caches: introduce invalidation as a safer ways to expire keys, deleting them are more problematic.

rhodecode/apps/_base/subscribers.py

@@ -1,55 +1,56 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2016-2020 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
 
 from rhodecode import events
 from rhodecode.lib import rc_cache
 
 log = logging.getLogger(__name__)
 
 # names of namespaces used for different permission related cached
 # during flush operation we need to take care of all those
 cache_namespaces = [
     'cache_user_auth.{}',
     'cache_user_repo_acl_ids.{}',
     'cache_user_user_group_acl_ids.{}',
     'cache_user_repo_group_acl_ids.{}'
 ]
 
 
 def trigger_user_permission_flush(event):
     """
     Subscriber to the `UserPermissionsChange`. This triggers the
     automatic flush of permission caches, so the users affected receive new permissions
     Right Away
     """
-
+    invalidate = True
     affected_user_ids = set(event.user_ids)
     for user_id in affected_user_ids:
         for cache_namespace_uid_tmpl in cache_namespaces:
             cache_namespace_uid = cache_namespace_uid_tmpl.format(user_id)
-            del_keys = rc_cache.clear_cache_namespace('cache_perms', cache_namespace_uid)
-            log.debug('Deleted %s cache keys for user_id: %s and namespace %s',
+            del_keys = rc_cache.clear_cache_namespace(
+                'cache_perms', cache_namespace_uid, invalidate=invalidate)
+            log.debug('Invalidated %s cache keys for user_id: %s and namespace %s',
                       del_keys, user_id, cache_namespace_uid)
 
 
 def includeme(config):
     config.add_subscriber(trigger_user_permission_flush, events.UserPermissionsChange)

rhodecode/lib/rc_cache/utils.py

@@ -1,415 +1,418 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2015-2020 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 time
 import logging
 import functools
 import threading
 
 from dogpile.cache import CacheRegion
 from dogpile.cache.util import compat
 
 import rhodecode
 from rhodecode.lib.utils import safe_str, sha1
 from rhodecode.lib.utils2 import safe_unicode, str2bool
 from rhodecode.model.db import Session, CacheKey, IntegrityError
 
 from rhodecode.lib.rc_cache import cache_key_meta
 from rhodecode.lib.rc_cache import region_meta
 
 log = logging.getLogger(__name__)
 
 
 def isCython(func):
     """
     Private helper that checks if a function is a cython function.
     """
     return func.__class__.__name__ == 'cython_function_or_method'
 
 
 class RhodeCodeCacheRegion(CacheRegion):
 
     def conditional_cache_on_arguments(
             self, namespace=None,
             expiration_time=None,
             should_cache_fn=None,
             to_str=compat.string_type,
             function_key_generator=None,
             condition=True):
         """
         Custom conditional decorator, that will not touch any dogpile internals if
         condition isn't meet. This works a bit different than should_cache_fn
         And it's faster in cases we don't ever want to compute cached values
         """
         expiration_time_is_callable = compat.callable(expiration_time)
 
         if function_key_generator is None:
             function_key_generator = self.function_key_generator
 
         # workaround for py2 and cython problems, this block should be removed
         # once we've migrated to py3
         if 'cython' == 'cython':
             def decorator(fn):
                 if to_str is compat.string_type:
                     # backwards compatible
                     key_generator = function_key_generator(namespace, fn)
                 else:
                     key_generator = function_key_generator(namespace, fn, to_str=to_str)
 
                 @functools.wraps(fn)
                 def decorate(*arg, **kw):
                     key = key_generator(*arg, **kw)
 
                     @functools.wraps(fn)
                     def creator():
                         return fn(*arg, **kw)
 
                     if not condition:
                         return creator()
 
                     timeout = expiration_time() if expiration_time_is_callable \
                         else expiration_time
 
                     return self.get_or_create(key, creator, timeout, should_cache_fn)
 
                 def invalidate(*arg, **kw):
                     key = key_generator(*arg, **kw)
                     self.delete(key)
 
                 def set_(value, *arg, **kw):
                     key = key_generator(*arg, **kw)
                     self.set(key, value)
 
                 def get(*arg, **kw):
                     key = key_generator(*arg, **kw)
                     return self.get(key)
 
                 def refresh(*arg, **kw):
                     key = key_generator(*arg, **kw)
                     value = fn(*arg, **kw)
                     self.set(key, value)
                     return value
 
                 decorate.set = set_
                 decorate.invalidate = invalidate
                 decorate.refresh = refresh
                 decorate.get = get
                 decorate.original = fn
                 decorate.key_generator = key_generator
                 decorate.__wrapped__ = fn
 
                 return decorate
             return decorator
 
         def get_or_create_for_user_func(key_generator, user_func, *arg, **kw):
 
             if not condition:
                 log.debug('Calling un-cached func:%s', user_func.func_name)
                 return user_func(*arg, **kw)
 
             key = key_generator(*arg, **kw)
 
             timeout = expiration_time() if expiration_time_is_callable \
                 else expiration_time
 
             log.debug('Calling cached fn:%s', user_func.func_name)
             return self.get_or_create(key, user_func, timeout, should_cache_fn, (arg, kw))
 
         def cache_decorator(user_func):
             if to_str is compat.string_type:
                 # backwards compatible
                 key_generator = function_key_generator(namespace, user_func)
             else:
                 key_generator = function_key_generator(namespace, user_func, to_str=to_str)
 
             def refresh(*arg, **kw):
                 """
                 Like invalidate, but regenerates the value instead
                 """
                 key = key_generator(*arg, **kw)
                 value = user_func(*arg, **kw)
                 self.set(key, value)
                 return value
 
             def invalidate(*arg, **kw):
                 key = key_generator(*arg, **kw)
                 self.delete(key)
 
             def set_(value, *arg, **kw):
                 key = key_generator(*arg, **kw)
                 self.set(key, value)
 
             def get(*arg, **kw):
                 key = key_generator(*arg, **kw)
                 return self.get(key)
 
             user_func.set = set_
             user_func.invalidate = invalidate
             user_func.get = get
             user_func.refresh = refresh
             user_func.key_generator = key_generator
             user_func.original = user_func
 
