we render diff
             raise HTTPFound(compare_url)
 
         try:
             node1 = self._get_file_node(diff1, path1)
             node2 = self._get_file_node(diff2, f_path)
         except (RepositoryError, NodeError):
             log.exception("Exception while trying to get node from repository")
             raise HTTPFound(
                 h.route_path('repo_files', repo_name=self.db_repo_name,
                              commit_id='tip', f_path=f_path))
 
         if all(isinstance(node.commit, EmptyCommit)
                for node in (node1, node2)):
             raise HTTPNotFound()
 
         c.commit_1 = node1.commit
         c.commit_2 = node2.commit
 
         if c.action == 'download':
             _diff = diffs.get_gitdiff(node1, node2,
                                       ignore_whitespace=ignore_whitespace,
                                       context=line_context)
             # NOTE: this was using diff_format='gitdiff'
             diff = diffs.DiffProcessor(_diff, diff_format='newdiff')
 
             response = Response(self.path_filter.get_raw_patch(diff))
             response.content_type = 'text/plain'
             response.content_disposition = (
                 f'attachment; filename={f_path}_{diff1}_vs_{diff2}.diff'
             )
             charset = self._get_default_encoding(c)
             if charset:
                 response.charset = charset
             return response
 
         elif c.action == 'raw':
             _diff = diffs.get_gitdiff(node1, node2,
                                       ignore_whitespace=ignore_whitespace,
                                       context=line_context)
             # NOTE: this was using diff_format='gitdiff'
             diff = diffs.DiffProcessor(_diff, diff_format='newdiff')
 
             response = Response(self.path_filter.get_raw_patch(diff))
             response.content_type = 'text/plain'
             charset = self._get_default_encoding(c)
             if charset:
                 response.charset = charset
             return response
 
         # in case we ever end up here
         raise HTTPNotFound()
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_files_diff_2way_redirect(self):
         """
         Kept only to make OLD links work
         """
         f_path = self._get_f_path_unchecked(self.request.matchdict)
         diff1 = self.request.GET.get('diff1', '')
         diff2 = self.request.GET.get('diff2', '')
 
         if not any((diff1, diff2)):
             h.flash(
                 'Need query parameter "diff1" or "diff2" to generate a diff.',
                 category='error')
             raise HTTPBadRequest()
 
         compare_url = h.route_path(
             'repo_compare',
             repo_name=self.db_repo_name,
             source_ref_type='rev',
             source_ref=diff1,
             target_ref_type='rev',
             target_ref=diff2,
             _query=dict(f_path=f_path, diffmode='sideside',
                         target_repo=self.db_repo_name,))
         raise HTTPFound(compare_url)
 
     @LoginRequired()
     def repo_files_default_commit_redirect(self):
         """
         Special page that redirects to the landing page of files based on the default
         commit for repository
         """
         c = self.load_default_context()
         ref_name = c.rhodecode_db_repo.landing_ref_name
         landing_url = h.repo_files_by_ref_url(
             c.rhodecode_db_repo.repo_name,
             c.rhodecode_db_repo.repo_type,
             f_path='',
             ref_name=ref_name,
             commit_id='tip',
             query=dict(at=ref_name)
         )
 
         raise HTTPFound(landing_url)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_files(self):
         c = self.load_default_context()
 
         view_name = getattr(self.request.matched_route, 'name', None)
 
         c.annotate = view_name == 'repo_files:annotated'
         # default is false, but .rst/.md files later are auto rendered, we can
         # overwrite auto rendering by setting this GET flag
         c.renderer = view_name == 'repo_files:rendered' or not self.request.GET.get('no-render', False)
 
         commit_id, f_path = self._get_commit_and_path()
 
         c.commit = self._get_commit_or_redirect(commit_id)
         c.branch = self.request.GET.get('branch', None)
         c.f_path = f_path
         at_rev = self.request.GET.get('at')
 
         # files or dirs
         try:
             c.file = c.commit.get_node(f_path, pre_load=['is_binary', 'size', 'data'])
 
             c.file_author = True
             c.file_tree = ''
 
             # prev link
             try:
                 prev_commit = c.commit.prev(c.branch)
                 c.prev_commit = prev_commit
                 c.url_prev = h.route_path(
                     'repo_files', repo_name=self.db_repo_name,
                     commit_id=prev_commit.raw_id, f_path=f_path)
                 if c.branch:
                     c.url_prev += '?branch=%s' % c.branch
             except (CommitDoesNotExistError, VCSError):
                 c.url_prev = '#'
                 c.prev_commit = EmptyCommit()
 
             # next link
             try:
                 next_commit = c.commit.next(c.branch)
                 c.next_commit = next_commit
                 c.url_next = h.route_path(
                     'repo_files', repo_name=self.db_repo_name,
                     commit_id=next_commit.raw_id, f_path=f_path)
                 if c.branch:
                     c.url_next += '?branch=%s' % c.branch
             except (CommitDoesNotExistError, VCSError):
                 c.url_next = '#'
                 c.next_commit = EmptyCommit()
 
             # load file content
             if c.file.is_file():
 
                 c.lf_node = {}
 
                 has_lf_enabled = self._is_lf_enabled(self.db_repo)
                 if has_lf_enabled:
                     c.lf_node = c.file.get_largefile_node()
 
                 c.file_source_page = 'true'
                 c.file_last_commit = c.file.last_commit
 
                 c.file_size_too_big = c.file.size > c.visual.cut_off_limit_file
 
                 if not (c.file_size_too_big or c.file.is_binary):
                     if c.annotate:  # annotation has precedence over renderer
                         c.annotated_lines = filenode_as_annotated_lines_tokens(
                             c.file
                         )
                     else:
                         c.renderer = (
                             c.renderer and h.renderer_from_filename(c.file.path)
                         )
                         if not c.renderer:
                             c.lines = filenode_as_lines_tokens(c.file)
 
                 _branch_name, _sha_commit_id, is_head = \
                     self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                         landing_ref=self.db_repo.landing_ref_name)
                 c.on_branch_head = is_head
 
                 branch = c.commit.branch if (
                     c.commit.branch and '/' not in c.commit.branch) else None
                 c.branch_or_raw_id = branch or c.commit.raw_id
                 c.branch_name = c.commit.branch or h.short_id(c.commit.raw_id)
 
                 author = c.file_last_commit.author
                 c.authors = [[
                     h.email(author),
                     h.person(author, 'username_or_name_or_email'),
                     1
                 ]]
 
             else:  # load tree content at path
                 c.file_source_page = 'false'
                 c.authors = []
                 # this loads a simple tree without metadata to speed things up
                 # later via ajax we call repo_nodetree_full and fetch whole
                 c.file_tree = self._get_tree_at_commit(c, c.commit.raw_id, f_path, at_rev=at_rev)
 
                 c.readme_data, c.readme_file = \
                     self._get_readme_data(self.db_repo, c.visual.default_renderer,
                                           c.commit.raw_id, f_path)
 
         except RepositoryError as e:
             h.flash(h.escape(safe_str(e)), category='error')
             raise HTTPNotFound()
 
         if self.request.environ.get('HTTP_X_PJAX'):
             html = render('rhodecode:templates/files/files_pjax.mako',
                           self._get_template_context(c), self.request)
         else:
             html = render('rhodecode:templates/files/files.mako',
                           self._get_template_context(c), self.request)
         return Response(html)
 
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_files_annotated_previous(self):
         self.load_default_context()
 
         commit_id, f_path = self._get_commit_and_path()
         commit = self._get_commit_or_redirect(commit_id)
         prev_commit_id = commit.raw_id
         line_anchor = self.request.GET.get('line_anchor')
         is_file = False
         try:
             _file = commit.get_node(f_path)
             is_file = _file.is_file()
         except (NodeDoesNotExistError, CommitDoesNotExistError, VCSError):
             pass
 
         if is_file:
             history = commit.get_path_history(f_path)
             prev_commit_id = history[1].raw_id \
                 if len(history) > 1 else prev_commit_id
         prev_url = h.route_path(
             'repo_files:annotated', repo_name=self.db_repo_name,
             commit_id=prev_commit_id, f_path=f_path,
             _anchor=f'L{line_anchor}')
 
         raise HTTPFound(prev_url)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_nodetree_full(self):
         """
         Returns rendered html of file tree that contains commit date,
         author, commit_id for the specified combination of
         repo, commit_id and file path
         """
         c = self.load_default_context()
 
         commit_id, f_path = self._get_commit_and_path()
         commit = self._get_commit_or_redirect(commit_id)
         try:
             dir_node = commit.get_node(f_path)
         except RepositoryError as e:
             return Response(f'error: {h.escape(safe_str(e))}')
 
         if dir_node.is_file():
             return Response('')
 
         c.file = dir_node
         c.commit = commit
         at_rev = self.request.GET.get('at')
 
         html = self._get_tree_at_commit(
             c, commit.raw_id, dir_node.path, full_load=True, at_rev=at_rev)
 
         return Response(html)
 
     def _get_attachement_headers(self, f_path):
         f_name = safe_str(f_path.split(Repository.NAME_SEP)[-1])
         safe_path = f_name.replace('"', '\\"')
         encoded_path = urllib.parse.quote(f_name)
 
         headers = "attachment; " \
                   "filename=\"{}\"; " \
                   "filename*=UTF-8\'\'{}".format(safe_path, encoded_path)
 
         return safe_bytes(headers).decode('latin-1', errors='replace')
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_file_raw(self):
         """
         Action for show as raw, some mimetypes are "rendered",
         those include images, icons.
         """
         c = self.load_default_context()
 
         commit_id, f_path = self._get_commit_and_path()
         commit = self._get_commit_or_redirect(commit_id)
         file_node = self._get_filenode_or_redirect(commit, f_path)
 
         raw_mimetype_mapping = {
             # map original mimetype to a mimetype used for "show as raw"
             # you can also provide a content-disposition to override the
             # default "attachment" disposition.
             # orig_type: (new_type, new_dispo)
 
