Mercurial > hg
view mercurial/upgrade.py @ 40918:3764330f76a6
sparse-revlog: enabled by default
The feature provides large benefits. It now seems mature enough to be enabled
by default.
* It solves catastrophic issues regarding delta storage in revlog,
* It allows for shorter delta chain in all repositories, improving
performances.
Running benchmark of a wide range of operation did not reveal problematic
impact. Performance gains are observed where expected.
The format is supported by Mercurial version 4.7. So it seems safe to enable
it by default now.
Here is a reminder of key numbers regarding this delta strategy effect on
repository size and performance.
Effect on Size:
===============
For repositories with a lot of branches, sparse-revlog significantly improve
size, fixing limitation associated with the span of a delta chain. In
addition, sparse-revlog, deal well with limitations of the delta chain length.
For large repositories, this allows for a stiff reduction of the delta chain
without a problematic impact on the repository size. This delta chain length
improvement helps all repositories, not just the ones with many branches.
As a reminder, here are the default chain limits for each "format":
* no-sparse: none
* sparse: 1000
Mercurial
---------
Manifest Size:
limit | none | 1000
------------|-------------|------------
no-sparse | 6 143 044 | 6 269 496
sparse | 5 798 796 | 5 827 025
Manifest Chain length data
limit || none || 1000
value || average | max || average | max
------------||---------|---------||---------|---------
no-sparse || 429 | 1 397 || 397 | 1 000
sparse || 326 | 1 290 || 313 | 1 000
Full Store Size
limit | none | 1000
------------|-------------|------------
no-sparse | 46 944 775 | 47 166 129
sparse | 46 622 445 | 46 723 774
pypy
----
Manifest Size:
limit | none | 1000
------------|-------------|------------
no-sparse | 52 941 760 | 56 200 970
sparse | 26 348 229 | 27 384 133
Manifest Chain length data
limit || none || 1000
value || average | max || average | max
------------||---------|---------||---------|---------
no-sparse || 769 | 3 889 || 390 | 1 000
sparse || 1 223 | 3 846 || 495 | 1 000
Full Store Size
limit | none | 1000
------------|-------------|------------
no-sparse | 336 050 203 | 339 309 413
sparse | 338 673 985 | 339 709 889
Mozilla
-------
Manifest Size:
limit | none | 1000
------------|----------------|---------------
no-sparse | 215 096 339 | 1 708 853 525
sparse | 188 947 271 | 278 894 170
Manifest Chain length data
limit || none || 1000
value || average | max || average | max
------------||---------|---------||---------|--------
no-sparse || 20 454 | 59 562 || 491 | 1 000
sparse || 23 509 | 69 891 || 489 | 1 000
Full Store Size
limit | none | 1000
------------|----------------|---------------
no-sparse | 2 377 578 715 | 3 876 258 798
sparse | 2 441 677 137 | 2 535 997 381
Netbeans
--------
Manifest Size:
limit | none | 1000
------------|----------------|---------------
no-sparse | 130 088 982 | 741 590 565
sparse | 118 836 887 | 159 161 207
Manifest Chain length data
limit || none || 1000
value || average | max || average | max
------------||---------|---------||---------|---------
no-sparse || 19 321 | 61 397 || 510 | 1 000
sparse || 21 240 | 61 583 || 503 | 1 000
Full Store Size
limit | none | 1000
------------|----------------|---------------
no-sparse | 1 160 013 008 | 1 771 514 591
sparse | 1 164 959 988 | 1 205 284 308
Private repo #1
---------------
Manifest Size:
limit | none | 1000
------------|-----------------|---------------
no-sparse | 33 725 285 081 | 33 724 834 190
sparse | 350 542 420 | 423 470 579
Manifest Chain length data
limit || none || 1000
value || average | max || average | max
------------||---------|---------||---------|---------
no-sparse || 282 | 8 885 || 113 | 1 000
snapshot || 3 655 | 8 951 || 530 | 1 000
Full Store Size
limit | none | 1000
------------|----------------|---------------
no-sparse | 41 544 149 652 | 41 543 698 761
sparse | 8 448 037 300 | 8 520 965 459
Effect on speed:
================
Performances are strongly impacted by the delta chain length. Longer chain
results in slower revision restoration. For this reason, the 1000 chain limit
introduced by sparse-revlog helps repository with previously large chains a
lot. In our corpus, this means `netbeans` and `mozilla-central` who suffered
from unreasonable manifest delta chain length.