             # Use `decorate` to preserve the signature of :param:`user_func`.
             return decorator.decorate(user_func, functools.partial(
                 get_or_create_for_user_func, key_generator))
 
         return cache_decorator
 
 
 def make_region(*arg, **kw):
     return RhodeCodeCacheRegion(*arg, **kw)
 
 
 def get_default_cache_settings(settings, prefixes=None):
     prefixes = prefixes or []
     cache_settings = {}
     for key in settings.keys():
         for prefix in prefixes:
             if key.startswith(prefix):
                 name = key.split(prefix)[1].strip()
                 val = settings[key]
                 if isinstance(val, compat.string_types):
                     val = val.strip()
                 cache_settings[name] = val
     return cache_settings
 
 
 def compute_key_from_params(*args):
     """
     Helper to compute key from given params to be used in cache manager
     """
     return sha1("_".join(map(safe_str, args)))
 
 
 def backend_key_generator(backend):
     """
     Special wrapper that also sends over the backend to the key generator
     """
     def wrapper(namespace, fn):
         return key_generator(backend, namespace, fn)
     return wrapper
 
 
 def key_generator(backend, namespace, fn):
     fname = fn.__name__
 
     def generate_key(*args):
         backend_prefix = getattr(backend, 'key_prefix', None) or 'backend_prefix'
         namespace_pref = namespace or 'default_namespace'
         arg_key = compute_key_from_params(*args)
         final_key = "{}:{}:{}_{}".format(backend_prefix, namespace_pref, fname, arg_key)
 
         return final_key
 
     return generate_key
 
 
 def get_or_create_region(region_name, region_namespace=None):
     from rhodecode.lib.rc_cache.backends import FileNamespaceBackend
     region_obj = region_meta.dogpile_cache_regions.get(region_name)
     if not region_obj:
         raise EnvironmentError(
             'Region `{}` not in configured: {}.'.format(
                 region_name, region_meta.dogpile_cache_regions.keys()))
 
     region_uid_name = '{}:{}'.format(region_name, region_namespace)
     if isinstance(region_obj.actual_backend, FileNamespaceBackend):
         region_exist = region_meta.dogpile_cache_regions.get(region_namespace)
         if region_exist:
             log.debug('Using already configured region: %s', region_namespace)
             return region_exist
         cache_dir = region_meta.dogpile_config_defaults['cache_dir']
         expiration_time = region_obj.expiration_time
 
         if not os.path.isdir(cache_dir):
             os.makedirs(cache_dir)
         new_region = make_region(
             name=region_uid_name,
             function_key_generator=backend_key_generator(region_obj.actual_backend)
         )
         namespace_filename = os.path.join(
             cache_dir, "{}.cache.dbm".format(region_namespace))
         # special type that allows 1db per namespace
         new_region.configure(
             backend='dogpile.cache.rc.file_namespace',
             expiration_time=expiration_time,
             arguments={"filename": namespace_filename}
         )
 
         # create and save in region caches
         log.debug('configuring new region: %s', region_uid_name)
         region_obj = region_meta.dogpile_cache_regions[region_namespace] = new_region
 
     return region_obj
 
 
-def clear_cache_namespace(cache_region, cache_namespace_uid):
+def clear_cache_namespace(cache_region, cache_namespace_uid, invalidate=False):
     region = get_or_create_region(cache_region, cache_namespace_uid)
     cache_keys = region.backend.list_keys(prefix=cache_namespace_uid)
     num_delete_keys = len(cache_keys)
+    if invalidate:
+        region.invalidate(hard=False)
+    else:
     if num_delete_keys:
         region.delete_multi(cache_keys)
     return num_delete_keys
 
 
 class ActiveRegionCache(object):
     def __init__(self, context, cache_data):
         self.context = context
         self.cache_data = cache_data
 
     def should_invalidate(self):
         return False
 
 
 class FreshRegionCache(object):
     def __init__(self, context, cache_data):
         self.context = context
         self.cache_data = cache_data
 
     def should_invalidate(self):
         return True
 
 
 class InvalidationContext(object):
     """
     usage::
 
         from rhodecode.lib import rc_cache
 
         cache_namespace_uid = CacheKey.SOME_NAMESPACE.format(1)
         region = rc_cache.get_or_create_region('cache_perms', cache_namespace_uid)
 
         @region.conditional_cache_on_arguments(namespace=cache_namespace_uid, condition=True)
         def heavy_compute(cache_name, param1, param2):
             print('COMPUTE {}, {}, {}'.format(cache_name, param1, param2))
 
         # invalidation namespace is shared namespace key for all process caches
         # we use it to send a global signal
         invalidation_namespace = 'repo_cache:1'
 
         inv_context_manager = rc_cache.InvalidationContext(
             uid=cache_namespace_uid, invalidation_namespace=invalidation_namespace)
         with inv_context_manager as invalidation_context:
             args = ('one', 'two')
             # re-compute and store cache if we get invalidate signal
             if invalidation_context.should_invalidate():
                 result = heavy_compute.refresh(*args)
             else:
                 result = heavy_compute(*args)
 
             compute_time = inv_context_manager.compute_time
             log.debug('result computed in %.4fs', compute_time)
 