             # show images inline:
             # Do not re-add SVG: it is unsafe and permits XSS attacks. One can
             # for example render an SVG with javascript inside or even render
             # HTML.
             'image/x-icon': ('image/x-icon', 'inline'),
             'image/png': ('image/png', 'inline'),
             'image/gif': ('image/gif', 'inline'),
             'image/jpeg': ('image/jpeg', 'inline'),
             'application/pdf': ('application/pdf', 'inline'),
         }
 
         mimetype = file_node.mimetype
         try:
             mimetype, disposition = raw_mimetype_mapping[mimetype]
         except KeyError:
             # we don't know anything special about this, handle it safely
             if file_node.is_binary:
                 # do same as download raw for binary files
                 mimetype, disposition = 'application/octet-stream', 'attachment'
             else:
                 # do not just use the original mimetype, but force text/plain,
                 # otherwise it would serve text/html and that might be unsafe.
                 # Note: underlying vcs library fakes text/plain mimetype if the
                 # mimetype can not be determined and it thinks it is not
                 # binary.This might lead to erroneous text display in some
                 # cases, but helps in other cases, like with text files
                 # without extension.
                 mimetype, disposition = 'text/plain', 'inline'
 
         if disposition == 'attachment':
             disposition = self._get_attachement_headers(f_path)
 
         stream_content = file_node.stream_bytes()
 
         response = Response(app_iter=stream_content)
         response.content_disposition = disposition
         response.content_type = mimetype
 
         charset = self._get_default_encoding(c)
         if charset:
             response.charset = charset
 
         return response
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_file_download(self):
         c = self.load_default_context()
 
         commit_id, f_path = self._get_commit_and_path()
         commit = self._get_commit_or_redirect(commit_id)
         file_node = self._get_filenode_or_redirect(commit, f_path)
 
         if self.request.GET.get('lf'):
             # only if lf get flag is passed, we download this file
             # as LFS/Largefile
             lf_node = file_node.get_largefile_node()
             if lf_node:
                 # overwrite our pointer with the REAL large-file
                 file_node = lf_node
 
         disposition = self._get_attachement_headers(f_path)
 
         stream_content = file_node.stream_bytes()
 
         response = Response(app_iter=stream_content)
         response.content_disposition = disposition
         response.content_type = file_node.mimetype
 
         charset = self._get_default_encoding(c)
         if charset:
             response.charset = charset
 
         return response
 
     def _get_nodelist_at_commit(self, repo_name, repo_id, commit_id, f_path):
 
         cache_seconds = safe_int(
             rhodecode.CONFIG.get('rc_cache.cache_repo.expiration_time'))
         cache_on = cache_seconds > 0
         log.debug(
             'Computing FILE SEARCH for repo_id %s commit_id `%s` and path `%s`'
             'with caching: %s[TTL: %ss]' % (
                 repo_id, commit_id, f_path, cache_on, cache_seconds or 0))
 
         cache_namespace_uid = f'repo.{repo_id}'
         region = rc_cache.get_or_create_region('cache_repo', cache_namespace_uid)
 
         @region.conditional_cache_on_arguments(namespace=cache_namespace_uid, condition=cache_on)
         def compute_file_search(_name_hash, _repo_id, _commit_id, _f_path):
             log.debug('Generating cached nodelist for repo_id:%s, %s, %s',
                       _repo_id, commit_id, f_path)
             try:
                 _d, _f = ScmModel().get_quick_filter_nodes(repo_name, _commit_id, _f_path)
             except (RepositoryError, CommitDoesNotExistError, Exception) as e:
                 log.exception(safe_str(e))
                 h.flash(h.escape(safe_str(e)), category='error')
                 raise HTTPFound(h.route_path(
                     'repo_files', repo_name=self.db_repo_name,
                     commit_id='tip', f_path='/'))
 
             return _d + _f
 
         result = compute_file_search(self.db_repo.repo_name_hash, self.db_repo.repo_id,
                                      commit_id, f_path)
         return filter(lambda n: self.path_filter.path_access_allowed(n['name']), result)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_nodelist(self):
         self.load_default_context()
 
         commit_id, f_path = self._get_commit_and_path()
         commit = self._get_commit_or_redirect(commit_id)
 
         metadata = self._get_nodelist_at_commit(
             self.db_repo_name, self.db_repo.repo_id, commit.raw_id, f_path)
         return {'nodes': [x for x in metadata]}
 
     def _create_references(self, branches_or_tags, symbolic_reference, f_path, ref_type):
         items = []
         for name, commit_id in branches_or_tags.items():
             sym_ref = symbolic_reference(commit_id, name, f_path, ref_type)
             items.append((sym_ref, name, ref_type))
         return items
 
     def _symbolic_reference(self, commit_id, name, f_path, ref_type):
         return commit_id
 
     def _symbolic_reference_svn(self, commit_id, name, f_path, ref_type):
         return commit_id
 
         # NOTE(dan): old code we used in "diff" mode compare
         new_f_path = vcspath.join(name, f_path)
         return f'{new_f_path}@{commit_id}'
 
     def _get_node_history(self, commit_obj, f_path, commits=None):
         """
         get commit history for given node
 
         :param commit_obj: commit to calculate history
         :param f_path: path for node to calculate history for
         :param commits: if passed don't calculate history and take
             commits defined in this list
         """
         _ = self.request.translate
 
         # calculate history based on tip
         tip = self.rhodecode_vcs_repo.get_commit()
         if commits is None:
             pre_load = ["author", "branch"]
             try:
                 commits = tip.get_path_history(f_path, pre_load=pre_load)
             except (NodeDoesNotExistError, CommitError):
                 # this node is not present at tip!
                 commits = commit_obj.get_path_history(f_path, pre_load=pre_load)
 
         history = []
         commits_group = ([], _("Changesets"))
         for commit in commits:
             branch = ' (%s)' % commit.branch if commit.branch else ''
             n_desc = f'r{commit.idx}:{commit.short_id}{branch}'
             commits_group[0].append((commit.raw_id, n_desc, 'sha'))
         history.append(commits_group)
 
         symbolic_reference = self._symbolic_reference
 
         if self.rhodecode_vcs_repo.alias == 'svn':
             adjusted_f_path = RepoFilesView.adjust_file_path_for_svn(
                 f_path, self.rhodecode_vcs_repo)
             if adjusted_f_path != f_path:
                 log.debug(
                     'Recognized svn tag or branch in file "%s", using svn '
                     'specific symbolic references', f_path)
                 f_path = adjusted_f_path
                 symbolic_reference = self._symbolic_reference_svn
 
         branches = self._create_references(
             self.rhodecode_vcs_repo.branches, symbolic_reference, f_path, 'branch')
         branches_group = (branches, _("Branches"))
 
         tags = self._create_references(
             self.rhodecode_vcs_repo.tags, symbolic_reference, f_path, 'tag')
         tags_group = (tags, _("Tags"))
 
         history.append(branches_group)
         history.append(tags_group)
 
         return history, commits
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_file_history(self):
         self.load_default_context()
 
         commit_id, f_path = self._get_commit_and_path()
         commit = self._get_commit_or_redirect(commit_id)
         file_node = self._get_filenode_or_redirect(commit, f_path)
 
         if file_node.is_file():
             file_history, _hist = self._get_node_history(commit, f_path)
 
             res = []
             for section_items, section in file_history:
                 items = []
                 for obj_id, obj_text, obj_type in section_items:
                     at_rev = ''
                     if obj_type in ['branch', 'bookmark', 'tag']:
                         at_rev = obj_text
                     entry = {
                         'id': obj_id,
                         'text': obj_text,
                         'type': obj_type,
                         'at_rev': at_rev
                     }
 
                     items.append(entry)
 
                 res.append({
                     'text': section,
                     'children': items
                 })
 
             data = {
                 'more': False,
                 'results': res
             }
             return data
 
         log.warning('Cannot fetch history for directory')
         raise HTTPBadRequest()
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator(
         'repository.read', 'repository.write', 'repository.admin')
     def repo_file_authors(self):
         c = self.load_default_context()
 
         commit_id, f_path = self._get_commit_and_path()
         commit = self._get_commit_or_redirect(commit_id)
         file_node = self._get_filenode_or_redirect(commit, f_path)
 
         if not file_node.is_file():
             raise HTTPBadRequest()
 
         c.file_last_commit = file_node.last_commit
         if self.request.GET.get('annotate') == '1':
             # use _hist from annotation if annotation mode is on
             commit_ids = {x[1] for x in file_node.annotate}
             _hist = (
                 self.rhodecode_vcs_repo.get_commit(commit_id)
                 for commit_id in commit_ids)
         else:
             _f_history, _hist = self._get_node_history(commit, f_path)
         c.file_author = False
 
         unique = collections.OrderedDict()
         for commit in _hist:
             author = commit.author
             if author not in unique:
                 unique[commit.author] = [
                     h.email(author),
                     h.person(author, 'username_or_name_or_email'),
                     1  # counter
                 ]
 
             else:
                 # increase counter
                 unique[commit.author][2] += 1
 
         c.authors = [val for val in unique.values()]
 
         return self._get_template_context(c)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     def repo_files_check_head(self):
         self.load_default_context()
 
         commit_id, f_path = self._get_commit_and_path()
         _branch_name, _sha_commit_id, is_head = \
             self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                 landing_ref=self.db_repo.landing_ref_name)
 
         new_path = self.request.POST.get('path')
         operation = self.request.POST.get('operation')
         path_exist = ''
 
         if new_path and operation in ['create', 'upload']:
             new_f_path = os.path.join(f_path.lstrip('/'), new_path)
             try:
                 commit_obj = self.rhodecode_vcs_repo.get_commit(commit_id)
                 # NOTE(dan): construct whole path without leading /
                 file_node = commit_obj.get_node(new_f_path)
                 if file_node is not None:
                     path_exist = new_f_path
             except EmptyRepositoryError:
                 pass
             except Exception:
                 pass
 
         return {
             'branch': _branch_name,
             'sha': _sha_commit_id,
             'is_head': is_head,
             'path_exists': path_exist
         }
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     def repo_files_remove_file(self):
         _ = self.request.translate
         c = self.load_default_context()
         commit_id, f_path = self._get_commit_and_path()
 
         self._ensure_not_locked()
         _branch_name, _sha_commit_id, is_head = \
             self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                 landing_ref=self.db_repo.landing_ref_name)
 
         self.forbid_non_head(is_head, f_path)
         self.check_branch_permission(_branch_name)
 