Another way sparse revlog helps, is by producing better delta's. For
repositories with many branches, the pathological patterns that resulted in
many sub-optimal deltas are gone. Smaller delta help with operations where
deltas are directly relevant, like bundle.
However, the sparse-revlog logic introduces some extra processing and a more
throughout testing of possible delta candidates. Adding an extra cost in some
cases. This cost is usually counterbalanced by the other performance gain.
However, for smaller repositories not affected by delta chain length issues or
branching related issues, this might make things a bit slower. However, these
are also repository where revlog performance is dwarfed by other costs.
Below are the summary of some timing from the performance test suite running
at `http://perf.octobus.net/` for a handful of key commands and operation.
It is important to keep in mind that most of this command works on the tip
part of the repository. The non-sparse and sparse version produce different
delta chains and the tip revision can end up at an arbitrary point of these
chains. This will impact some performance number listed in this summary.
For the record: here is the delta chain length for the tip revision of
manifest log in the benchmarked repository:
| no-sparse | sparse |
mercurial | 94 | 904 |
pypy | 23 | 673 |
netbeans | 4158 | 258 |
mozilla | 63263 | 781 |
As you can see, the chain length for mercurial and pypy turn out to be
significantly longer. The netbeans and mozilla one get shorter because these
repositories benefit from the maximum chain length.
Timing for `hg commit`:
-----------------------
The time taken by `hg commit` does not varies significantly, no drawback for
using sparse here.
| no-sparse | sparse |
mercurial | 68.1ms | 66.7ms |
pypy | 95.0ms | 94.1ms |
netbeans | 614.0ms | 611.0ms |
mozilla | 1340.0ms | 1.320.0ms |
Check the final section for statistics on a wider array of write.
Timing for bundling 10 000 changesets
-------------------------------------
The repository that benefits from better delta see a good performance boost.
The other ones are not significantly affected.
| no-sparse | sparse |
mercurial | 3.1s | 3.0s |
pypy | 25.1s | 7.5s |
netbeans | 24.2s | 17.0s |
mozilla | 23.7s | 25.0s |
Timing for unbundling 1 000 changesets
--------------------------------------
Mercurial and mozilla are unaffected. The pypy repository benefit well from
the better delta.
However, the netbeans repository takes a visible hit. Digging that difference
reveals that it comes from the sparse-revlog bundle having to deal with a
snapshot that was re-encoded in the bundle. The slow path for adding new a revision
had to be triggered for it, slowing things down. The Sparse versions do not have
such snapshot to handle similar cases in the tested configuration.
| no-sparse | sparse |
mercurial | 519ms | 502ms |
pypy | 1.270ms | 886ms |
netbeans | 1.370ms | 2.250ms |
mozilla | 3.230ms | 3.210ms |
Netbeans benefits from the better deltas in other dimensions too. For
example, the produced bundle is significantly smaller:
* netbeans-no-sparse.hg: 2.3MB
* netbeans-sparse.hg: 1.9MB
Timing to restore the tip most manifest entry:
----------------------------------------------
Nothing surprising here. The timing for mercurial and pypy are within a small
range where they won't affect performance much. In our tested case, they are
slower as they use a longer chain.
Timing for netbeans and mozilla improves a lot. Removing a significant amount
of time.
| no-sparse | sparse |
mercurial | 1.09ms | 3.15ms |
pypy | 4.11ms | 10.70ms |
netbeans | 239.00ms | 112.00ms |
mozilla | 688.00ms | 198.00ms |
Reading 100 revision in descending order:
-----------------------------------------
We see the same kind of effect when reading the last 100 revisions. Large
boost for netbeans and mozilla, as they use much smaller delta chain.
Mercurial and pypy longer chain means slower reads, but nothing gets out of
control.
| no-sparse | sparse |
mercurial | 0.089s | 0.268s |
pypy | 0.259s | 0.698s |
netbeans | 125.000s | 20.600s |
mozilla | 23.000s | 11.400s |
Writing from full text: statistic for the last 30K revisions
------------------------------------------------------------
This benchmark adds revisions to revlog from their full text. This is similar
to the work done during a commit, but for a large amount of revisions so that
we get a more relevant view.