         # To send global invalidation signal, simply run
         CacheKey.set_invalidate(invalidation_namespace)
 
     """
 
     def __repr__(self):
         return '<InvalidationContext:{}[{}]>'.format(
             safe_str(self.cache_key), safe_str(self.uid))
 
     def __init__(self, uid, invalidation_namespace='',
                  raise_exception=False, thread_scoped=None):
         self.uid = uid
         self.invalidation_namespace = invalidation_namespace
         self.raise_exception = raise_exception
         self.proc_id = safe_unicode(rhodecode.CONFIG.get('instance_id') or 'DEFAULT')
         self.thread_id = 'global'
 
         if thread_scoped is None:
             # if we set "default" we can override this via .ini settings
             thread_scoped = str2bool(rhodecode.CONFIG.get('cache_thread_scoped'))
 
         # Append the thread id to the cache key if this invalidation context
         # should be scoped to the current thread.
         if thread_scoped is True:
             self.thread_id = threading.current_thread().ident
 
         self.cache_key = compute_key_from_params(uid)
         self.cache_key = 'proc:{}|thread:{}|params:{}'.format(
             self.proc_id, self.thread_id, self.cache_key)
         self.compute_time = 0
 
     def get_or_create_cache_obj(self, cache_type, invalidation_namespace=''):
         invalidation_namespace = invalidation_namespace or self.invalidation_namespace
         # fetch all cache keys for this namespace and convert them to a map to find if we
         # have specific cache_key object registered. We do this because we want to have
         # all consistent cache_state_uid for newly registered objects
         cache_obj_map = CacheKey.get_namespace_map(invalidation_namespace)
         cache_obj = cache_obj_map.get(self.cache_key)
         log.debug('Fetched cache obj %s using %s cache key.', cache_obj, self.cache_key)
         if not cache_obj:
             new_cache_args = invalidation_namespace
             first_cache_obj = next(cache_obj_map.itervalues()) if cache_obj_map else None
             cache_state_uid = None
             if first_cache_obj:
                 cache_state_uid = first_cache_obj.cache_state_uid
             cache_obj = CacheKey(self.cache_key, cache_args=new_cache_args,
                                  cache_state_uid=cache_state_uid)
             cache_key_meta.cache_keys_by_pid.append(self.cache_key)
 
         return cache_obj
 
     def __enter__(self):
         """
         Test if current object is valid, and return CacheRegion function
         that does invalidation and calculation
         """
         log.debug('Entering cache invalidation check context: %s', self.invalidation_namespace)
         # register or get a new key based on uid
         self.cache_obj = self.get_or_create_cache_obj(cache_type=self.uid)
         cache_data = self.cache_obj.get_dict()
         self._start_time = time.time()
         if self.cache_obj.cache_active:
             # means our cache obj is existing and marked as it's
             # cache is not outdated, we return ActiveRegionCache
             self.skip_cache_active_change = True
 
             return ActiveRegionCache(context=self, cache_data=cache_data)
 
         # the key is either not existing or set to False, we return
         # the real invalidator which re-computes value. We additionally set
         # the flag to actually update the Database objects
         self.skip_cache_active_change = False
         return FreshRegionCache(context=self, cache_data=cache_data)
 
     def __exit__(self, exc_type, exc_val, exc_tb):
         # save compute time
         self.compute_time = time.time() - self._start_time
 
         if self.skip_cache_active_change:
             return
 
         try:
             self.cache_obj.cache_active = True
             Session().add(self.cache_obj)
             Session().commit()
         except IntegrityError:
             # if we catch integrity error, it means we inserted this object
             # assumption is that's really an edge race-condition case and
             # it's safe is to skip it
             Session().rollback()
         except Exception:
             log.exception('Failed to commit on cache key update')
             Session().rollback()
             if self.raise_exception:
                 raise

rhodecode/model/scm.py

@@ -1,1024 +1,1025 b''
 # -*- coding: utf-8 -*-
 
 # Copyright (C) 2010-2020 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/
 
 """
 Scm model for RhodeCode
 """
 
 import os.path
 import traceback
 import logging
 import cStringIO
 
 from sqlalchemy import func
 from zope.cachedescriptors.property import Lazy as LazyProperty
 
 import rhodecode
 from rhodecode.lib.vcs import get_backend
 from rhodecode.lib.vcs.exceptions import RepositoryError, NodeNotChangedError
 from rhodecode.lib.vcs.nodes import FileNode
 from rhodecode.lib.vcs.backends.base import EmptyCommit
 from rhodecode.lib import helpers as h, rc_cache
 from rhodecode.lib.auth import (
     HasRepoPermissionAny, HasRepoGroupPermissionAny,
     HasUserGroupPermissionAny)
 from rhodecode.lib.exceptions import NonRelativePathError, IMCCommitError
 from rhodecode.lib import hooks_utils
 from rhodecode.lib.utils import (
     get_filesystem_repos, make_db_config)
 from rhodecode.lib.utils2 import (safe_str, safe_unicode)
 from rhodecode.lib.system_info import get_system_info
 from rhodecode.model import BaseModel
 from rhodecode.model.db import (
     or_, false,
     Repository, CacheKey, UserFollowing, UserLog, User, RepoGroup,
     PullRequest, FileStore)
 from rhodecode.model.settings import VcsSettingsModel
 from rhodecode.model.validation_schema.validators import url_validator, InvalidCloneUrl
 
 log = logging.getLogger(__name__)
 
 
 class UserTemp(object):
     def __init__(self, user_id):
         self.user_id = user_id
 
     def __repr__(self):
         return "<%s('id:%s')>" % (self.__class__.__name__, self.user_id)
 
 
 class RepoTemp(object):
     def __init__(self, repo_id):
         self.repo_id = repo_id
 
     def __repr__(self):
         return "<%s('id:%s')>" % (self.__class__.__name__, self.repo_id)
 
 
 class SimpleCachedRepoList(object):
     """
     Lighter version of of iteration of repos without the scm initialisation,
     and with cache usage
     """
     def __init__(self, db_repo_list, repos_path, order_by=None, perm_set=None):
         self.db_repo_list = db_repo_list
         self.repos_path = repos_path
         self.order_by = order_by
         self.reversed = (order_by or '').startswith('-')
         if not perm_set:
             perm_set = ['repository.read', 'repository.write',
                         'repository.admin']
         self.perm_set = perm_set
 
