##// END OF EJS Templates
compression: introduce a `storage.revlog.zlib.level` configuration...
compression: introduce a `storage.revlog.zlib.level` configuration This option control the zlib compression level used when compression revlog chunk. This is also a good excuse to pave the way for a similar configuration option for the zstd compression engine. Having a dedicated option for each compression algorithm is useful because they don't support the same range of values. Using a higher zlib compression impact CPU consumption at compression time, but does not directly affected decompression time. However dealing with small compressed chunk can directly help decompression and indirectly help other revlog logic. I ran some basic test on repositories using different level. I am using the mercurial, pypy, netbeans and mozilla-central clone from our benchmark suite. All tested repository use sparse-revlog and got all their delta recomputed. The different compression level has a small effect on the repository size (about 10% variation in the total range). My quick analysis is that revlog mostly store small delta, that are not affected by the compression level much. So the variation probably mostly comes from better compression of the snapshots revisions, and snapshot revision only represent a small portion of the repository content. I also made some basic timings measurements. The "read" timings are gathered using simple run of `hg perfrevlogrevisions`, the "write" timings using `hg perfrevlogwrite` (restricted to the last 5000 revisions for netbeans and mozilla central). The timings are gathered on a generic machine, (not one of our performance locked machine), so small variation might not be meaningful. However large trend remains relevant. Keep in mind that these numbers are not pure compression/decompression time. They also involve the full revlog logic. In particular the difference in chunk size has an impact on the delta chain structure, affecting performance when writing or reading them. On read/write performance, the compression level has a bigger impact. Counter-intuitively, the higher compression levels improve "write" performance for the large repositories in our tested setting. Maybe because the last 5000 delta chain end up having a very different shape in this specific spot? Or maybe because of a more general trend of better delta chains thanks to the smaller chunk and snapshot. This series does not intend to change the default compression level. However, these result call for a deeper analysis of this performance difference in the future. Full data ========= repo level .hg/store size 00manifest.d read write ---------------------------------------------------------------- mercurial 1 49,402,813 5,963,475 0.170159 53.250304 mercurial 6 47,197,397 5,875,730 0.182820 56.264320 mercurial 9 47,121,596 5,849,781 0.189219 56.293612 pypy 1 370,830,572 28,462,425 2.679217 460.721984 pypy 6 340,112,317 27,648,747 2.768691 467.537158 pypy 9 338,360,736 27,639,003 2.763495 476.589918 netbeans 1 1,281,847,810 165,495,457 122.477027 520.560316 netbeans 6 1,205,284,353 159,161,207 139.876147 715.930400 netbeans 9 1,197,135,671 155,034,586 141.620281 678.297064 mozilla 1 2,775,497,186 298,527,987 147.867662 751.263721 mozilla 6 2,596,856,420 286,597,671 170.572118 987.056093 mozilla 9 2,587,542,494 287,018,264 163.622338 739.803002

File last commit:

r41825:0531dff7 default
r42210:1fac9b93 default
Show More
shallowrepo.py
305 lines | 11.4 KiB | text/x-python | PythonLexer
# shallowrepo.py - shallow repository that uses remote filelogs
#
# Copyright 2013 Facebook, Inc.
#
# This software may be used and distributed according to the terms of the
# GNU General Public License version 2 or any later version.
from __future__ import absolute_import
import os
from mercurial.i18n import _
from mercurial.node import hex, nullid, nullrev
from mercurial import (
encoding,
error,
localrepo,
match,
scmutil,
sparse,
util,
)
from mercurial.utils import procutil
from . import (
connectionpool,
constants,
contentstore,
datapack,
fileserverclient,
historypack,
metadatastore,
remotefilectx,
remotefilelog,
shallowutil,
)
if util.safehasattr(util, '_hgexecutable'):
# Before 5be286db
_hgexecutable = util.hgexecutable
else:
from mercurial.utils import procutil
_hgexecutable = procutil.hgexecutable
# These make*stores functions are global so that other extensions can replace
# them.
def makelocalstores(repo):
"""In-repo stores, like .hg/store/data; can not be discarded."""
localpath = os.path.join(repo.svfs.vfs.base, 'data')
if not os.path.exists(localpath):
os.makedirs(localpath)
# Instantiate local data stores
localcontent = contentstore.remotefilelogcontentstore(
repo, localpath, repo.name, shared=False)
localmetadata = metadatastore.remotefilelogmetadatastore(
repo, localpath, repo.name, shared=False)
return localcontent, localmetadata
def makecachestores(repo):
"""Typically machine-wide, cache of remote data; can be discarded."""