We see better overall performances with sparse-revlog. The very worst case is
usually slower with sparse-revlog, but does not gets out of control. For the
vast majorities of the other writes, sparse-revlog is significantly faster for
larger repositories. This is reflected in the accumulated rewrite time for
netbeans and mozilla.
The notable exception is the pypy repository where things get slower. The
extra processing is not balanced by shorter delta chain. However, this is to
be seen as a blocking issue. First, the overall time spend dealing with revlog
for the repository pypy size is small compared to the other costs, so we get
slower on operations that matter less than for other larger repository.
Second, we still get nice size benefit from using sparse-revlog, smaller repo
size brings other usability and speed benefit (eg: bundle size).
max time | no-sparse | sparse |
mercurial | 0.010143s | 0.045280s |
pypy | 0.034924s | 0.243288s |
netbeans | 0.605371s | 2.130876s |
mozilla | 1.478342s | 3.424541s |
99% time | no-sparse | sparse |
mercurial | 0.003774s | 0.003758s |
pypy | 0.017387s | 0.025310s |
netbeans | 0.576913s | 0.271195s |
mozilla | 1.478342s | 0.449661s |
95% time | no-sparse | sparse |
mercurial | 0.002069s | 0.002120s |
pypy | 0.010141s | 0.014797s |
netbeans | 0.540202s | 0.258644s |
mozilla | 0.654830s | 0.243440s |
full time | no-sparse | sparse |
mercurial | 14.15s | 14.87s |
pypy | 90.50s | 137.12s |
netbeans | 6401.06s | 3411.14s |
mozilla | 3086.89s | 1991.97s |
Differential Revision: https://phab.mercurial-scm.org/D5345
author | Boris Feld <boris.feld@octobus.net> |
---|---|
date | Mon, 12 Nov 2018 01:22:38 +0100 |
parents | 64051af15596 |
children | 038108a9811c |
line wrap: on
line source
# upgrade.py - functions for in place upgrade of Mercurial repository # # Copyright (c) 2016-present, Gregory Szorc # # 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 stat from .i18n import _ from . import ( changelog, error, filelog, hg, localrepo, manifest, pycompat, revlog, scmutil, util, vfs as vfsmod, ) def requiredsourcerequirements(repo): """Obtain requirements required to be present to upgrade a repo. An upgrade will not be allowed if the repository doesn't have the requirements returned by this function. """ return { # Introduced in Mercurial 0.9.2. 'revlogv1', # Introduced in Mercurial 0.9.2. 'store', } def blocksourcerequirements(repo): """Obtain requirements that will prevent an upgrade from occurring. An upgrade cannot be performed if the source repository contains a requirements in the returned set. """ return { # The upgrade code does not yet support these experimental features. # This is an artificial limitation. 'treemanifest', # This was a precursor to generaldelta and was never enabled by default. # It should (hopefully) not exist in the wild. 'parentdelta', # Upgrade should operate on the actual store, not the shared link. 'shared', } def supportremovedrequirements(repo): """Obtain requirements that can be removed during an upgrade. If an upgrade were to create a repository that dropped a requirement, the dropped requirement must appear in the returned set for the upgrade to be allowed. """ return { localrepo.SPARSEREVLOG_REQUIREMENT, } def supporteddestrequirements(repo): """Obtain requirements that upgrade supports in the destination. If the result of the upgrade would create requirements not in this set, the upgrade is disallowed. Extensions should monkeypatch this to add their custom requirements. """ return { 'dotencode', 'fncache', 'generaldelta', 'revlogv1', 'store', localrepo.SPARSEREVLOG_REQUIREMENT, } def allowednewrequirements(repo): """Obtain requirements that can be added to a repository during upgrade. This is used to disallow proposed requirements from being added when they weren't present before. We use a list of allowed requirement additions instead of a list of known bad additions because the whitelist approach is safer and will prevent future, unknown requirements from accidentally being added. """ return { 'dotencode', 'fncache', 'generaldelta', localrepo.SPARSEREVLOG_REQUIREMENT, } def preservedrequirements(repo): return set() deficiency = 'deficiency' optimisation = 'optimization' class improvement(object): """Represents an improvement that can be made as part of an upgrade. The following attributes