     def __len__(self):
         return len(self.db_repo_list)
 
     def __repr__(self):
         return '<%s (%s)>' % (self.__class__.__name__, self.__len__())
 
     def __iter__(self):
         for dbr in self.db_repo_list:
             # check permission at this level
             has_perm = HasRepoPermissionAny(*self.perm_set)(
                 dbr.repo_name, 'SimpleCachedRepoList check')
             if not has_perm:
                 continue
 
             tmp_d = {
                 'name': dbr.repo_name,
                 'dbrepo': dbr.get_dict(),
                 'dbrepo_fork': dbr.fork.get_dict() if dbr.fork else {}
             }
             yield tmp_d
 
 
 class _PermCheckIterator(object):
 
     def __init__(
             self, obj_list, obj_attr, perm_set, perm_checker,
             extra_kwargs=None):
         """
         Creates iterator from given list of objects, additionally
         checking permission for them from perm_set var
 
         :param obj_list: list of db objects
         :param obj_attr: attribute of object to pass into perm_checker
         :param perm_set: list of permissions to check
         :param perm_checker: callable to check permissions against
         """
         self.obj_list = obj_list
         self.obj_attr = obj_attr
         self.perm_set = perm_set
         self.perm_checker = perm_checker(*self.perm_set)
         self.extra_kwargs = extra_kwargs or {}
 
     def __len__(self):
         return len(self.obj_list)
 
     def __repr__(self):
         return '<%s (%s)>' % (self.__class__.__name__, self.__len__())
 
     def __iter__(self):
         for db_obj in self.obj_list:
             # check permission at this level
             # NOTE(marcink): the __dict__.get() is ~4x faster then getattr()
             name = db_obj.__dict__.get(self.obj_attr, None)
             if not self.perm_checker(name, self.__class__.__name__, **self.extra_kwargs):
                 continue
 
             yield db_obj
 
 
 class RepoList(_PermCheckIterator):
 
     def __init__(self, db_repo_list, perm_set=None, extra_kwargs=None):
         if not perm_set:
             perm_set = ['repository.read', 'repository.write', 'repository.admin']
 
         super(RepoList, self).__init__(
             obj_list=db_repo_list,
             obj_attr='_repo_name', perm_set=perm_set,
             perm_checker=HasRepoPermissionAny,
             extra_kwargs=extra_kwargs)
 
 
 class RepoGroupList(_PermCheckIterator):
 
     def __init__(self, db_repo_group_list, perm_set=None, extra_kwargs=None):
         if not perm_set:
             perm_set = ['group.read', 'group.write', 'group.admin']
 
         super(RepoGroupList, self).__init__(
             obj_list=db_repo_group_list,
             obj_attr='_group_name', perm_set=perm_set,
             perm_checker=HasRepoGroupPermissionAny,
             extra_kwargs=extra_kwargs)
 
 
 class UserGroupList(_PermCheckIterator):
 
     def __init__(self, db_user_group_list, perm_set=None, extra_kwargs=None):
         if not perm_set:
             perm_set = ['usergroup.read', 'usergroup.write', 'usergroup.admin']
 
         super(UserGroupList, self).__init__(
             obj_list=db_user_group_list,
             obj_attr='users_group_name', perm_set=perm_set,
             perm_checker=HasUserGroupPermissionAny,
             extra_kwargs=extra_kwargs)
 
 
 class ScmModel(BaseModel):
     """
     Generic Scm Model
     """
 
     @LazyProperty
     def repos_path(self):
         """
         Gets the repositories root path from database
         """
 
         settings_model = VcsSettingsModel(sa=self.sa)
         return settings_model.get_repos_location()
 
     def repo_scan(self, repos_path=None):
         """
         Listing of repositories in given path. This path should not be a
         repository itself. Return a dictionary of repository objects
 
         :param repos_path: path to directory containing repositories
         """
 
         if repos_path is None:
             repos_path = self.repos_path
 
         log.info('scanning for repositories in %s', repos_path)
 
         config = make_db_config()
         config.set('extensions', 'largefiles', '')
         repos = {}
 
         for name, path in get_filesystem_repos(repos_path, recursive=True):
             # name need to be decomposed and put back together using the /
             # since this is internal storage separator for rhodecode
             name = Repository.normalize_repo_name(name)
 
             try:
                 if name in repos:
                     raise RepositoryError('Duplicate repository name %s '
                                           'found in %s' % (name, path))
                 elif path[0] in rhodecode.BACKENDS:
                     backend = get_backend(path[0])
                     repos[name] = backend(path[1], config=config,
                                           with_wire={"cache": False})
             except OSError:
                 continue
             except RepositoryError:
                 log.exception('Failed to create a repo')
                 continue
 
         log.debug('found %s paths with repositories', len(repos))
         return repos
 
     def get_repos(self, all_repos=None, sort_key=None):
         """
         Get all repositories from db and for each repo create it's
         backend instance and fill that backed with information from database
 
         :param all_repos: list of repository names as strings
             give specific repositories list, good for filtering
 
         :param sort_key: initial sorting of repositories
         """
         if all_repos is None:
             all_repos = self.sa.query(Repository)\
                 .filter(Repository.group_id == None)\
                 .order_by(func.lower(Repository.repo_name)).all()
         repo_iter = SimpleCachedRepoList(
             all_repos, repos_path=self.repos_path, order_by=sort_key)
         return repo_iter
 
     def get_repo_groups(self, all_groups=None):
         if all_groups is None:
             all_groups = RepoGroup.query()\
                 .filter(RepoGroup.group_parent_id == None).all()
         return [x for x in RepoGroupList(all_groups)]
 
     def mark_for_invalidation(self, repo_name, delete=False):
         """
         Mark caches of this repo invalid in the database. `delete` flag
         removes the cache entries
 
