##// END OF EJS Templates
repair: migrate revlogs during upgrade...
repair: migrate revlogs during upgrade Our next step for in-place upgrade is to migrate store data. Revlogs are the biggest source of data within the store and a store is useless without them, so we implement their migration first. Our strategy for migrating revlogs is to walk the store and call `revlog.clone()` on each revlog. There are some minor complications. Because revlogs have different storage options (e.g. changelog has generaldelta and delta chains disabled), we need to obtain the correct class of revlog so inserted data is encoded properly for its type. Various attempts at implementing progress indicators that didn't lead to frustration from false "it's almost done" indicators were made. I initially used a single progress bar based on number of revlogs. However, this quickly churned through all filelogs, got to 99% then effectively froze at 99.99% when it got to the manifest. So I converted the progress bar to total revision count. This was a little bit better. But the manifest was still significantly slower than filelogs and it took forever to process the last few percent. I then tried both revision/chunk bytes and raw bytes as the denominator. This had the opposite effect: because so much data is in manifests, it would churn through filelogs without showing much progress. When it got to manifests, it would fill in 90+% of the progress bar. I finally gave up having a unified progress bar and instead implemented 3 progress bars: 1 for filelog revisions, 1 for manifest revisions, and 1 for changelog revisions. I added extra messages indicating the total number of revisions of each so users know there are more progress bars coming. I also added extra messages before and after each stage to give extra details about what is happening. Strictly speaking, this isn't necessary. But the numbers are impressive. For example, when converting a non-generaldelta mozilla-central repository, the messages you see are: migrating 2475593 total revisions (1833043 in filelogs, 321156 in manifests, 321394 in changelog) migrating 1.67 GB in store; 2508 GB tracked data migrating 267868 filelogs containing 1833043 revisions (1.09 GB in store; 57.3 GB tracked data) finished migrating 1833043 filelog revisions across 267868 filelogs; change in size: -415776 bytes migrating 1 manifests containing 321156 revisions (518 MB in store; 2451 GB tracked data) That "2508 GB" figure really blew me away. I had no clue that the raw tracked data in mozilla-central was that large. Granted, 2451 GB is in the manifest and "only" 57.3 GB is in filelogs. But still. It's worth noting that gratuitous loading of source revlogs in order to display numbers and progress bars does serve a purpose: it ensures we can open all source revlogs. We don't want to spend several minutes copying revlogs only to encounter a permissions error or similar later. As part of this commit, we also add swapping of the store directory to the upgrade function. After revlogs are converted, we move the old store into the backup directory then move the temporary repo's store into the old store's location. On well-behaved systems, this should be 2 atomic operations and the window of inconsistency show be very narrow. There are still a few improvements to be made to store copying and upgrading. But this commit gets the bulk of the work out of the way.

File last commit:

r30743:2df98312 default
r30779:38aa1ca9 default
Show More
unionrepo.py
260 lines | 9.0 KiB | text/x-python | PythonLexer
# unionrepo.py - repository class for viewing union of repository changesets
#
# Derived from bundlerepo.py
# Copyright 2006, 2007 Benoit Boissinot <bboissin@gmail.com>
# Copyright 2013 Unity Technologies, Mads Kiilerich <madski@unity3d.com>
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
"""Repository class for "in-memory pull" of one local repository to another,
allowing operations like diff and log with revsets.
"""
from __future__ import absolute_import
from .i18n import _
from .node import nullid
from . import (
changelog,
cmdutil,
error,
filelog,
localrepo,
manifest,
mdiff,
pathutil,
pycompat,
revlog,
scmutil,
util,
)
class unionrevlog(revlog.revlog):
def __init__(self, opener, indexfile, revlog2, linkmapper):
# How it works:
# To retrieve a revision, we just need to know the node id so we can
# look it up in revlog2.
#
# To differentiate a rev in the second revlog from a rev in the revlog,
# we check revision against repotiprev.
opener = scmutil.readonlyvfs(opener)
revlog.revlog.__init__(self, opener, indexfile)
self.revlog2 = revlog2
n = len(self)
self.repotiprev = n - 1
self.bundlerevs = set() # used by 'bundle()' revset expression
for rev2 in self.revlog2:
rev = self.revlog2.index[rev2]
# rev numbers - in revlog2, very different from self.rev
_start, _csize, _rsize, base, linkrev, p1rev, p2rev, node = rev
flags = _start & 0xFFFF
if linkmapper is None: # link is to same revlog
assert linkrev == rev2 # we never link back
link = n
else: # rev must be mapped from repo2 cl to unified cl by linkmapper
link = linkmapper(linkrev)
if linkmapper is not None: # link is to same revlog
base = linkmapper(base)
if node in self.nodemap:
# this happens for the common revlog revisions
self.bundlerevs.add(self.nodemap[node])
continue
p1node = self.revlog2.node(p1rev)
p2node = self.revlog2.node(p2rev)
e = (flags, None, None, base,
link, self.rev(p1node), self.rev(p2node), node)
self.index.insert(-1, e)
self.nodemap[node] = n
self.bundlerevs.add(n)
n += 1
def _chunk(self, rev):
if rev <= self.repotiprev:
return revlog.revlog._chunk(self, rev)
return self.revlog2._chunk(self.node(rev))
def revdiff(self, rev1, rev2):
"""return or calculate a delta between two revisions"""
if rev1 > self.repotiprev and rev2 > self.repotiprev:
return self.revlog2.revdiff(
self.revlog2.rev(self.node(rev1)),
self.revlog2.rev(self.node(rev2)))
elif rev1 <= self.repotiprev and rev2 <= self.repotiprev:
return self.baserevdiff(rev1, rev2)
return mdiff.textdiff(self.revision(self.node(rev1)),
self.revision(self.node(rev2)))
def revision(self, nodeorrev, raw=False):
"""return an uncompressed revision of a given node or revision
number.