are defined on each instance: name Machine-readable string uniquely identifying this improvement. It will be mapped to an action later in the upgrade process. type Either ``deficiency`` or ``optimisation``. A deficiency is an obvious problem. An optimization is an action (sometimes optional) that can be taken to further improve the state of the repository. description Message intended for humans explaining the improvement in more detail, including the implications of it. For ``deficiency`` types, should be worded in the present tense. For ``optimisation`` types, should be worded in the future tense. upgrademessage Message intended for humans explaining what an upgrade addressing this issue will do. Should be worded in the future tense. """ def __init__(self, name, type, description, upgrademessage): self.name = name self.type = type self.description = description self.upgrademessage = upgrademessage def __eq__(self, other): if not isinstance(other, improvement): # This is what python tell use to do return NotImplemented return self.name == other.name def __ne__(self, other): return not self == other def __hash__(self): return hash(self.name) allformatvariant = [] def registerformatvariant(cls): allformatvariant.append(cls) return cls class formatvariant(improvement): """an improvement subclass dedicated to repository format""" type = deficiency ### The following attributes should be defined for each class: # machine-readable string uniquely identifying this improvement. it will be # mapped to an action later in the upgrade process. name = None # message intended for humans explaining the improvement in more detail, # including the implications of it ``deficiency`` types, should be worded # in the present tense. description = None # message intended for humans explaining what an upgrade addressing this # issue will do. should be worded in the future tense. upgrademessage = None # value of current Mercurial default for new repository default = None def __init__(self): raise NotImplementedError() @staticmethod def fromrepo(repo): """current value of the variant in the repository""" raise NotImplementedError() @staticmethod def fromconfig(repo): """current value of the variant in the configuration""" raise NotImplementedError() class requirementformatvariant(formatvariant): """formatvariant based on a 'requirement' name. Many format variant are controlled by a 'requirement'. We define a small subclass to factor the code. """ # the requirement that control this format variant _requirement = None @staticmethod def _newreporequirements(ui): return localrepo.newreporequirements( ui, localrepo.defaultcreateopts(ui)) @classmethod def fromrepo(cls, repo): assert cls._requirement is not None return cls._requirement in repo.requirements @classmethod def fromconfig(cls, repo): assert cls._requirement is not None return cls._requirement in cls._newreporequirements(repo.ui) @registerformatvariant class fncache(requirementformatvariant): name = 'fncache' _requirement = 'fncache' default = True description = _('long and reserved filenames may not work correctly; ' 'repository performance is sub-optimal') upgrademessage = _('repository will be more resilient to storing ' 'certain paths and performance of certain ' 'operations should be improved') @registerformatvariant class dotencode(requirementformatvariant): name = 'dotencode' _requirement = 'dotencode' default = True description = _('storage of filenames beginning with a period or ' 'space may not work correctly') upgrademessage = _('repository will be better able to store files ' 'beginning with a space or period') @registerformatvariant class generaldelta(requirementformatvariant): name = 'generaldelta' _requirement = 'generaldelta' default = True description = _('deltas within internal storage are unable to ' 'choose optimal revisions; repository is larger and ' 'slower than it could be; interaction with other ' 'repositories may require extra network and CPU ' 'resources, making "hg push" and "hg pull" slower') upgrademessage = _('repository storage will be able to create ' 'optimal deltas; new repository data will be ' 'smaller and read times should decrease; ' 'interacting with other repositories using this ' 'storage model should require less network and ' 'CPU resources, making "hg push" and "hg pull" ' 