         :param repo_name: the repo_name for which caches should be marked
             invalid, or deleted
         :param delete: delete the entry keys instead of setting bool
             flag on them, and also purge caches used by the dogpile
         """
         repo = Repository.get_by_repo_name(repo_name)
 
         if repo:
             invalidation_namespace = CacheKey.REPO_INVALIDATION_NAMESPACE.format(
                 repo_id=repo.repo_id)
             CacheKey.set_invalidate(invalidation_namespace, delete=delete)
 
             repo_id = repo.repo_id
             config = repo._config
             config.set('extensions', 'largefiles', '')
             repo.update_commit_cache(config=config, cs_cache=None)
             if delete:
                 cache_namespace_uid = 'cache_repo.{}'.format(repo_id)
-                rc_cache.clear_cache_namespace('cache_repo', cache_namespace_uid)
+                rc_cache.clear_cache_namespace(
+                    'cache_repo', cache_namespace_uid, invalidate=True)
 
     def toggle_following_repo(self, follow_repo_id, user_id):
 
         f = self.sa.query(UserFollowing)\
             .filter(UserFollowing.follows_repo_id == follow_repo_id)\
             .filter(UserFollowing.user_id == user_id).scalar()
 
         if f is not None:
             try:
                 self.sa.delete(f)
                 return
             except Exception:
                 log.error(traceback.format_exc())
                 raise
 
         try:
             f = UserFollowing()
             f.user_id = user_id
             f.follows_repo_id = follow_repo_id
             self.sa.add(f)
         except Exception:
             log.error(traceback.format_exc())
             raise
 
     def toggle_following_user(self, follow_user_id, user_id):
         f = self.sa.query(UserFollowing)\
             .filter(UserFollowing.follows_user_id == follow_user_id)\
             .filter(UserFollowing.user_id == user_id).scalar()
 
         if f is not None:
             try:
                 self.sa.delete(f)
                 return
             except Exception:
                 log.error(traceback.format_exc())
                 raise
 
         try:
             f = UserFollowing()
             f.user_id = user_id
             f.follows_user_id = follow_user_id
             self.sa.add(f)
         except Exception:
             log.error(traceback.format_exc())
             raise
 
     def is_following_repo(self, repo_name, user_id, cache=False):
         r = self.sa.query(Repository)\
             .filter(Repository.repo_name == repo_name).scalar()
 
         f = self.sa.query(UserFollowing)\
             .filter(UserFollowing.follows_repository == r)\
             .filter(UserFollowing.user_id == user_id).scalar()
 
         return f is not None
 
     def is_following_user(self, username, user_id, cache=False):
         u = User.get_by_username(username)
 
         f = self.sa.query(UserFollowing)\
             .filter(UserFollowing.follows_user == u)\
             .filter(UserFollowing.user_id == user_id).scalar()
 
         return f is not None
 
     def get_followers(self, repo):
         repo = self._get_repo(repo)
 
         return self.sa.query(UserFollowing)\
             .filter(UserFollowing.follows_repository == repo).count()
 
     def get_forks(self, repo):
         repo = self._get_repo(repo)
         return self.sa.query(Repository)\
             .filter(Repository.fork == repo).count()
 
     def get_pull_requests(self, repo):
         repo = self._get_repo(repo)
         return self.sa.query(PullRequest)\
             .filter(PullRequest.target_repo == repo)\
             .filter(PullRequest.status != PullRequest.STATUS_CLOSED).count()
 
     def get_artifacts(self, repo):
         repo = self._get_repo(repo)
         return self.sa.query(FileStore)\
             .filter(FileStore.repo == repo)\
             .filter(or_(FileStore.hidden == None, FileStore.hidden == false())).count()
 
     def mark_as_fork(self, repo, fork, user):
         repo = self._get_repo(repo)
         fork = self._get_repo(fork)
         if fork and repo.repo_id == fork.repo_id:
             raise Exception("Cannot set repository as fork of itself")
 
         if fork and repo.repo_type != fork.repo_type:
             raise RepositoryError(
                 "Cannot set repository as fork of repository with other type")
 
         repo.fork = fork
         self.sa.add(repo)
         return repo
 
     def pull_changes(self, repo, username, remote_uri=None, validate_uri=True):
         dbrepo = self._get_repo(repo)
         remote_uri = remote_uri or dbrepo.clone_uri
         if not remote_uri:
             raise Exception("This repository doesn't have a clone uri")
 
         repo = dbrepo.scm_instance(cache=False)
         repo.config.clear_section('hooks')
 
         try:
             # NOTE(marcink): add extra validation so we skip invalid urls
             # this is due this tasks can be executed via scheduler without
             # proper validation of remote_uri
             if validate_uri:
                 config = make_db_config(clear_session=False)
                 url_validator(remote_uri, dbrepo.repo_type, config)
         except InvalidCloneUrl:
             raise
 
         repo_name = dbrepo.repo_name
         try:
             # TODO: we need to make sure those operations call proper hooks !
             repo.fetch(remote_uri)
 
             self.mark_for_invalidation(repo_name)
         except Exception:
             log.error(traceback.format_exc())
             raise
 
     def push_changes(self, repo, username, remote_uri=None, validate_uri=True):
         dbrepo = self._get_repo(repo)
         remote_uri = remote_uri or dbrepo.push_uri
         if not remote_uri:
             raise Exception("This repository doesn't have a clone uri")
 
         repo = dbrepo.scm_instance(cache=False)
         repo.config.clear_section('hooks')
 
         try:
             # NOTE(marcink): add extra validation so we skip invalid urls
             # this is due this tasks can be executed via scheduler without
             # proper validation of remote_uri
             if validate_uri:
                 config = make_db_config(clear_session=False)
                 url_validator(remote_uri, dbrepo.repo_type, config)
         except InvalidCloneUrl:
             raise
 
         try:
             repo.push(remote_uri)
         except Exception:
             log.error(traceback.format_exc())
             raise
 
     def commit_change(self, repo, repo_name, commit, user, author, message,
                       content, f_path):
         """
         Commits changes
 