# Instantiate shared cache stores
cachepath = shallowutil.getcachepath(repo.ui)
cachecontent = contentstore.remotefilelogcontentstore(
repo, cachepath, repo.name, shared=True)
cachemetadata = metadatastore.remotefilelogmetadatastore(
repo, cachepath, repo.name, shared=True)
repo.sharedstore = cachecontent
repo.shareddatastores.append(cachecontent)
repo.sharedhistorystores.append(cachemetadata)
return cachecontent, cachemetadata
def makeremotestores(repo, cachecontent, cachemetadata):
"""These stores fetch data from a remote server."""
# Instantiate remote stores
repo.fileservice = fileserverclient.fileserverclient(repo)
remotecontent = contentstore.remotecontentstore(
repo.ui, repo.fileservice, cachecontent)
remotemetadata = metadatastore.remotemetadatastore(
repo.ui, repo.fileservice, cachemetadata)
return remotecontent, remotemetadata
def makepackstores(repo):
"""Packs are more efficient (to read from) cache stores."""
# Instantiate pack stores
packpath = shallowutil.getcachepackpath(repo,
constants.FILEPACK_CATEGORY)
packcontentstore = datapack.datapackstore(repo.ui, packpath)
packmetadatastore = historypack.historypackstore(repo.ui, packpath)
repo.shareddatastores.append(packcontentstore)
repo.sharedhistorystores.append(packmetadatastore)
shallowutil.reportpackmetrics(repo.ui, 'filestore', packcontentstore,
packmetadatastore)
return packcontentstore, packmetadatastore
def makeunionstores(repo):
"""Union stores iterate the other stores and return the first result."""
repo.shareddatastores = []
repo.sharedhistorystores = []
packcontentstore, packmetadatastore = makepackstores(repo)
cachecontent, cachemetadata = makecachestores(repo)
localcontent, localmetadata = makelocalstores(repo)
remotecontent, remotemetadata = makeremotestores(repo, cachecontent,
cachemetadata)
# Instantiate union stores
repo.contentstore = contentstore.unioncontentstore(
packcontentstore, cachecontent,
localcontent, remotecontent, writestore=localcontent)
repo.metadatastore = metadatastore.unionmetadatastore(
packmetadatastore, cachemetadata, localmetadata, remotemetadata,
writestore=localmetadata)
fileservicedatawrite = cachecontent
fileservicehistorywrite = cachemetadata
repo.fileservice.setstore(repo.contentstore, repo.metadatastore,
fileservicedatawrite, fileservicehistorywrite)
shallowutil.reportpackmetrics(repo.ui, 'filestore',
packcontentstore, packmetadatastore)
def wraprepo(repo):
class shallowrepository(repo.__class__):
@util.propertycache
def name(self):
return self.ui.config('remotefilelog', 'reponame')
@util.propertycache
def fallbackpath(self):
path = repo.ui.config("remotefilelog", "fallbackpath",
repo.ui.config('paths', 'default'))
if not path:
raise error.Abort("no remotefilelog server "
"configured - is your .hg/hgrc trusted?")
return path
def maybesparsematch(self, *revs, **kwargs):
'''
A wrapper that allows the remotefilelog to invoke sparsematch() if
this is a sparse repository, or returns None if this is not a
sparse repository.
'''
if revs:
ret = sparse.matcher(repo, revs=revs)
else:
ret = sparse.matcher(repo)
if ret.always():
return None
return ret
def file(self, f):
if f[0] == '/':
f = f[1:]
if self.shallowmatch(f):
return remotefilelog.remotefilelog(self.svfs, f, self)
else:
return super(shallowrepository, self).file(f)
def filectx(self, path, *args, **kwargs):
if self.shallowmatch(path):
return remotefilectx.remotefilectx(self, path, *args, **kwargs)
else:
return super(shallowrepository, self).filectx(path, *args,
**kwargs)
@localrepo.unfilteredmethod
def commitctx(self, ctx, error=False):
"""Add a new revision to current repository.