         c.commit = self._get_commit_or_redirect(commit_id)
         c.file = self._get_filenode_or_redirect(c.commit, f_path)
 
         c.default_message = _(
             'Deleted file {} via RhodeCode Enterprise').format(f_path)
         c.f_path = f_path
 
         return self._get_template_context(c)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     @CSRFRequired()
     def repo_files_delete_file(self):
         _ = self.request.translate
 
         c = self.load_default_context()
         commit_id, f_path = self._get_commit_and_path()
 
         self._ensure_not_locked()
         _branch_name, _sha_commit_id, is_head = \
             self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                 landing_ref=self.db_repo.landing_ref_name)
 
         self.forbid_non_head(is_head, f_path)
         self.check_branch_permission(_branch_name)
 
         c.commit = self._get_commit_or_redirect(commit_id)
         c.file = self._get_filenode_or_redirect(c.commit, f_path)
 
         c.default_message = _(
             'Deleted file {} via RhodeCode Enterprise').format(f_path)
         c.f_path = f_path
         node_path = f_path
         author = self._rhodecode_db_user.full_contact
         message = self.request.POST.get('message') or c.default_message
         try:
             nodes = {
                 safe_bytes(node_path): {
                     'content': b''
                 }
             }
             ScmModel().delete_nodes(
                 user=self._rhodecode_db_user.user_id, repo=self.db_repo,
                 message=message,
                 nodes=nodes,
                 parent_commit=c.commit,
                 author=author,
             )
 
             h.flash(
                 _('Successfully deleted file `{}`').format(
                     h.escape(f_path)), category='success')
         except Exception:
             log.exception('Error during commit operation')
             h.flash(_('Error occurred during commit'), category='error')
         raise HTTPFound(
             h.route_path('repo_commit', repo_name=self.db_repo_name,
                          commit_id='tip'))
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     def repo_files_edit_file(self):
         _ = self.request.translate
         c = self.load_default_context()
         commit_id, f_path = self._get_commit_and_path()
 
         self._ensure_not_locked()
         _branch_name, _sha_commit_id, is_head = \
             self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                 landing_ref=self.db_repo.landing_ref_name)
 
         self.forbid_non_head(is_head, f_path, commit_id=commit_id)
         self.check_branch_permission(_branch_name, commit_id=commit_id)
 
         c.commit = self._get_commit_or_redirect(commit_id)
         c.file = self._get_filenode_or_redirect(c.commit, f_path)
 
         if c.file.is_binary:
             files_url = h.route_path(
                 'repo_files',
                 repo_name=self.db_repo_name,
                 commit_id=c.commit.raw_id, f_path=f_path)
             raise HTTPFound(files_url)
 
         c.default_message = _('Edited file {} via RhodeCode Enterprise').format(f_path)
         c.f_path = f_path
 
         return self._get_template_context(c)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     @CSRFRequired()
     def repo_files_update_file(self):
         _ = self.request.translate
         c = self.load_default_context()
         commit_id, f_path = self._get_commit_and_path()
 
         self._ensure_not_locked()
 
         c.commit = self._get_commit_or_redirect(commit_id)
         c.file = self._get_filenode_or_redirect(c.commit, f_path)
 
         if c.file.is_binary:
             raise HTTPFound(h.route_path('repo_files', repo_name=self.db_repo_name,
                                          commit_id=c.commit.raw_id, f_path=f_path))
 
         _branch_name, _sha_commit_id, is_head = \
             self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                 landing_ref=self.db_repo.landing_ref_name)
 
         self.forbid_non_head(is_head, f_path, commit_id=commit_id)
         self.check_branch_permission(_branch_name, commit_id=commit_id)
 
         c.default_message = _('Edited file {} via RhodeCode Enterprise').format(f_path)
         c.f_path = f_path
 
         old_content = c.file.str_content
         sl = old_content.splitlines(1)
         first_line = sl[0] if sl else ''
 
         r_post = self.request.POST
         # line endings:  0 - Unix, 1 - Mac, 2 - DOS
         line_ending_mode = detect_mode(first_line, 0)
         content = convert_line_endings(r_post.get('content', ''), line_ending_mode)
 
         message = r_post.get('message') or c.default_message
 
         org_node_path = c.file.str_path
         filename = r_post['filename']
 
         root_path = c.file.dir_path
         pure_path = self.create_pure_path(root_path, filename)
         node_path = pure_path.as_posix()
 
         default_redirect_url = h.route_path('repo_commit', repo_name=self.db_repo_name,
                                             commit_id=commit_id)
         if content == old_content and node_path == org_node_path:
             h.flash(_('No changes detected on {}').format(h.escape(org_node_path)),
                     category='warning')
             raise HTTPFound(default_redirect_url)
 
         try:
             mapping = {
                 c.file.bytes_path: {
                     'org_filename': org_node_path,
                     'filename': safe_bytes(node_path),
                     'content': safe_bytes(content),
                     'lexer': '',
                     'op': 'mod',
                     'mode': c.file.mode
                 }
             }
 
             commit = ScmModel().update_nodes(
                 user=self._rhodecode_db_user.user_id,
                 repo=self.db_repo,
                 message=message,
                 nodes=mapping,
                 parent_commit=c.commit,
             )
 
             h.flash(_('Successfully committed changes to file `{}`').format(
                     h.escape(f_path)), category='success')
             default_redirect_url = h.route_path(
                 'repo_commit', repo_name=self.db_repo_name, commit_id=commit.raw_id)
 
         except Exception:
             log.exception('Error occurred during commit')
             h.flash(_('Error occurred during commit'), category='error')
 
         raise HTTPFound(default_redirect_url)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     def repo_files_add_file(self):
         _ = self.request.translate
         c = self.load_default_context()
         commit_id, f_path = self._get_commit_and_path()
 
         self._ensure_not_locked()
 
         # Check if we need to use this page to upload binary
         upload_binary = str2bool(self.request.params.get('upload_binary', False))
 
         c.commit = self._get_commit_or_redirect(commit_id, redirect_after=False)
         if c.commit is None:
             c.commit = EmptyCommit(alias=self.rhodecode_vcs_repo.alias)
 
         if self.rhodecode_vcs_repo.is_empty():
             # for empty repository we cannot check for current branch, we rely on
             # c.commit.branch instead
             _branch_name, _sha_commit_id, is_head = c.commit.branch, '', True
         else:
             _branch_name, _sha_commit_id, is_head = \
                 self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                     landing_ref=self.db_repo.landing_ref_name)
 
         self.forbid_non_head(is_head, f_path, commit_id=commit_id)
         self.check_branch_permission(_branch_name, commit_id=commit_id)
 
         c.default_message = (_('Added file via RhodeCode Enterprise')) \
             if not upload_binary else (_('Edited file {} via RhodeCode Enterprise').format(f_path))
         c.f_path = f_path.lstrip('/')  # ensure not relative path
         c.replace_binary = upload_binary
 
         return self._get_template_context(c)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     @CSRFRequired()
     def repo_files_create_file(self):
         _ = self.request.translate
         c = self.load_default_context()
         commit_id, f_path = self._get_commit_and_path()
 
         self._ensure_not_locked()
 
         c.commit = self._get_commit_or_redirect(commit_id, redirect_after=False)
         if c.commit is None:
             c.commit = EmptyCommit(alias=self.rhodecode_vcs_repo.alias)
 
         # calculate redirect URL
         if self.rhodecode_vcs_repo.is_empty():
             default_redirect_url = h.route_path(
                 'repo_summary', repo_name=self.db_repo_name)
         else:
             default_redirect_url = h.route_path(
                 'repo_commit', repo_name=self.db_repo_name, commit_id='tip')
 
         if self.rhodecode_vcs_repo.is_empty():
             # for empty repository we cannot check for current branch, we rely on
             # c.commit.branch instead
             _branch_name, _sha_commit_id, is_head = c.commit.branch, '', True
         else:
             _branch_name, _sha_commit_id, is_head = \
                 self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                     landing_ref=self.db_repo.landing_ref_name)
 
         self.forbid_non_head(is_head, f_path, commit_id=commit_id)
         self.check_branch_permission(_branch_name, commit_id=commit_id)
 
         c.default_message = (_('Added file via RhodeCode Enterprise'))
         c.f_path = f_path
 
         r_post = self.request.POST
         message = r_post.get('message') or c.default_message
         filename = r_post.get('filename')
         unix_mode = 0
 
         if not filename:
             # If there's no commit, redirect to repo summary
             if type(c.commit) is EmptyCommit:
                 redirect_url = h.route_path(
                     'repo_summary', repo_name=self.db_repo_name)
             else:
                 redirect_url = default_redirect_url
             h.flash(_('No filename specified'), category='warning')
             raise HTTPFound(redirect_url)
 
         root_path = f_path
         pure_path = self.create_pure_path(root_path, filename)
         node_path = pure_path.as_posix().lstrip('/')
 
         author = self._rhodecode_db_user.full_contact
         content = convert_line_endings(r_post.get('content', ''), unix_mode)
         nodes = {
             safe_bytes(node_path): {
                 'content': safe_bytes(content)
             }
         }
 
         try:
 
             commit = ScmModel().create_nodes(
                 user=self._rhodecode_db_user.user_id,
                 repo=self.db_repo,
                 message=message,
                 nodes=nodes,
                 parent_commit=c.commit,
                 author=author,
             )
 
             h.flash(_('Successfully committed new file `{}`').format(
                     h.escape(node_path)), category='success')
 
             default_redirect_url = h.route_path(
                 'repo_commit', repo_name=self.db_repo_name, commit_id=commit.raw_id)
 
         except NonRelativePathError:
             log.exception('Non Relative path found')
             h.flash(_('The location specified must be a relative path and must not '
                       'contain .. in the path'), category='warning')
             raise HTTPFound(default_redirect_url)
         except (NodeError, NodeAlreadyExistsError) as e:
             h.flash(h.escape(safe_str(e)), category='error')
         except Exception:
             log.exception('Error occurred during commit')
             h.flash(_('Error occurred during commit'), category='error')
 
         raise HTTPFound(default_redirect_url)
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     @CSRFRequired()
     def repo_files_upload_file(self):
         _ = self.request.translate
         c = self.load_default_context()
         commit_id, f_path = self._get_commit_and_path()
 
         self._ensure_not_locked()
 
         c.commit = self._get_commit_or_redirect(commit_id, redirect_after=False)
         if c.commit is None:
             c.commit = EmptyCommit(alias=self.rhodecode_vcs_repo.alias)
 