         :param repo: SCM instance
 
         """
         user = self._get_user(user)
 
         # decoding here will force that we have proper encoded values
         # in any other case this will throw exceptions and deny commit
         content = safe_str(content)
         path = safe_str(f_path)
         # message and author needs to be unicode
         # proper backend should then translate that into required type
         message = safe_unicode(message)
         author = safe_unicode(author)
         imc = repo.in_memory_commit
         imc.change(FileNode(path, content, mode=commit.get_file_mode(f_path)))
         try:
             # TODO: handle pre-push action !
             tip = imc.commit(
                 message=message, author=author, parents=[commit],
                 branch=commit.branch)
         except Exception as e:
             log.error(traceback.format_exc())
             raise IMCCommitError(str(e))
         finally:
             # always clear caches, if commit fails we want fresh object also
             self.mark_for_invalidation(repo_name)
 
         # We trigger the post-push action
         hooks_utils.trigger_post_push_hook(
             username=user.username, action='push_local', hook_type='post_push',
             repo_name=repo_name, repo_type=repo.alias, commit_ids=[tip.raw_id])
         return tip
 
     def _sanitize_path(self, f_path):
         if f_path.startswith('/') or f_path.startswith('./') or '../' in f_path:
             raise NonRelativePathError('%s is not an relative path' % f_path)
         if f_path:
             f_path = os.path.normpath(f_path)
         return f_path
 
     def get_dirnode_metadata(self, request, commit, dir_node):
         if not dir_node.is_dir():
             return []
 
         data = []
         for node in dir_node:
             if not node.is_file():
                 # we skip file-nodes
                 continue
 
             last_commit = node.last_commit
             last_commit_date = last_commit.date
             data.append({
                 'name': node.name,
                 'size': h.format_byte_size_binary(node.size),
                 'modified_at': h.format_date(last_commit_date),
                 'modified_ts': last_commit_date.isoformat(),
                 'revision': last_commit.revision,
                 'short_id': last_commit.short_id,
                 'message': h.escape(last_commit.message),
                 'author': h.escape(last_commit.author),
                 'user_profile': h.gravatar_with_user(
                     request, last_commit.author),
             })
 
         return data
 
     def get_nodes(self, repo_name, commit_id, root_path='/', flat=True,
                   extended_info=False, content=False, max_file_bytes=None):
         """
         recursive walk in root dir and return a set of all path in that dir
         based on repository walk function
 
         :param repo_name: name of repository
         :param commit_id: commit id for which to list nodes
         :param root_path: root path to list
         :param flat: return as a list, if False returns a dict with description
         :param extended_info: show additional info such as md5, binary, size etc
         :param content: add nodes content to the return data
         :param max_file_bytes: will not return file contents over this limit
 
         """
         _files = list()
         _dirs = list()
         try:
             _repo = self._get_repo(repo_name)
             commit = _repo.scm_instance().get_commit(commit_id=commit_id)
             root_path = root_path.lstrip('/')
             for __, dirs, files in commit.walk(root_path):
 
                 for f in files:
                     _content = None
                     _data = f_name = f.unicode_path
 
                     if not flat:
                         _data = {
                             "name": h.escape(f_name),
                             "type": "file",
                             }
                         if extended_info:
                             _data.update({
                                 "md5": f.md5,
                                 "binary": f.is_binary,
                                 "size": f.size,
                                 "extension": f.extension,
                                 "mimetype": f.mimetype,
                                 "lines": f.lines()[0]
                             })
 
                         if content:
                             over_size_limit = (max_file_bytes is not None
                                                and f.size > max_file_bytes)
                             full_content = None
                             if not f.is_binary and not over_size_limit:
                                 full_content = safe_str(f.content)
 
                             _data.update({
                                 "content": full_content,
                             })
                     _files.append(_data)
 
                 for d in dirs:
                     _data = d_name = d.unicode_path
                     if not flat:
                         _data = {
                             "name": h.escape(d_name),
                             "type": "dir",
                             }
                     if extended_info:
                         _data.update({
                             "md5": None,
                             "binary": None,
                             "size": None,
                             "extension": None,
                         })
                     if content:
                         _data.update({
                             "content": None
                         })
                     _dirs.append(_data)
         except RepositoryError:
             log.exception("Exception in get_nodes")
             raise
 
         return _dirs, _files
 
     def get_quick_filter_nodes(self, repo_name, commit_id, root_path='/'):
         """
         Generate files for quick filter in files view
         """
 
         _files = list()
         _dirs = list()
         try:
             _repo = self._get_repo(repo_name)
             commit = _repo.scm_instance().get_commit(commit_id=commit_id)
             root_path = root_path.lstrip('/')
             for __, dirs, files in commit.walk(root_path):
 
                 for f in files:
 
                     _data = {
                         "name": h.escape(f.unicode_path),
                         "type": "file",
                         }
 
                     _files.append(_data)
 
                 for d in dirs:
 
                     _data = {
                         "name": h.escape(d.unicode_path),
                         "type": "dir",
                         }
 
                     _dirs.append(_data)
         except RepositoryError:
             log.exception("Exception in get_quick_filter_nodes")
             raise
 
         return _dirs, _files
 
     def get_node(self, repo_name, commit_id, file_path,
                  extended_info=False, content=False, max_file_bytes=None, cache=True):
         """
         retrieve single node from commit
         """
         try:
 