'faster') @registerformatvariant class sparserevlog(requirementformatvariant): name = 'sparserevlog' _requirement = localrepo.SPARSEREVLOG_REQUIREMENT default = True description = _('in order to limit disk reading and memory usage on older ' 'version, the span of a delta chain from its root to its ' 'end is limited, whatever the relevant data in this span. ' 'This can severly limit Mercurial ability to build good ' 'chain of delta resulting is much more storage space being ' 'taken and limit reusability of on disk delta during ' 'exchange.' ) upgrademessage = _('Revlog supports delta chain with more unused data ' 'between payload. These gaps will be skipped at read ' 'time. This allows for better delta chains, making a ' 'better compression and faster exchange with server.') @registerformatvariant class removecldeltachain(formatvariant): name = 'plain-cl-delta' default = True description = _('changelog storage is using deltas instead of ' 'raw entries; changelog reading and any ' 'operation relying on changelog data are slower ' 'than they could be') upgrademessage = _('changelog storage will be reformated to ' 'store raw entries; changelog reading will be ' 'faster; changelog size may be reduced') @staticmethod def fromrepo(repo): # Mercurial 4.0 changed changelogs to not use delta chains. Search for # changelogs with deltas. cl = repo.changelog chainbase = cl.chainbase return all(rev == chainbase(rev) for rev in cl) @staticmethod def fromconfig(repo): return True @registerformatvariant class compressionengine(formatvariant): name = 'compression' default = 'zlib' description = _('Compresion algorithm used to compress data. ' 'Some engine are faster than other') upgrademessage = _('revlog content will be recompressed with the new ' 'algorithm.') @classmethod def fromrepo(cls, repo): for req in repo.requirements: if req.startswith('exp-compression-'): return req.split('-', 2)[2] return 'zlib' @classmethod def fromconfig(cls, repo): return repo.ui.config('experimental', 'format.compression') def finddeficiencies(repo): """returns a list of deficiencies that the repo suffer from""" deficiencies = [] # We could detect lack of revlogv1 and store here, but they were added # in 0.9.2 and we don't support upgrading repos without these # requirements, so let's not bother. for fv in allformatvariant: if not fv.fromrepo(repo): deficiencies.append(fv) return deficiencies def findoptimizations(repo): """Determine optimisation that could be used during upgrade""" # These are unconditionally added. There is logic later that figures out # which ones to apply. optimizations = [] optimizations.append(improvement( name='redeltaparent', type=optimisation, description=_('deltas within internal storage will be recalculated to ' 'choose an optimal base revision where this was not ' 'already done; the size of the repository may shrink and ' 'various operations may become faster; the first time ' 'this optimization is performed could slow down upgrade ' 'execution considerably; subsequent invocations should ' 'not run noticeably slower'), upgrademessage=_('deltas within internal storage will choose a new ' 'base revision if needed'))) optimizations.append(improvement( name='redeltamultibase', type=optimisation, description=_('deltas within internal storage will be recalculated ' 'against multiple base revision and the smallest ' 'difference will be used; the size of the repository may ' 'shrink significantly when there are many merges; this ' 'optimization will slow down execution in proportion to ' 'the number of merges in the repository and the amount ' 'of files in the repository; this slow down should not ' 'be significant unless there are tens of thousands of ' 'files and thousands of merges'), upgrademessage=_('deltas within internal storage will choose an ' 'optimal delta by computing deltas against multiple ' 'parents; may slow down execution time ' 'significantly'))) optimizations.append(improvement( name='redeltaall', type=optimisation, description=_('deltas within internal storage will always be ' 'recalculated without reusing prior deltas; this will ' 'likely make execution run several times slower; this ' 'optimization is typically not needed'), upgrademessage=_('deltas within internal storage will be fully ' 'recomputed; this will likely drastically slow down ' 'execution