         # calculate redirect URL
         if self.rhodecode_vcs_repo.is_empty():
             default_redirect_url = h.route_path(
                 'repo_summary', repo_name=self.db_repo_name)
         else:
             default_redirect_url = h.route_path(
                 'repo_commit', repo_name=self.db_repo_name, commit_id='tip')
 
         if self.rhodecode_vcs_repo.is_empty():
             # for empty repository we cannot check for current branch, we rely on
             # c.commit.branch instead
             _branch_name, _sha_commit_id, is_head = c.commit.branch, '', True
         else:
             _branch_name, _sha_commit_id, is_head = \
                 self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                     landing_ref=self.db_repo.landing_ref_name)
 
         error = self.forbid_non_head(is_head, f_path, json_mode=True)
         if error:
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
         error = self.check_branch_permission(_branch_name, json_mode=True)
         if error:
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
 
         c.default_message = (_('Added file via RhodeCode Enterprise'))
         c.f_path = f_path
 
         r_post = self.request.POST
 
         message = c.default_message
         user_message = r_post.getall('message')
         if isinstance(user_message, list) and user_message:
             # we take the first from duplicated results if it's not empty
             message = user_message[0] if user_message[0] else message
 
         nodes = {}
 
         for file_obj in r_post.getall('files_upload') or []:
             content = file_obj.file
             filename = file_obj.filename
 
             root_path = f_path
             pure_path = self.create_pure_path(root_path, filename)
             node_path = pure_path.as_posix().lstrip('/')
 
             nodes[safe_bytes(node_path)] = {
                 'content': content
             }
 
         if not nodes:
             error = 'missing files'
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
 
         author = self._rhodecode_db_user.full_contact
 
         try:
             commit = ScmModel().create_nodes(
                 user=self._rhodecode_db_user.user_id,
                 repo=self.db_repo,
                 message=message,
                 nodes=nodes,
                 parent_commit=c.commit,
                 author=author,
             )
             if len(nodes) == 1:
                 flash_message = _('Successfully committed {} new files').format(len(nodes))
             else:
                 flash_message = _('Successfully committed 1 new file')
 
             h.flash(flash_message, category='success')
 
             default_redirect_url = h.route_path(
                 'repo_commit', repo_name=self.db_repo_name, commit_id=commit.raw_id)
 
         except NonRelativePathError:
             log.exception('Non Relative path found')
             error = _('The location specified must be a relative path and must not '
                       'contain .. in the path')
             h.flash(error, category='warning')
 
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
         except (NodeError, NodeAlreadyExistsError) as e:
             error = h.escape(e)
             h.flash(error, category='error')
 
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
         except Exception:
             log.exception('Error occurred during commit')
             error = _('Error occurred during commit')
             h.flash(error, category='error')
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
 
         return {
             'error': None,
             'redirect_url': default_redirect_url
         }
 
     @LoginRequired()
     @HasRepoPermissionAnyDecorator('repository.write', 'repository.admin')
     @CSRFRequired()
     def repo_files_replace_file(self):
         _ = self.request.translate
         c = self.load_default_context()
         commit_id, f_path = self._get_commit_and_path()
 
         self._ensure_not_locked()
 
         c.commit = self._get_commit_or_redirect(commit_id, redirect_after=False)
         if c.commit is None:
             c.commit = EmptyCommit(alias=self.rhodecode_vcs_repo.alias)
 
         if self.rhodecode_vcs_repo.is_empty():
             default_redirect_url = h.route_path(
                 'repo_summary', repo_name=self.db_repo_name)
         else:
             default_redirect_url = h.route_path(
                 'repo_commit', repo_name=self.db_repo_name, commit_id='tip')
 
         if self.rhodecode_vcs_repo.is_empty():
             # for empty repository we cannot check for current branch, we rely on
             # c.commit.branch instead
             _branch_name, _sha_commit_id, is_head = c.commit.branch, '', True
         else:
             _branch_name, _sha_commit_id, is_head = \
                 self._is_valid_head(commit_id, self.rhodecode_vcs_repo,
                                     landing_ref=self.db_repo.landing_ref_name)
 
         error = self.forbid_non_head(is_head, f_path, json_mode=True)
         if error:
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
         error = self.check_branch_permission(_branch_name, json_mode=True)
         if error:
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
 
         c.default_message = (_('Edited file {} via RhodeCode Enterprise').format(f_path))
         c.f_path = f_path
 
         r_post = self.request.POST
 
         message = c.default_message
         user_message = r_post.getall('message')
         if isinstance(user_message, list) and user_message:
             # we take the first from duplicated results if it's not empty
             message = user_message[0] if user_message[0] else message
 
         data_for_replacement = r_post.getall('files_upload') or []
         if (objects_count := len(data_for_replacement)) > 1:
             return {
                 'error': 'too many files for replacement',
                 'redirect_url': default_redirect_url
             }
         elif not objects_count:
             return {
                 'error': 'missing files',
                 'redirect_url': default_redirect_url
             }
 
         content = data_for_replacement[0].file
         retrieved_filename = data_for_replacement[0].filename
 
         if retrieved_filename.split('.')[-1] != f_path.split('.')[-1]:
             return {
                 'error': 'file extension of uploaded file doesn\'t match an original file\'s extension',
                 'redirect_url': default_redirect_url
             }
 
         author = self._rhodecode_db_user.full_contact
 
         try:
             commit = ScmModel().update_binary_node(
                 user=self._rhodecode_db_user.user_id,
                 repo=self.db_repo,
                 message=message,
                 node={
                     'content': content,
                     'file_path': f_path.encode(),
                 },
                 parent_commit=c.commit,
                 author=author,
             )
 
             h.flash(_('Successfully committed 1 new file'), category='success')
 
             default_redirect_url = h.route_path(
                 'repo_commit', repo_name=self.db_repo_name, commit_id=commit.raw_id)
 
         except (NodeError, NodeAlreadyExistsError) as e:
             error = h.escape(e)
             h.flash(error, category='error')
 
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
         except Exception:
             log.exception('Error occurred during commit')
             error = _('Error occurred during commit')
             h.flash(error, category='error')
             return {
                 'error': error,
                 'redirect_url': default_redirect_url
             }
 
         return {
             'error': None,
             'redirect_url': default_redirect_url
         }

rhodecode/config/config_maker.py

@@ -1,201 +1,205 b''
 # Copyright (C) 2010-2023 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 import os
 import tempfile
 import logging
 
 from pyramid.settings import asbool
 
 from rhodecode.config.settings_maker import SettingsMaker
 from rhodecode.config import utils as config_utils
 
 log = logging.getLogger(__name__)
 
 
 def sanitize_settings_and_apply_defaults(global_config, settings):
     """
     Applies settings defaults and does all type conversion.
 
     We would move all settings parsing and preparation into this place, so that
     we have only one place left which deals with this part. The remaining parts
     of the application would start to rely fully on well-prepared settings.
 
     This piece would later be split up per topic to avoid a big fat monster
     function.
     """
     jn = os.path.join
 
     global_settings_maker = SettingsMaker(global_config)
     global_settings_maker.make_setting('debug', default=False, parser='bool')
     debug_enabled = asbool(global_config.get('debug'))
 
     settings_maker = SettingsMaker(settings)
 
     settings_maker.make_setting(
         'logging.autoconfigure',
         default=False,
         parser='bool')
 
     logging_conf = jn(os.path.dirname(global_config.get('__file__')), 'logging.ini')
     settings_maker.enable_logging(logging_conf, level='INFO' if debug_enabled else 'DEBUG')
 
     # Default includes, possible to change as a user
     pyramid_includes = settings_maker.make_setting('pyramid.includes', [], parser='list:newline')
     log.debug(
         "Using the following pyramid.includes: %s",
         pyramid_includes)
 
     settings_maker.make_setting('rhodecode.edition', 'Community Edition')
     settings_maker.make_setting('rhodecode.edition_id', 'CE')
 
     if 'mako.default_filters' not in settings:
         # set custom default filters if we don't have it defined
         settings['mako.imports'] = 'from rhodecode.lib.base import h_filter'
         settings['mako.default_filters'] = 'h_filter'
 
     if 'mako.directories' not in settings:
         mako_directories = settings.setdefault('mako.directories', [
             # Base templates of the original application
             'rhodecode:templates',
         ])
         log.debug(
             "Using the following Mako template directories: %s",
             mako_directories)
 
     # NOTE(marcink): fix redis requirement for schema of connection since 3.X
     if 'beaker.session.type' in settings and settings['beaker.session.type'] == 'ext:redis':
         raw_url = settings['beaker.session.url']
         if not raw_url.startswith(('redis://', 'rediss://', 'unix://')):
             settings['beaker.session.url'] = 'redis://' + raw_url
 
     settings_maker.make_setting('__file__', global_config.get('__file__'))
 
     # TODO: johbo: Re-think this, usually the call to config.include
     # should allow to pass in a prefix.
     settings_maker.make_setting('rhodecode.api.url', '/_admin/api')
 
     # Sanitize generic settings.
     settings_maker.make_setting('default_encoding', 'UTF-8', parser='list')
     settings_maker.make_setting('gzip_responses', False, parser='bool')
     settings_maker.make_setting('startup.import_repos', 'false', parser='bool')
 
     # statsd
     settings_maker.make_setting('statsd.enabled', False, parser='bool')
     settings_maker.make_setting('statsd.statsd_host', 'statsd-exporter', parser='string')
     settings_maker.make_setting('statsd.statsd_port', 9125, parser='int')
     settings_maker.make_setting('statsd.statsd_prefix', '')
     settings_maker.make_setting('statsd.statsd_ipv6', False, parser='bool')
 
     settings_maker.make_setting('vcs.svn.compatible_version', '')
     settings_maker.make_setting('vcs.svn.proxy.enabled', True, parser='bool')
     settings_maker.make_setting('vcs.svn.proxy.host', 'http://svn:8090', parser='string')
     settings_maker.make_setting('vcs.hooks.protocol', 'http')
     settings_maker.make_setting('vcs.hooks.host', '*')
     settings_maker.make_setting('vcs.scm_app_implementation', 'http')
     settings_maker.make_setting('vcs.server', '')
     settings_maker.make_setting('vcs.server.protocol', 'http')
     settings_maker.make_setting('vcs.server.enable', 'true', parser='bool')
     settings_maker.make_setting('vcs.hooks.direct_calls', 'false', parser='bool')
     settings_maker.make_setting('vcs.start_server', 'false', parser='bool')
     settings_maker.make_setting('vcs.backends', 'hg, git, svn', parser='list')
     settings_maker.make_setting('vcs.connection_timeout', 3600, parser='int')
 
     settings_maker.make_setting('vcs.methods.cache', True, parser='bool')
 