             _repo = self._get_repo(repo_name)
             commit = _repo.scm_instance().get_commit(commit_id=commit_id)
 
             file_node = commit.get_node(file_path)
             if file_node.is_dir():
                 raise RepositoryError('The given path is a directory')
 
             _content = None
             f_name = file_node.unicode_path
 
             file_data = {
                 "name": h.escape(f_name),
                 "type": "file",
             }
 
             if extended_info:
                 file_data.update({
                     "extension": file_node.extension,
                     "mimetype": file_node.mimetype,
                 })
 
                 if cache:
                     md5 = file_node.md5
                     is_binary = file_node.is_binary
                     size = file_node.size
                 else:
                     is_binary, md5, size, _content = file_node.metadata_uncached()
 
                 file_data.update({
                     "md5": md5,
                     "binary": is_binary,
                     "size": size,
                 })
 
             if content and cache:
                 # get content + cache
                 size = file_node.size
                 over_size_limit = (max_file_bytes is not None and size > max_file_bytes)
                 full_content = None
                 all_lines = 0
                 if not file_node.is_binary and not over_size_limit:
                     full_content = safe_unicode(file_node.content)
                     all_lines, empty_lines = file_node.count_lines(full_content)
 
                 file_data.update({
                     "content": full_content,
                     "lines": all_lines
                 })
             elif content:
                 # get content *without* cache
                 if _content is None:
                     is_binary, md5, size, _content = file_node.metadata_uncached()
 
                 over_size_limit = (max_file_bytes is not None and size > max_file_bytes)
                 full_content = None
                 all_lines = 0
                 if not is_binary and not over_size_limit:
                     full_content = safe_unicode(_content)
                     all_lines, empty_lines = file_node.count_lines(full_content)
 
                 file_data.update({
                     "content": full_content,
                     "lines": all_lines
                 })
 
         except RepositoryError:
             log.exception("Exception in get_node")
             raise
 
         return file_data
 
     def get_fts_data(self, repo_name, commit_id, root_path='/'):
         """
         Fetch node tree for usage in full text search
         """
 
         tree_info = list()
 
         try:
             _repo = self._get_repo(repo_name)
             commit = _repo.scm_instance().get_commit(commit_id=commit_id)
             root_path = root_path.lstrip('/')
             for __, dirs, files in commit.walk(root_path):
 
                 for f in files:
                     is_binary, md5, size, _content = f.metadata_uncached()
                     _data = {
                         "name": f.unicode_path,
                         "md5": md5,
                         "extension": f.extension,
                         "binary": is_binary,
                         "size": size
                     }
 
                     tree_info.append(_data)
 
         except RepositoryError:
             log.exception("Exception in get_nodes")
             raise
 
         return tree_info
 
     def create_nodes(self, user, repo, message, nodes, parent_commit=None,
                      author=None, trigger_push_hook=True):
         """
         Commits given multiple nodes into repo
 
         :param user: RhodeCode User object or user_id, the commiter
         :param repo: RhodeCode Repository object
         :param message: commit message
         :param nodes: mapping {filename:{'content':content},...}
         :param parent_commit: parent commit, can be empty than it's
            initial commit
         :param author: author of commit, cna be different that commiter
            only for git
         :param trigger_push_hook: trigger push hooks
 
         :returns: new commited commit
         """
 
         user = self._get_user(user)
         scm_instance = repo.scm_instance(cache=False)
 
         processed_nodes = []
         for f_path in nodes:
             f_path = self._sanitize_path(f_path)
             content = nodes[f_path]['content']
             f_path = safe_str(f_path)
             # decoding here will force that we have proper encoded values
             # in any other case this will throw exceptions and deny commit
             if isinstance(content, (basestring,)):
                 content = safe_str(content)
             elif isinstance(content, (file, cStringIO.OutputType,)):
                 content = content.read()
             else:
                 raise Exception('Content is of unrecognized type %s' % (
                     type(content)
                 ))
             processed_nodes.append((f_path, content))
 
         message = safe_unicode(message)
         commiter = user.full_contact
         author = safe_unicode(author) if author else commiter
 
         imc = scm_instance.in_memory_commit
 
         if not parent_commit:
             parent_commit = EmptyCommit(alias=scm_instance.alias)
 
         if isinstance(parent_commit, EmptyCommit):
             # EmptyCommit means we we're editing empty repository
             parents = None
         else:
             parents = [parent_commit]
         # add multiple nodes
         for path, content in processed_nodes:
             imc.add(FileNode(path, content=content))
         # TODO: handle pre push scenario
         tip = imc.commit(message=message,
                          author=author,
                          parents=parents,
                          branch=parent_commit.branch)
 
         self.mark_for_invalidation(repo.repo_name)
         if trigger_push_hook:
             hooks_utils.trigger_post_push_hook(
                 username=user.username, action='push_local',
                 repo_name=repo.repo_name, repo_type=scm_instance.alias,
                 hook_type='post_push',
                 commit_ids=[tip.raw_id])
         return tip
 
     def update_nodes(self, user, repo, message, nodes, parent_commit=None,
                      author=None, trigger_push_hook=True):
         user = self._get_user(user)
         scm_instance = repo.scm_instance(cache=False)
 
         message = safe_unicode(message)
         commiter = user.full_contact
         author = safe_unicode(author) if author else commiter
 
         imc = scm_instance.in_memory_commit
 
         if not parent_commit:
             parent_commit = EmptyCommit(alias=scm_instance.alias)
 
         if isinstance(parent_commit, EmptyCommit):
             # EmptyCommit means we we're editing empty repository
             parents = None
         else:
             parents = [parent_commit]
 