time'))) optimizations.append(improvement( name='redeltafulladd', type=optimisation, description=_('every revision will be re-added as if it was new ' 'content. It will go through the full storage ' 'mechanism giving extensions a chance to process it ' '(eg. lfs). This is similar to "redeltaall" but even ' 'slower since more logic is involved.'), upgrademessage=_('each revision will be added as new content to the ' 'internal storage; this will likely drastically slow ' 'down execution time, but some extensions might need ' 'it'))) return optimizations def determineactions(repo, deficiencies, sourcereqs, destreqs): """Determine upgrade actions that will be performed. Given a list of improvements as returned by ``finddeficiencies`` and ``findoptimizations``, determine the list of upgrade actions that will be performed. The role of this function is to filter improvements if needed, apply recommended optimizations from the improvements list that make sense, etc. Returns a list of action names. """ newactions = [] knownreqs = supporteddestrequirements(repo) for d in deficiencies: name = d.name # If the action is a requirement that doesn't show up in the # destination requirements, prune the action. if name in knownreqs and name not in destreqs: continue newactions.append(d) # FUTURE consider adding some optimizations here for certain transitions. # e.g. adding generaldelta could schedule parent redeltas. return newactions def _revlogfrompath(repo, path): """Obtain a revlog from a repo path. An instance of the appropriate class is returned. """ if path == '00changelog.i': return changelog.changelog(repo.svfs) elif path.endswith('00manifest.i'): mandir = path[:-len('00manifest.i')] return manifest.manifestrevlog(repo.svfs, tree=mandir) else: #reverse of "/".join(("data", path + ".i")) return filelog.filelog(repo.svfs, path[5:-2]) def _copyrevlogs(ui, srcrepo, dstrepo, tr, deltareuse, forcedeltabothparents): """Copy revlogs between 2 repos.""" revcount = 0 srcsize = 0 srcrawsize = 0 dstsize = 0 fcount = 0 frevcount = 0 fsrcsize = 0 frawsize = 0 fdstsize = 0 mcount = 0 mrevcount = 0 msrcsize = 0 mrawsize = 0 mdstsize = 0 crevcount = 0 csrcsize = 0 crawsize = 0 cdstsize = 0 # Perform a pass to collect metadata. This validates we can open all # source files and allows a unified progress bar to be displayed. for unencoded, encoded, size in srcrepo.store.walk(): if unencoded.endswith('.d'): continue rl = _revlogfrompath(srcrepo, unencoded) info = rl.storageinfo(exclusivefiles=True, revisionscount=True, trackedsize=True, storedsize=True) revcount += info['revisionscount'] or 0 datasize = info['storedsize'] or 0 rawsize = info['trackedsize'] or 0 srcsize += datasize srcrawsize += rawsize # This is for the separate progress bars. if isinstance(rl, changelog.changelog): crevcount += len(rl) csrcsize += datasize crawsize += rawsize elif isinstance(rl, manifest.manifestrevlog): mcount += 1 mrevcount += len(rl) msrcsize += datasize mrawsize += rawsize elif isinstance(rl, filelog.filelog): fcount += 1 frevcount += len(rl) fsrcsize += datasize frawsize += rawsize else: error.ProgrammingError('unknown revlog type') if not revcount: return ui.write(_('migrating %d total revisions (%d in filelogs, %d in manifests, ' '%d in changelog)\n') % (revcount, frevcount, mrevcount, crevcount)) ui.write(_('migrating %s in store; %s tracked data\n') % ( (util.bytecount(srcsize), util.bytecount(srcrawsize)))) # Used to keep track of progress. progress = None def oncopiedrevision(rl, rev, node): progress.increment() # Do the actual copying. # FUTURE this operation can be farmed off to worker processes. seen = set() for unencoded, encoded, size in srcrepo.store.walk(): if unencoded.endswith('.d'): continue oldrl = _revlogfrompath(srcrepo, unencoded) newrl = _revlogfrompath(dstrepo, unencoded) if isinstance(oldrl, changelog.changelog) and 'c' not in seen: ui.write(_('finished migrating %d manifest revisions across %d ' 'manifests; change in size: %s\n') % (mrevcount, mcount, util.bytecount(mdstsize - msrcsize))) ui.write(_('migrating changelog containing %d revisions ' '(%s in store; %s tracked data)\n') % (crevcount, util.bytecount(csrcsize), util.bytecount(crawsize))) seen.add('c') progress = srcrepo.ui.makeprogress(_('changelog revisions'), total=crevcount) elif isinstance(oldrl, manifest.manifestrevlog) and 