     # repo_store path
     settings_maker.make_setting('repo_store.path', '/var/opt/rhodecode_repo_store')
     # Support legacy values of vcs.scm_app_implementation. Legacy
     # configurations may use 'rhodecode.lib.middleware.utils.scm_app_http', or
     # disabled since 4.13 'vcsserver.scm_app' which is now mapped to 'http'.
     scm_app_impl = settings['vcs.scm_app_implementation']
     if scm_app_impl in ['rhodecode.lib.middleware.utils.scm_app_http', 'vcsserver.scm_app']:
         settings['vcs.scm_app_implementation'] = 'http'
 
     settings_maker.make_setting('appenlight', False, parser='bool')
 
     temp_store = tempfile.gettempdir()
     tmp_cache_dir = jn(temp_store, 'rc_cache')
 
     # save default, cache dir, and use it for all backends later.
     default_cache_dir = settings_maker.make_setting(
         'cache_dir',
         default=tmp_cache_dir, default_when_empty=True,
         parser='dir:ensured')
 
     # exception store cache
     settings_maker.make_setting(
         'exception_tracker.store_path',
         default=jn(default_cache_dir, 'exc_store'), default_when_empty=True,
         parser='dir:ensured'
     )
 
     settings_maker.make_setting(
         'celerybeat-schedule.path',
         default=jn(default_cache_dir, 'celerybeat_schedule', 'celerybeat-schedule.db'), default_when_empty=True,
         parser='file:ensured'
     )
 
     settings_maker.make_setting('exception_tracker.send_email', False, parser='bool')
     settings_maker.make_setting('exception_tracker.email_prefix', '[RHODECODE ERROR]', default_when_empty=True)
 
     # sessions, ensure file since no-value is memory
     settings_maker.make_setting('beaker.session.type', 'file')
     settings_maker.make_setting('beaker.session.data_dir',  jn(default_cache_dir, 'session_data'))
 
     # cache_general
     settings_maker.make_setting('rc_cache.cache_general.backend', 'dogpile.cache.rc.file_namespace')
     settings_maker.make_setting('rc_cache.cache_general.expiration_time', 60 * 60 * 12, parser='int')
     settings_maker.make_setting('rc_cache.cache_general.arguments.filename', jn(default_cache_dir, 'rhodecode_cache_general.db'))
 
     # cache_perms
     settings_maker.make_setting('rc_cache.cache_perms.backend', 'dogpile.cache.rc.file_namespace')
     settings_maker.make_setting('rc_cache.cache_perms.expiration_time', 60 * 60, parser='int')
     settings_maker.make_setting('rc_cache.cache_perms.arguments.filename', jn(default_cache_dir, 'rhodecode_cache_perms_db'))
 
     # cache_repo
     settings_maker.make_setting('rc_cache.cache_repo.backend', 'dogpile.cache.rc.file_namespace')
     settings_maker.make_setting('rc_cache.cache_repo.expiration_time', 60 * 60 * 24 * 30, parser='int')
     settings_maker.make_setting('rc_cache.cache_repo.arguments.filename', jn(default_cache_dir, 'rhodecode_cache_repo_db'))
 
     # cache_license
     settings_maker.make_setting('rc_cache.cache_license.backend', 'dogpile.cache.rc.file_namespace')
     settings_maker.make_setting('rc_cache.cache_license.expiration_time', 60 * 5, parser='int')
     settings_maker.make_setting('rc_cache.cache_license.arguments.filename', jn(default_cache_dir, 'rhodecode_cache_license_db'))
 
     # cache_repo_longterm memory, 96H
     settings_maker.make_setting('rc_cache.cache_repo_longterm.backend', 'dogpile.cache.rc.memory_lru')
     settings_maker.make_setting('rc_cache.cache_repo_longterm.expiration_time', 345600, parser='int')
     settings_maker.make_setting('rc_cache.cache_repo_longterm.max_size', 10000, parser='int')
 
     # sql_cache_short
     settings_maker.make_setting('rc_cache.sql_cache_short.backend', 'dogpile.cache.rc.memory_lru')
     settings_maker.make_setting('rc_cache.sql_cache_short.expiration_time', 30, parser='int')
     settings_maker.make_setting('rc_cache.sql_cache_short.max_size', 10000, parser='int')
 
     # archive_cache
-    settings_maker.make_setting('archive_cache.store_dir', jn(default_cache_dir, 'archive_cache'), default_when_empty=True,)
-    settings_maker.make_setting('archive_cache.cache_size_gb', 10, parser='float')
-    settings_maker.make_setting('archive_cache.cache_shards', 10, parser='int')
+    settings_maker.make_setting('archive_cache.locking.url', 'redis://redis:6379/1')
+    settings_maker.make_setting('archive_cache.backend.type', 'filesystem')
+
+    settings_maker.make_setting('archive_cache.filesystem.store_dir', jn(default_cache_dir, 'archive_cache'), default_when_empty=True,)
+    settings_maker.make_setting('archive_cache.filesystem.cache_size_gb', 10, parser='float')
+    settings_maker.make_setting('archive_cache.filesystem.cache_shards', 8, parser='int')
+    settings_maker.make_setting('archive_cache.filesystem.eviction_policy', 'least-recently-stored')
 
     settings_maker.env_expand()
 
     # configure instance id
     config_utils.set_instance_id(settings)
 
     return settings

rhodecode/lib/rc_cache/archive_cache/fanout_cache.py

@@ -1,88 +1,264 b''
-# Copyright (C) 2015-2023 RhodeCode GmbH
+# Copyright (C) 2015-2024 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
-import logging
+import codecs
+import contextlib
+import functools
 import os
-import diskcache
-from diskcache import RLock
+import logging
+import time
+import typing
+import zlib
+
+from rhodecode.lib.ext_json import json
+from .lock import GenerationLock
 
 log = logging.getLogger(__name__)
 
 cache_meta = None
 
+UNKNOWN = -241
+NO_VAL = -917
 
-class ReentrantLock(RLock):
-    def __enter__(self):
-        reentrant_lock_key = self._key
+MODE_BINARY = 'BINARY'
+
+
+class FileSystemCache:
+
+    def __init__(self, index, directory, **settings):
+        self._index = index
+        self._directory = directory
+
+    def _write_file(self, full_path, iterator, mode, encoding=None):
+        full_dir, _ = os.path.split(full_path)
+
+        for count in range(1, 11):
+            with contextlib.suppress(OSError):
+                os.makedirs(full_dir)
+
+            try:
+                # Another cache may have deleted the directory before
+                # the file could be opened.
+                writer = open(full_path, mode, encoding=encoding)
+            except OSError:
+                if count == 10:
+                    # Give up after 10 tries to open the file.
+                    raise
+                continue
+
+            with writer:
+                size = 0
+                for chunk in iterator:
+                    size += len(chunk)
+                    writer.write(chunk)
+                return size
+
+    def _get_keyfile(self, key):
+        return os.path.join(self._directory, f'{key}.key')
+
+    def store(self, key, value_reader, metadata):
+        filename, full_path = self.random_filename()
+        key_file = self._get_keyfile(key)
+
+        # STORE METADATA
+        _metadata = {
+            "version": "v1",
+            "timestamp": time.time(),
+            "filename": filename,
+            "full_path": full_path,
+            "key_file": key_file,
+        }
+        if metadata:
+            _metadata.update(metadata)
+
+        reader = functools.partial(value_reader.read, 2**22)
+
+        iterator = iter(reader, b'')
+        size = self._write_file(full_path, iterator, 'xb')
+
+        # after archive is finished, we create a key to save the presence of the binary file
+        with open(key_file, 'wb') as f:
+            f.write(json.dumps(_metadata))
+
+        return key, size, MODE_BINARY, filename, _metadata
+
+    def fetch(self, key) -> tuple[typing.BinaryIO, dict]:
+        if key not in self:
+            raise KeyError(key)
+
+        key_file = self._get_keyfile(key)
+        with open(key_file, 'rb') as f:
+            metadata = json.loads(f.read())
+
+        filename = metadata['filename']
+
+        return open(os.path.join(self._directory, filename), 'rb'), metadata
+
+    def random_filename(self):
+        """Return filename and full-path tuple for file storage.
+
+        Filename will be a randomly generated 28 character hexadecimal string
+        with ".archive_cache" suffixed. Two levels of sub-directories will be used to
+        reduce the size of directories. On older filesystems, lookups in
+        directories with many files may be slow.
+        """
 
-        log.debug('Acquire ReentrantLock(key=%s) for archive cache generation...', reentrant_lock_key)
-        #self.acquire()
-        log.debug('Lock for key=%s acquired', reentrant_lock_key)
+        hex_name = codecs.encode(os.urandom(16), 'hex').decode('utf-8')
+        sub_dir = os.path.join(hex_name[:2], hex_name[2:4])
+        name = hex_name[4:] + '.archive_cache'
+        filename = os.path.join(sub_dir, name)
+        full_path = os.path.join(self._directory, filename)
+        return filename, full_path
+
+    def hash(self, key):
+        """Compute portable hash for `key`.
+
+        :param key: key to hash
+        :return: hash value
+
+        """
+        mask = 0xFFFFFFFF
+        return zlib.adler32(key.encode('utf-8')) & mask  # noqa
+
+    def __contains__(self, key):
+        """Return `True` if `key` matching item is found in cache.
+
+        :param key: key matching item
+        :return: True if key matching item
+
+        """
+        key_file = self._get_keyfile(key)
+        return os.path.exists(key_file)
+
+
+class FanoutCache:
+    """Cache that shards keys and values."""
+
+    def __init__(
+        self, directory=None, **settings
+    ):
+        """Initialize cache instance.
+
+        :param str directory: cache directory
+        :param settings: settings dict
+
+        """
+        if directory is None:
+            raise ValueError('directory cannot be None')
+
+        directory = str(directory)
+        directory = os.path.expanduser(directory)
+        directory = os.path.expandvars(directory)
+        self._directory = directory
 