Revision information is passed via the context argument.
"""
# some contexts already have manifest nodes, they don't need any
# prefetching (for example if we're just editing a commit message
# we can reuse manifest
if not ctx.manifestnode():
# prefetch files that will likely be compared
m1 = ctx.p1().manifest()
files = []
for f in ctx.modified() + ctx.added():
fparent1 = m1.get(f, nullid)
if fparent1 != nullid:
files.append((f, hex(fparent1)))
self.fileservice.prefetch(files)
return super(shallowrepository, self).commitctx(ctx,
error=error)
def backgroundprefetch(self, revs, base=None, repack=False, pats=None,
opts=None):
"""Runs prefetch in background with optional repack
"""
cmd = [_hgexecutable(), '-R', repo.origroot, 'prefetch']
if repack:
cmd.append('--repack')
if revs:
cmd += ['-r', revs]
procutil.runbgcommand(cmd, encoding.environ)
def prefetch(self, revs, base=None, pats=None, opts=None):
"""Prefetches all the necessary file revisions for the given revs
Optionally runs repack in background
"""
with repo._lock(repo.svfs, 'prefetchlock', True, None, None,
_('prefetching in %s') % repo.origroot):
self._prefetch(revs, base, pats, opts)
def _prefetch(self, revs, base=None, pats=None, opts=None):
fallbackpath = self.fallbackpath
if fallbackpath:
# If we know a rev is on the server, we should fetch the server
# version of those files, since our local file versions might
# become obsolete if the local commits are stripped.
localrevs = repo.revs('outgoing(%s)', fallbackpath)
if base is not None and base != nullrev:
serverbase = list(repo.revs('first(reverse(::%s) - %ld)',
base, localrevs))
if serverbase:
base = serverbase[0]
else:
localrevs = repo
mfl = repo.manifestlog
mfrevlog = mfl.getstorage('')
if base is not None:
mfdict = mfl[repo[base].manifestnode()].read()
skip = set(mfdict.iteritems())
else:
skip = set()
# Copy the skip set to start large and avoid constant resizing,
# and since it's likely to be very similar to the prefetch set.
files = skip.copy()
serverfiles = skip.copy()
visited = set()
visited.add(nullrev)
revcount = len(revs)
progress = self.ui.makeprogress(_('prefetching'), total=revcount)
progress.update(0)
for rev in sorted(revs):
ctx = repo[rev]
if pats:
m = scmutil.match(ctx, pats, opts)
sparsematch = repo.maybesparsematch(rev)
mfnode = ctx.manifestnode()
mfrev = mfrevlog.rev(mfnode)
# Decompressing manifests is expensive.
# When possible, only read the deltas.
p1, p2 = mfrevlog.parentrevs(mfrev)
if p1 in visited and p2 in visited:
mfdict = mfl[mfnode].readfast()
else:
mfdict = mfl[mfnode].read()
diff = mfdict.iteritems()
if pats:
diff = (pf for pf in diff if m(pf[0]))
if sparsematch:
diff = (pf for pf in diff if sparsematch(pf[0]))
if rev not in localrevs:
serverfiles.update(diff)
else:
files.update(diff)
visited.add(mfrev)
progress.increment()
files.difference_update(skip)
serverfiles.difference_update(skip)
progress.complete()
# Fetch files known to be on the server
if serverfiles:
results = [(path, hex(fnode)) for (path, fnode) in serverfiles]
repo.fileservice.prefetch(results, force=True)
# Fetch files that may or may not be on the server
if files:
results = [(path, hex(fnode)) for (path, fnode) in files]
repo.fileservice.prefetch(results)
def close(self):
super(shallowrepository, self).close()
self.connectionpool.close()
repo.__class__ = shallowrepository
repo.shallowmatch = match.always()
makeunionstores(repo)
repo.includepattern = repo.ui.configlist("remotefilelog", "includepattern",
None)
repo.excludepattern = repo.ui.configlist("remotefilelog", "excludepattern",
None)
if not util.safehasattr(repo, 'connectionpool'):
repo.connectionpool = connectionpool.connectionpool(repo)
if repo.includepattern or repo.excludepattern:
repo.shallowmatch = match.match(repo.root, '', None,
repo.includepattern, repo.excludepattern)