         # add multiple nodes
         for _filename, data in nodes.items():
             # new filename, can be renamed from the old one, also sanitaze
             # the path for any hack around relative paths like ../../ etc.
             filename = self._sanitize_path(data['filename'])
             old_filename = self._sanitize_path(_filename)
             content = data['content']
             file_mode = data.get('mode')
             filenode = FileNode(old_filename, content=content, mode=file_mode)
             op = data['op']
             if op == 'add':
                 imc.add(filenode)
             elif op == 'del':
                 imc.remove(filenode)
             elif op == 'mod':
                 if filename != old_filename:
                     # TODO: handle renames more efficient, needs vcs lib changes
                     imc.remove(filenode)
                     imc.add(FileNode(filename, content=content, mode=file_mode))
                 else:
                     imc.change(filenode)
 
         try:
             # TODO: handle pre push scenario commit changes
             tip = imc.commit(message=message,
                              author=author,
                              parents=parents,
                              branch=parent_commit.branch)
         except NodeNotChangedError:
             raise
         except Exception as e:
             log.exception("Unexpected exception during call to imc.commit")
             raise IMCCommitError(str(e))
         finally:
             # always clear caches, if commit fails we want fresh object also
             self.mark_for_invalidation(repo.repo_name)
 
         if trigger_push_hook:
             hooks_utils.trigger_post_push_hook(
                 username=user.username, action='push_local', hook_type='post_push',
                 repo_name=repo.repo_name, repo_type=scm_instance.alias,
                 commit_ids=[tip.raw_id])
 
         return tip
 
     def delete_nodes(self, user, repo, message, nodes, parent_commit=None,
                      author=None, trigger_push_hook=True):
         """
         Deletes given multiple nodes into `repo`
 
         :param user: RhodeCode User object or user_id, the committer
         :param repo: RhodeCode Repository object
         :param message: commit message
         :param nodes: mapping {filename:{'content':content},...}
         :param parent_commit: parent commit, can be empty than it's initial
            commit
         :param author: author of commit, cna be different that commiter only
            for git
         :param trigger_push_hook: trigger push hooks
 
         :returns: new commit after deletion
         """
 
         user = self._get_user(user)
         scm_instance = repo.scm_instance(cache=False)
 
         processed_nodes = []
         for f_path in nodes:
             f_path = self._sanitize_path(f_path)
             # content can be empty but for compatabilty it allows same dicts
             # structure as add_nodes
             content = nodes[f_path].get('content')
             processed_nodes.append((f_path, content))
 
         message = safe_unicode(message)
         commiter = user.full_contact
         author = safe_unicode(author) if author else commiter
 
         imc = scm_instance.in_memory_commit
 
         if not parent_commit:
             parent_commit = EmptyCommit(alias=scm_instance.alias)
 
         if isinstance(parent_commit, EmptyCommit):
             # EmptyCommit means we we're editing empty repository
             parents = None
         else:
             parents = [parent_commit]
         # add multiple nodes
         for path, content in processed_nodes:
             imc.remove(FileNode(path, content=content))
 
         # TODO: handle pre push scenario
         tip = imc.commit(message=message,
                          author=author,
                          parents=parents,
                          branch=parent_commit.branch)
 
         self.mark_for_invalidation(repo.repo_name)
         if trigger_push_hook:
             hooks_utils.trigger_post_push_hook(
                 username=user.username, action='push_local', hook_type='post_push',
                 repo_name=repo.repo_name, repo_type=scm_instance.alias,
                 commit_ids=[tip.raw_id])
         return tip
 
     def strip(self, repo, commit_id, branch):
         scm_instance = repo.scm_instance(cache=False)
         scm_instance.config.clear_section('hooks')
         scm_instance.strip(commit_id, branch)
         self.mark_for_invalidation(repo.repo_name)
 
     def get_unread_journal(self):
         return self.sa.query(UserLog).count()
 
     @classmethod
     def backend_landing_ref(cls, repo_type):
         """
         Return a default landing ref based on a repository type.
         """
 
         landing_ref = {
             'hg': ('branch:default', 'default'),
             'git': ('branch:master', 'master'),
             'svn': ('rev:tip', 'latest tip'),
             'default': ('rev:tip', 'latest tip'),
         }
 
         return landing_ref.get(repo_type) or landing_ref['default']
 
     def get_repo_landing_revs(self, translator, repo=None):
         """
         Generates select option with tags branches and bookmarks (for hg only)
         grouped by type
 
         :param repo:
         """
         _ = translator
         repo = self._get_repo(repo)
 
         if repo:
             repo_type = repo.repo_type
         else:
             repo_type = 'default'
 
         default_landing_ref, landing_ref_lbl = self.backend_landing_ref(repo_type)
 
         default_ref_options = [
             [default_landing_ref, landing_ref_lbl]
         ]
         default_choices = [
             default_landing_ref
         ]
 
         if not repo:
             return default_choices, default_ref_options
 
         repo = repo.scm_instance()
 
         ref_options = [('rev:tip', 'latest tip')]
         choices = ['rev:tip']
 
         # branches
         branch_group = [(u'branch:%s' % safe_unicode(b), safe_unicode(b)) for b in repo.branches]
         if not branch_group:
             # new repo, or without maybe a branch?
             branch_group = default_ref_options
 
         branches_group = (branch_group, _("Branches"))
         ref_options.append(branches_group)
         choices.extend([x[0] for x in branches_group[0]])
 
         # bookmarks for HG
         if repo.alias == 'hg':
             bookmarks_group = (
                 [(u'book:%s' % safe_unicode(b), safe_unicode(b))
                     for b in repo.bookmarks],
                 _("Bookmarks"))
             ref_options.append(bookmarks_group)
             choices.extend([x[0] for x in bookmarks_group[0]])
 
         # tags
         tags_group = (
             [(u'tag:%s' % safe_unicode(t), safe_unicode(t))
                 for t in repo.tags],
             _("Tags"))
         ref_options.append(tags_group)
         choices.extend([x[0] for x in tags_group[0]])
 
         return choices, ref_options
 
     def get_server_info(self, environ=None):
         server_info = get_system_info(environ)
         return server_info