-    def __exit__(self, *exc_info):
-        #self.release()
-        pass
+        self._count = settings.pop('cache_shards')
+        self._locking_url = settings.pop('locking_url')
+
+        self._shards = tuple(
+            FileSystemCache(
+                index=num,
+                directory=os.path.join(directory, 'shard_%03d' % num),
+                **settings,
+            )
+            for num in range(self._count)
+        )
+        self._hash = self._shards[0].hash
+
+    def get_lock(self, lock_key):
+        return GenerationLock(lock_key, self._locking_url)
+
+    def _get_shard(self, key) -> FileSystemCache:
+        index = self._hash(key) % self._count
+        shard = self._shards[index]
+        return shard
+
+    def store(self, key, value_reader, metadata=None):
+        shard = self._get_shard(key)
+        return shard.store(key, value_reader, metadata)
+
+    def fetch(self, key):
+        """Return file handle corresponding to `key` from cache.
+        """
+        shard = self._get_shard(key)
+        return shard.fetch(key)
+
+    def has_key(self, key):
+        """Return `True` if `key` matching item is found in cache.
+
+        :param key: key for item
+        :return: True if key is found
+
+        """
+        shard = self._get_shard(key)
+        return key in shard
+
+    def __contains__(self, item):
+        return self.has_key(item)
+
+    def evict(self):
+        """Remove old items based on the conditions"""
+        # TODO: Implement this...
+        return
 
 
 def get_archival_config(config):
 
     final_config = {
-        'archive_cache.eviction_policy': 'least-frequently-used'
+
     }
 
     for k, v in config.items():
         if k.startswith('archive_cache'):
             final_config[k] = v
 
     return final_config
 
 
 def get_archival_cache_store(config):
 
     global cache_meta
     if cache_meta is not None:
         return cache_meta
 
     config = get_archival_config(config)
+    backend = config['archive_cache.backend.type']
+    if backend != 'filesystem':
+        raise ValueError('archive_cache.backend.type only supports "filesystem"')
 
-    archive_cache_dir = config['archive_cache.store_dir']
-    archive_cache_size_gb = config['archive_cache.cache_size_gb']
-    archive_cache_shards = config['archive_cache.cache_shards']
-    archive_cache_eviction_policy = config['archive_cache.eviction_policy']
+    archive_cache_locking_url = config['archive_cache.locking.url']
+    archive_cache_dir = config['archive_cache.filesystem.store_dir']
+    archive_cache_size_gb = config['archive_cache.filesystem.cache_size_gb']
+    archive_cache_shards = config['archive_cache.filesystem.cache_shards']
+    archive_cache_eviction_policy = config['archive_cache.filesystem.eviction_policy']
 
     log.debug('Initializing archival cache instance under %s', archive_cache_dir)
 
     # check if it's ok to write, and re-create the archive cache
     if not os.path.isdir(archive_cache_dir):
         os.makedirs(archive_cache_dir, exist_ok=True)
 
-    d_cache = diskcache.FanoutCache(
-        archive_cache_dir, shards=archive_cache_shards,
-        cull_limit=0,  # manual eviction required
-        size_limit=archive_cache_size_gb * 1024 * 1024 * 1024,
-        eviction_policy=archive_cache_eviction_policy,
-        timeout=30
+    d_cache = FanoutCache(
+        archive_cache_dir,
+        locking_url=archive_cache_locking_url,
+        cache_shards=archive_cache_shards,
+        cache_size_limit=archive_cache_size_gb * 1024 * 1024 * 1024,
+        cache_eviction_policy=archive_cache_eviction_policy
     )
     cache_meta = d_cache
     return cache_meta
 
-
-def includeme(config):
-    # init our cache at start
-    settings = config.get_settings()
-    get_archival_cache_store(settings)

rhodecode/lib/system_info.py

@@ -1,844 +1,851 b''
 # Copyright (C) 2017-2023 RhodeCode GmbH
 #
 # This program is free software: you can redistribute it and/or modify
 # it under the terms of the GNU Affero General Public License, version 3
 # (only), as published by the Free Software Foundation.
 #
 # This program is distributed in the hope that it will be useful,
 # but WITHOUT ANY WARRANTY; without even the implied warranty of
 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 # GNU General Public License for more details.
 #
 # You should have received a copy of the GNU Affero General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 # This program is dual-licensed. If you wish to learn more about the
 # RhodeCode Enterprise Edition, including its added features, Support services,
 # and proprietary license terms, please see https://rhodecode.com/licenses/
 
 
 import os
 import sys
 import time
 import platform
 import collections
 import psutil
 from functools import wraps
 
 import pkg_resources
 import logging
 import resource
 
 import configparser
 
 from rc_license.models import LicenseModel
 from rhodecode.lib.str_utils import safe_str
 
 log = logging.getLogger(__name__)
 
 
 _NA = 'NOT AVAILABLE'
 _NA_FLOAT = 0.0
 
 STATE_OK = 'ok'
 STATE_ERR = 'error'
 STATE_WARN = 'warning'
 
 STATE_OK_DEFAULT = {'message': '', 'type': STATE_OK}
 
 
 registered_helpers = {}
 
 
 def register_sysinfo(func):
     """
     @register_helper
     def db_check():
         pass
 
     db_check == registered_helpers['db_check']
     """
     global registered_helpers
     registered_helpers[func.__name__] = func
 
     @wraps(func)
     def _wrapper(*args, **kwargs):
         return func(*args, **kwargs)
     return _wrapper
 
 
 # HELPERS
 def percentage(part: (int, float), whole: (int, float)):
     whole = float(whole)
     if whole > 0:
         return round(100 * float(part) / whole, 1)
     return 0.0
 
 
 def get_storage_size(storage_path):
     sizes = []
     for file_ in os.listdir(storage_path):
         storage_file = os.path.join(storage_path, file_)
         if os.path.isfile(storage_file):
             try:
                 sizes.append(os.path.getsize(storage_file))
             except OSError:
                 log.exception('Failed to get size of storage file %s', storage_file)
                 pass
 
     return sum(sizes)
 
 
 def get_resource(resource_type):
     try:
         return resource.getrlimit(resource_type)
     except Exception:
         return 'NOT_SUPPORTED'
 
 
 def get_cert_path(ini_path):
     default = '/etc/ssl/certs/ca-certificates.crt'
     control_ca_bundle = os.path.join(
         os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(ini_path)))),
         '.rccontrol-profile/etc/ca-bundle.crt')
     if os.path.isfile(control_ca_bundle):
         default = control_ca_bundle
 
     return default
 
 
 class SysInfoRes(object):
     def __init__(self, value, state=None, human_value=None):
         self.value = value
         self.state = state or STATE_OK_DEFAULT
         self.human_value = human_value or value
 
     def __json__(self):
         return {
             'value': self.value,
             'state': self.state,
             'human_value': self.human_value,
         }
 
     def get_value(self):
         return self.__json__()
 
     def __str__(self):
         return f'<SysInfoRes({self.__json__()})>'
 
 
 class SysInfo(object):
 
     def __init__(self, func_name, **kwargs):
         self.function_name = func_name
         self.value = _NA
         self.state = None
         self.kwargs = kwargs or {}
 
     def __call__(self):
         computed = self.compute(**self.kwargs)
         if not isinstance(computed, SysInfoRes):
             raise ValueError(
                 'computed value for {} is not instance of '
                 '{}, got {} instead'.format(
                     self.function_name, SysInfoRes, type(computed)))
         return computed.__json__()
 
     def __str__(self):
         return f'<SysInfo({self.function_name})>'
 
     def compute(self, **kwargs):
         return self.function_name(**kwargs)
 
 
 # SysInfo functions
 @register_sysinfo
 def python_info():
     value = dict(version=f'{platform.python_version()}:{platform.python_implementation()}',
                  executable=sys.executable)
     return SysInfoRes(value=value)
 
 
 @register_sysinfo
 def py_modules():
     mods = dict([(p.project_name, {'version': p.version, 'location': p.location})
                  for p in pkg_resources.working_set])
 
     value = sorted(mods.items(), key=lambda k: k[0].lower())
     return SysInfoRes(value=value)
 
 
 @register_sysinfo
 def platform_type():
     from rhodecode.lib.utils import generate_platform_uuid
 
     value = dict(
         name=safe_str(platform.platform()),
         uuid=generate_platform_uuid()
     )
     return SysInfoRes(value=value)
 
 
 @register_sysinfo
 def locale_info():
     import locale
 
     def safe_get_locale(locale_name):
         try:
             locale.getlocale(locale_name)
         except TypeError:
             return f'FAILED_LOCALE_GET:{locale_name}'
 
     value = dict(
         locale_default=locale.getlocale(),
         locale_lc_all=safe_get_locale(locale.LC_ALL),
         locale_lc_ctype=safe_get_locale(locale.LC_CTYPE),
         lang_env=os.environ.get('LANG'),
         lc_all_env=os.environ.get('LC_ALL'),
         local_archive_env=os.environ.get('LOCALE_ARCHIVE'),
     )
     human_value = \
         f"LANG: {value['lang_env']}, \
           locale LC_ALL: {value['locale_lc_all']},  \
           locale LC_CTYPE: {value['locale_lc_ctype']},  \
           Default locales: {value['locale_default']}"
 
     return SysInfoRes(value=value, human_value=human_value)
 
 
 @register_sysinfo
 def ulimit_info():
     data = collections.OrderedDict([
         ('cpu time (seconds)', get_resource(resource.RLIMIT_CPU)),
         ('file size', get_resource(resource.RLIMIT_FSIZE)),
         ('stack size', get_resource(resource.RLIMIT_STACK)),
         ('core file size', get_resource(resource.RLIMIT_CORE)),
         ('address space size', get_resource(resource.RLIMIT_AS)),
         ('locked in mem size', get_resource(resource.RLIMIT_MEMLOCK)),
         ('heap size', get_resource(resource.RLIMIT_DATA)),
         ('rss size', get_resource(resource.RLIMIT_RSS)),
         ('number of processes', get_resource(resource.RLIMIT_NPROC)),
         ('open files', get_resource(resource.RLIMIT_NOFILE)),
     ])
 
     text = ', '.join(f'{k}:{v}' for k, v in data.items())
 
     value = {
         'limits': data,
         'text': text,
     }
     return SysInfoRes(value=value)
 