"""
if isinstance(nodeorrev, int):
rev = nodeorrev
node = self.node(rev)
else:
node = nodeorrev
rev = self.rev(node)
if node == nullid:
return ""
if rev > self.repotiprev:
text = self.revlog2.revision(node)
self._cache = (node, rev, text)
else:
text = self.baserevision(rev)
# already cached
return text
def baserevision(self, nodeorrev):
# Revlog subclasses may override 'revision' method to modify format of
# content retrieved from revlog. To use unionrevlog with such class one
# needs to override 'baserevision' and make more specific call here.
return revlog.revlog.revision(self, nodeorrev)
def baserevdiff(self, rev1, rev2):
# Exists for the same purpose as baserevision.
return revlog.revlog.revdiff(self, rev1, rev2)
def addrevision(self, text, transaction, link, p1=None, p2=None, d=None):
raise NotImplementedError
def addgroup(self, revs, linkmapper, transaction):
raise NotImplementedError
def strip(self, rev, minlink):
raise NotImplementedError
def checksize(self):
raise NotImplementedError
class unionchangelog(unionrevlog, changelog.changelog):
def __init__(self, opener, opener2):
changelog.changelog.__init__(self, opener)
linkmapper = None
changelog2 = changelog.changelog(opener2)
unionrevlog.__init__(self, opener, self.indexfile, changelog2,
linkmapper)
def baserevision(self, nodeorrev):
# Although changelog doesn't override 'revision' method, some extensions
# may replace this class with another that does. Same story with
# manifest and filelog classes.
return changelog.changelog.revision(self, nodeorrev)
def baserevdiff(self, rev1, rev2):
return changelog.changelog.revdiff(self, rev1, rev2)
class unionmanifest(unionrevlog, manifest.manifestrevlog):
def __init__(self, opener, opener2, linkmapper):
manifest.manifestrevlog.__init__(self, opener)
manifest2 = manifest.manifestrevlog(opener2)
unionrevlog.__init__(self, opener, self.indexfile, manifest2,
linkmapper)
def baserevision(self, nodeorrev):
return manifest.manifestrevlog.revision(self, nodeorrev)
def baserevdiff(self, rev1, rev2):
return manifest.manifestrevlog.revdiff(self, rev1, rev2)
class unionfilelog(unionrevlog, filelog.filelog):
def __init__(self, opener, path, opener2, linkmapper, repo):
filelog.filelog.__init__(self, opener, path)
filelog2 = filelog.filelog(opener2, path)
unionrevlog.__init__(self, opener, self.indexfile, filelog2,
linkmapper)
self._repo = repo
def baserevision(self, nodeorrev):
return filelog.filelog.revision(self, nodeorrev)
def baserevdiff(self, rev1, rev2):
return filelog.filelog.revdiff(self, rev1, rev2)
def iscensored(self, rev):
"""Check if a revision is censored."""
if rev <= self.repotiprev:
return filelog.filelog.iscensored(self, rev)
node = self.node(rev)
return self.revlog2.iscensored(self.revlog2.rev(node))
class unionpeer(localrepo.localpeer):
def canpush(self):
return False
class unionrepository(localrepo.localrepository):
def __init__(self, ui, path, path2):
localrepo.localrepository.__init__(self, ui, path)
self.ui.setconfig('phases', 'publish', False, 'unionrepo')
self._url = 'union:%s+%s' % (util.expandpath(path),
util.expandpath(path2))
self.repo2 = localrepo.localrepository(ui, path2)
@localrepo.unfilteredpropertycache
def changelog(self):
return unionchangelog(self.svfs, self.repo2.svfs)
def _clrev(self, rev2):
"""map from repo2 changelog rev to temporary rev in self.changelog"""
node = self.repo2.changelog.node(rev2)
return self.changelog.rev(node)
def _constructmanifest(self):
return unionmanifest(self.svfs, self.repo2.svfs,
self.unfiltered()._clrev)
def url(self):
return self._url
def file(self, f):
return unionfilelog(self.svfs, f, self.repo2.svfs,
self.unfiltered()._clrev, self)
def close(self):
self.repo2.close()
def cancopy(self):
return False
def peer(self):
return unionpeer(self)
def getcwd(self):
return pycompat.getcwd() # always outside the repo
def instance(ui, path, create):
if create:
raise error.Abort(_('cannot create new union repository'))
parentpath = ui.config("bundle", "mainreporoot", "")
if not parentpath:
# try to find the correct path to the working directory repo
parentpath = cmdutil.findrepo(pycompat.getcwd())
if parentpath is None:
parentpath = ''
if parentpath:
# Try to make the full path relative so we get a nice, short URL.
# In particular, we don't want temp dir names in test outputs.
cwd = pycompat.getcwd()
if parentpath == cwd:
parentpath = ''
else:
cwd = pathutil.normasprefix(cwd)
if parentpath.startswith(cwd):
parentpath = parentpath[len(cwd):]
if path.startswith('union:'):
s = path.split(":", 1)[1].split("+", 1)
if len(s) == 1:
repopath, repopath2 = parentpath, s[0]
else:
repopath, repopath2 = s
else:
repopath, repopath2 = parentpath, path
return unionrepository(ui, repopath, repopath2)