 
 @register_sysinfo
 def uptime():
     from rhodecode.lib.helpers import age, time_to_datetime
     from rhodecode.translation import TranslationString
 
     value = dict(boot_time=0, uptime=0, text='')
     state = STATE_OK_DEFAULT
 
     boot_time = psutil.boot_time()
     value['boot_time'] = boot_time
     value['uptime'] = time.time() - boot_time
 
     date_or_age = age(time_to_datetime(boot_time))
     if isinstance(date_or_age, TranslationString):
         date_or_age = date_or_age.interpolate()
 
     human_value = value.copy()
     human_value['boot_time'] = time_to_datetime(boot_time)
     human_value['uptime'] = age(time_to_datetime(boot_time), show_suffix=False)
 
     human_value['text'] = f'Server started {date_or_age}'
     return SysInfoRes(value=value, human_value=human_value)
 
 
 @register_sysinfo
 def memory():
     from rhodecode.lib.helpers import format_byte_size_binary
     value = dict(available=0, used=0, used_real=0, cached=0, percent=0,
                  percent_used=0, free=0, inactive=0, active=0, shared=0,
                  total=0, buffers=0, text='')
 
     state = STATE_OK_DEFAULT
 
     value.update(dict(psutil.virtual_memory()._asdict()))
     value['used_real'] = value['total'] - value['available']
     value['percent_used'] = psutil._common.usage_percent(value['used_real'], value['total'], 1)
 
     human_value = value.copy()
     human_value['text'] = '{}/{}, {}% used'.format(
         format_byte_size_binary(value['used_real']),
         format_byte_size_binary(value['total']),
         value['percent_used'])
 
     keys = list(value.keys())[::]
     keys.pop(keys.index('percent'))
     keys.pop(keys.index('percent_used'))
     keys.pop(keys.index('text'))
     for k in keys:
         human_value[k] = format_byte_size_binary(value[k])
 
     if state['type'] == STATE_OK and value['percent_used'] > 90:
         msg = 'Critical: your available RAM memory is very low.'
         state = {'message': msg, 'type': STATE_ERR}
 
     elif state['type'] == STATE_OK and value['percent_used'] > 70:
         msg = 'Warning: your available RAM memory is running low.'
         state = {'message': msg, 'type': STATE_WARN}
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def machine_load():
     value = {'1_min': _NA_FLOAT, '5_min': _NA_FLOAT, '15_min': _NA_FLOAT, 'text': ''}
     state = STATE_OK_DEFAULT
 
     # load averages
     if hasattr(psutil.os, 'getloadavg'):
         value.update(dict(
             list(zip(['1_min', '5_min', '15_min'], psutil.os.getloadavg()))
         ))
 
     human_value = value.copy()
     human_value['text'] = '1min: {}, 5min: {}, 15min: {}'.format(
         value['1_min'], value['5_min'], value['15_min'])
 
     if state['type'] == STATE_OK and value['15_min'] > 5.0:
         msg = 'Warning: your machine load is very high.'
         state = {'message': msg, 'type': STATE_WARN}
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def cpu():
     value = {'cpu': 0, 'cpu_count': 0, 'cpu_usage': []}
     state = STATE_OK_DEFAULT
 
     value['cpu'] = psutil.cpu_percent(0.5)
     value['cpu_usage'] = psutil.cpu_percent(0.5, percpu=True)
     value['cpu_count'] = psutil.cpu_count()
 
     human_value = value.copy()
     human_value['text'] = '{} cores at {} %'.format(value['cpu_count'], value['cpu'])
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def storage():
     from rhodecode.lib.helpers import format_byte_size_binary
     from rhodecode.lib.utils import get_rhodecode_repo_store_path
     path = get_rhodecode_repo_store_path()
 
     value = dict(percent=0, used=0, total=0, path=path, text='')
     state = STATE_OK_DEFAULT
 
     try:
         value.update(dict(psutil.disk_usage(path)._asdict()))
     except Exception as e:
         log.exception('Failed to fetch disk info')
         state = {'message': str(e), 'type': STATE_ERR}
 
     human_value = value.copy()
     human_value['used'] = format_byte_size_binary(value['used'])
     human_value['total'] = format_byte_size_binary(value['total'])
     human_value['text'] = "{}/{}, {}% used".format(
         format_byte_size_binary(value['used']),
         format_byte_size_binary(value['total']),
         value['percent'])
 
     if state['type'] == STATE_OK and value['percent'] > 90:
         msg = 'Critical: your disk space is very low.'
         state = {'message': msg, 'type': STATE_ERR}
 
     elif state['type'] == STATE_OK and value['percent'] > 70:
         msg = 'Warning: your disk space is running low.'
         state = {'message': msg, 'type': STATE_WARN}
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def storage_inodes():
     from rhodecode.lib.utils import get_rhodecode_repo_store_path
     path = get_rhodecode_repo_store_path()
 
     value = dict(percent=0.0, free=0, used=0, total=0, path=path, text='')
     state = STATE_OK_DEFAULT
 
     try:
         i_stat = os.statvfs(path)
         value['free'] = i_stat.f_ffree
         value['used'] = i_stat.f_files-i_stat.f_favail
         value['total'] = i_stat.f_files
         value['percent'] = percentage(value['used'], value['total'])
     except Exception as e:
         log.exception('Failed to fetch disk inodes info')
         state = {'message': str(e), 'type': STATE_ERR}
 
     human_value = value.copy()
     human_value['text'] = "{}/{}, {}% used".format(
         value['used'], value['total'], value['percent'])
 
     if state['type'] == STATE_OK and value['percent'] > 90:
         msg = 'Critical: your disk free inodes are very low.'
         state = {'message': msg, 'type': STATE_ERR}
 
     elif state['type'] == STATE_OK and value['percent'] > 70:
         msg = 'Warning: your disk free inodes are running low.'
         state = {'message': msg, 'type': STATE_WARN}
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def storage_archives():
     import rhodecode
     from rhodecode.lib.utils import safe_str
     from rhodecode.lib.helpers import format_byte_size_binary
 
-    msg = 'Archive cache storage is controlled by ' \
-          'archive_cache.store_dir=/path/to/cache option in the .ini file'
-    path = safe_str(rhodecode.CONFIG.get('archive_cache.store_dir', msg))
+    storage_type = rhodecode.ConfigGet().get_str('archive_cache.backend.type')
+    storage_key = 'archive_cache.filesystem.store_dir'
+
+    default_msg = 'Archive cache storage is controlled by '\
+                  f'{storage_key}=/path/to/cache option in the .ini file'
+    path = rhodecode.ConfigGet().get_str(storage_key, missing=default_msg)
 
     value = dict(percent=0, used=0, total=0, items=0, path=path, text='')
     state = STATE_OK_DEFAULT
     try:
+        if storage_type != 'filesystem':
+            # raise Exc to stop reporting on different type
+            raise ValueError('Storage type must be "filesystem"')
+
         items_count = 0
         used = 0
         for root, dirs, files in os.walk(path):
             if root == path:
                 items_count = len(dirs)
 
             for f in files:
                 try:
                     used += os.path.getsize(os.path.join(root, f))
                 except OSError:
                     pass
         value.update({
             'percent': 100,
             'used': used,
             'total': used,
             'items': items_count
         })
 
     except Exception as e:
         log.exception('failed to fetch archive cache storage')
         state = {'message': str(e), 'type': STATE_ERR}
 
     human_value = value.copy()
     human_value['used'] = format_byte_size_binary(value['used'])
     human_value['total'] = format_byte_size_binary(value['total'])
     human_value['text'] = "{} ({} items)".format(
         human_value['used'], value['items'])
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def storage_gist():
     from rhodecode.model.gist import GIST_STORE_LOC
     from rhodecode.lib.utils import safe_str, get_rhodecode_repo_store_path
     from rhodecode.lib.helpers import format_byte_size_binary
     path = safe_str(os.path.join(
         get_rhodecode_repo_store_path(), GIST_STORE_LOC))
 
     # gist storage
     value = dict(percent=0, used=0, total=0, items=0, path=path, text='')
     state = STATE_OK_DEFAULT
 
     try:
         items_count = 0
         used = 0
         for root, dirs, files in os.walk(path):
             if root == path:
                 items_count = len(dirs)
 
             for f in files:
                 try:
                     used += os.path.getsize(os.path.join(root, f))
                 except OSError:
                     pass
         value.update({
             'percent': 100,
             'used': used,
             'total': used,
             'items': items_count
         })
     except Exception as e:
         log.exception('failed to fetch gist storage items')
         state = {'message': str(e), 'type': STATE_ERR}
 
     human_value = value.copy()
     human_value['used'] = format_byte_size_binary(value['used'])
     human_value['total'] = format_byte_size_binary(value['total'])
     human_value['text'] = "{} ({} items)".format(
         human_value['used'], value['items'])
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def storage_temp():
     import tempfile
     from rhodecode.lib.helpers import format_byte_size_binary
 
     path = tempfile.gettempdir()
     value = dict(percent=0, used=0, total=0, items=0, path=path, text='')
     state = STATE_OK_DEFAULT
 
     if not psutil:
         return SysInfoRes(value=value, state=state)
 
     try:
         value.update(dict(psutil.disk_usage(path)._asdict()))
     except Exception as e:
         log.exception('Failed to fetch temp dir info')
         state = {'message': str(e), 'type': STATE_ERR}
 
     human_value = value.copy()
     human_value['used'] = format_byte_size_binary(value['used'])
     human_value['total'] = format_byte_size_binary(value['total'])
     human_value['text'] = "{}/{}, {}% used".format(
         format_byte_size_binary(value['used']),
         format_byte_size_binary(value['total']),
         value['percent'])
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def search_info():
     import rhodecode
     from rhodecode.lib.index import searcher_from_config
 
     backend = rhodecode.CONFIG.get('search.module', '')
     location = rhodecode.CONFIG.get('search.location', '')
 
     try:
         searcher = searcher_from_config(rhodecode.CONFIG)
         searcher = searcher.__class__.__name__
     except Exception:
         searcher = None
 
     value = dict(
         backend=backend, searcher=searcher, location=location, text='')
     state = STATE_OK_DEFAULT
 
     human_value = value.copy()
     human_value['text'] = "backend:`{}`".format(human_value['backend'])
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def git_info():
     from rhodecode.lib.vcs.backends import git
     state = STATE_OK_DEFAULT
     value = human_value = ''
     try:
         value = git.discover_git_version(raise_on_exc=True)
         human_value = f'version reported from VCSServer: {value}'
     except Exception as e:
         state = {'message': str(e), 'type': STATE_ERR}
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def hg_info():
     from rhodecode.lib.vcs.backends import hg
     state = STATE_OK_DEFAULT
     value = human_value = ''
     try:
         value = hg.discover_hg_version(raise_on_exc=True)
         human_value = f'version reported from VCSServer: {value}'
     except Exception as e:
         state = {'message': str(e), 'type': STATE_ERR}
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def svn_info():
     from rhodecode.lib.vcs.backends import svn
     state = STATE_OK_DEFAULT
     value = human_value = ''
     try:
         value = svn.discover_svn_version(raise_on_exc=True)
         human_value = f'version reported from VCSServer: {value}'
     except Exception as e:
         state = {'message': str(e), 'type': STATE_ERR}
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def vcs_backends():
     import rhodecode
     value = rhodecode.CONFIG.get('vcs.backends')
     human_value = 'Enabled backends in order: {}'.format(','.join(value))
     return SysInfoRes(value=value, human_value=human_value)
 
 
 @register_sysinfo
 def vcs_server():
     import rhodecode
     from rhodecode.lib.vcs.backends import get_vcsserver_service_data
 
     server_url = rhodecode.CONFIG.get('vcs.server')
     enabled = rhodecode.CONFIG.get('vcs.server.enable')
     protocol = rhodecode.CONFIG.get('vcs.server.protocol') or 'http'
     state = STATE_OK_DEFAULT
     version = None
     workers = 0
 
     try:
         data = get_vcsserver_service_data()
         if data and 'version' in data:
             version = data['version']
 
         if data and 'config' in data:
             conf = data['config']
             workers = conf.get('workers', 'NOT AVAILABLE')
 
         connection = 'connected'
     except Exception as e:
         connection = 'failed'
         state = {'message': str(e), 'type': STATE_ERR}
 
     value = dict(
         url=server_url,
         enabled=enabled,
         protocol=protocol,
         connection=connection,
         version=version,
         text='',
     )
 
     human_value = value.copy()
     human_value['text'] = \
         '{url}@ver:{ver} via {mode} mode[workers:{workers}], connection:{conn}'.format(
             url=server_url, ver=version, workers=workers, mode=protocol,
             conn=connection)
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def vcs_server_config():
     from rhodecode.lib.vcs.backends import get_vcsserver_service_data
     state = STATE_OK_DEFAULT
 
     value = {}
     try:
         data = get_vcsserver_service_data()
         value = data['app_config']
     except Exception as e:
         state = {'message': str(e), 'type': STATE_ERR}
 
     human_value = value.copy()
     human_value['text'] = 'VCS Server config'
 
     return SysInfoRes(value=value, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def rhodecode_app_info():
     import rhodecode
     edition = rhodecode.CONFIG.get('rhodecode.edition')
 
     value = dict(
         rhodecode_version=rhodecode.__version__,
         rhodecode_lib_path=os.path.abspath(rhodecode.__file__),
         text=''
     )
     human_value = value.copy()
     human_value['text'] = 'RhodeCode {edition}, version {ver}'.format(
         edition=edition, ver=value['rhodecode_version']
     )
     return SysInfoRes(value=value, human_value=human_value)
 
 
 @register_sysinfo
 def rhodecode_config():
     import rhodecode
     path = rhodecode.CONFIG.get('__file__')
     rhodecode_ini_safe = rhodecode.CONFIG.copy()
     cert_path = get_cert_path(path)
 
     try:
         config = configparser.ConfigParser()
         config.read(path)
         parsed_ini = config
         if parsed_ini.has_section('server:main'):
             parsed_ini = dict(parsed_ini.items('server:main'))
     except Exception:
         log.exception('Failed to read .ini file for display')
         parsed_ini = {}
 
     rhodecode_ini_safe['server:main'] = parsed_ini
 
     blacklist = [
         f'rhodecode_{LicenseModel.LICENSE_DB_KEY}',
         'routes.map',
         'sqlalchemy.db1.url',
         'channelstream.secret',
         'beaker.session.secret',
         'rhodecode.encrypted_values.secret',
         'rhodecode_auth_github_consumer_key',
         'rhodecode_auth_github_consumer_secret',
         'rhodecode_auth_google_consumer_key',
         'rhodecode_auth_google_consumer_secret',
         'rhodecode_auth_bitbucket_consumer_secret',
         'rhodecode_auth_bitbucket_consumer_key',
         'rhodecode_auth_twitter_consumer_secret',
         'rhodecode_auth_twitter_consumer_key',
 
         'rhodecode_auth_twitter_secret',
         'rhodecode_auth_github_secret',
         'rhodecode_auth_google_secret',
         'rhodecode_auth_bitbucket_secret',
 
         'appenlight.api_key',
         ('app_conf', 'sqlalchemy.db1.url')
     ]
     for k in blacklist:
         if isinstance(k, tuple):
             section, key = k
             if section in rhodecode_ini_safe:
                 rhodecode_ini_safe[section] = '**OBFUSCATED**'
         else:
             rhodecode_ini_safe.pop(k, None)
 
     # TODO: maybe put some CONFIG checks here ?
     return SysInfoRes(value={'config': rhodecode_ini_safe,
                              'path': path, 'cert_path': cert_path})
 
 
 @register_sysinfo
 def database_info():
     import rhodecode
     from sqlalchemy.engine import url as engine_url
     from rhodecode.model import meta
     from rhodecode.model.meta import Session
     from rhodecode.model.db import DbMigrateVersion
 
     state = STATE_OK_DEFAULT
 
     db_migrate = DbMigrateVersion.query().filter(
         DbMigrateVersion.repository_id == 'rhodecode_db_migrations').one()
 
     db_url_obj = engine_url.make_url(rhodecode.CONFIG['sqlalchemy.db1.url'])
 
     try:
         engine = meta.get_engine()
         db_server_info = engine.dialect._get_server_version_info(
             Session.connection(bind=engine))
         db_version = '.'.join(map(str, db_server_info))
     except Exception:
         log.exception('failed to fetch db version')
         db_version = 'UNKNOWN'
 
     db_info = dict(
         migrate_version=db_migrate.version,
         type=db_url_obj.get_backend_name(),
         version=db_version,
         url=repr(db_url_obj)
     )
     current_version = db_migrate.version
     expected_version = rhodecode.__dbversion__
     if state['type'] == STATE_OK and current_version != expected_version:
         msg = 'Critical: database schema mismatch, ' \
               'expected version {}, got {}. ' \
               'Please run migrations on your database.'.format(
             expected_version, current_version)
         state = {'message': msg, 'type': STATE_ERR}
 
     human_value = db_info.copy()
     human_value['url'] = "{} @ migration version: {}".format(
         db_info['url'], db_info['migrate_version'])
     human_value['version'] = "{} {}".format(db_info['type'], db_info['version'])
     return SysInfoRes(value=db_info, state=state, human_value=human_value)
 
 
 @register_sysinfo
 def server_info(environ):
     import rhodecode
     from rhodecode.lib.base import get_server_ip_addr, get_server_port
 
     value = {
         'server_ip': '{}:{}'.format(
             get_server_ip_addr(environ, log_errors=False),
             get_server_port(environ)
         ),
         'server_id': rhodecode.CONFIG.get('instance_id'),
     }
     return SysInfoRes(value=value)
 
 
 @register_sysinfo
 def usage_info():
     from rhodecode.model.db import User, Repository, true
     value = {
         'users': User.query().count(),
         'users_active': User.query().filter(User.active == true()).count(),
         'repositories': Repository.query().count(),
         'repository_types': {
             'hg': Repository.query().filter(
                 Repository.repo_type == 'hg').count(),
             'git': Repository.query().filter(
                 Repository.repo_type == 'git').count(),
             'svn': Repository.query().filter(
                 Repository.repo_type == 'svn').count(),
         },
     }
     return SysInfoRes(value=value)
 
 
 def get_system_info(environ):
     environ = environ or {}
     return {
         'rhodecode_app': SysInfo(rhodecode_app_info)(),
         'rhodecode_config': SysInfo(rhodecode_config)(),
         'rhodecode_usage': SysInfo(usage_info)(),
         'python': SysInfo(python_info)(),
         'py_modules': SysInfo(py_modules)(),
 
         'platform': SysInfo(platform_type)(),
         'locale': SysInfo(locale_info)(),
         'server': SysInfo(server_info, environ=environ)(),
         'database': SysInfo(database_info)(),
         'ulimit': SysInfo(ulimit_info)(),
         'storage': SysInfo(storage)(),
         'storage_inodes': SysInfo(storage_inodes)(),
         'storage_archive': SysInfo(storage_archives)(),
         'storage_gist': SysInfo(storage_gist)(),
         'storage_temp': SysInfo(storage_temp)(),
 
         'search': SysInfo(search_info)(),
 
         'uptime': SysInfo(uptime)(),
         'load': SysInfo(machine_load)(),
         'cpu': SysInfo(cpu)(),
         'memory': SysInfo(memory)(),
 
         'vcs_backends': SysInfo(vcs_backends)(),
         'vcs_server': SysInfo(vcs_server)(),
 
         'vcs_server_config': SysInfo(vcs_server_config)(),
 
         'git': SysInfo(git_info)(),
         'hg': SysInfo(hg_info)(),
         'svn': SysInfo(svn_info)(),
     }
 
 
 def load_system_info(key):
     """
     get_sys_info('vcs_server')
     get_sys_info('database')
     """
     return SysInfo(registered_helpers[key])()