'm' not in seen: ui.write(_('finished migrating %d filelog revisions across %d ' 'filelogs; change in size: %s\n') % (frevcount, fcount, util.bytecount(fdstsize - fsrcsize))) ui.write(_('migrating %d manifests containing %d revisions ' '(%s in store; %s tracked data)\n') % (mcount, mrevcount, util.bytecount(msrcsize), util.bytecount(mrawsize))) seen.add('m') if progress: progress.complete() progress = srcrepo.ui.makeprogress(_('manifest revisions'), total=mrevcount) elif 'f' not in seen: ui.write(_('migrating %d filelogs containing %d revisions ' '(%s in store; %s tracked data)\n') % (fcount, frevcount, util.bytecount(fsrcsize), util.bytecount(frawsize))) seen.add('f') if progress: progress.complete() progress = srcrepo.ui.makeprogress(_('file revisions'), total=frevcount) ui.note(_('cloning %d revisions from %s\n') % (len(oldrl), unencoded)) oldrl.clone(tr, newrl, addrevisioncb=oncopiedrevision, deltareuse=deltareuse, forcedeltabothparents=forcedeltabothparents) info = newrl.storageinfo(storedsize=True) datasize = info['storedsize'] or 0 dstsize += datasize if isinstance(newrl, changelog.changelog): cdstsize += datasize elif isinstance(newrl, manifest.manifestrevlog): mdstsize += datasize else: fdstsize += datasize progress.complete() ui.write(_('finished migrating %d changelog revisions; change in size: ' '%s\n') % (crevcount, util.bytecount(cdstsize - csrcsize))) ui.write(_('finished migrating %d total revisions; total change in store ' 'size: %s\n') % (revcount, util.bytecount(dstsize - srcsize))) def _filterstorefile(srcrepo, dstrepo, requirements, path, mode, st): """Determine whether to copy a store file during upgrade. This function is called when migrating store files from ``srcrepo`` to ``dstrepo`` as part of upgrading a repository. Args: srcrepo: repo we are copying from dstrepo: repo we are copying to requirements: set of requirements for ``dstrepo`` path: store file being examined mode: the ``ST_MODE`` file type of ``path`` st: ``stat`` data structure for ``path`` Function should return ``True`` if the file is to be copied. """ # Skip revlogs. if path.endswith(('.i', '.d')): return False # Skip transaction related files. if path.startswith('undo'): return False # Only copy regular files. if mode != stat.S_IFREG: return False # Skip other skipped files. if path in ('lock', 'fncache'): return False return True def _finishdatamigration(ui, srcrepo, dstrepo, requirements): """Hook point for extensions to perform additional actions during upgrade. This function is called after revlogs and store files have been copied but before the new store is swapped into the original location. """ def _upgraderepo(ui, srcrepo, dstrepo, requirements, actions): """Do the low-level work of upgrading a repository. The upgrade is effectively performed as a copy between a source repository and a temporary destination repository. The source repository is unmodified for as long as possible so the upgrade can abort at any time without causing loss of service for readers and without corrupting the source repository. """ assert srcrepo.currentwlock() assert dstrepo.currentwlock() ui.write(_('(it is safe to interrupt this process any time before ' 'data migration completes)\n')) if 'redeltaall' in actions: deltareuse = revlog.revlog.DELTAREUSENEVER elif 'redeltaparent' in actions: deltareuse = revlog.revlog.DELTAREUSESAMEREVS elif 'redeltamultibase' in actions: deltareuse = revlog.revlog.DELTAREUSESAMEREVS elif 'redeltafulladd' in actions: deltareuse = revlog.revlog.DELTAREUSEFULLADD else: deltareuse = revlog.revlog.DELTAREUSEALWAYS with dstrepo.transaction('upgrade') as tr: _copyrevlogs(ui, srcrepo, dstrepo, tr, deltareuse, 'redeltamultibase' in actions) # Now copy other files in the store directory. # The sorted() makes execution deterministic. for p, kind, st in sorted(srcrepo.store.vfs.readdir('', stat=True)): if not _filterstorefile(srcrepo, dstrepo, requirements, p, kind, st): continue srcrepo.ui.write(_('copying %s\n') % p) src = srcrepo.store.rawvfs.join(p) dst = dstrepo.store.rawvfs.join(p) util.copyfile(src, dst, copystat=True) _finishdatamigration(ui, srcrepo, dstrepo, requirements) ui.write(_('data fully migrated to temporary repository\n')) backuppath = pycompat.mkdtemp(prefix='upgradebackup.', dir=srcrepo.path) backupvfs = vfsmod.vfs(backuppath) # Make a backup of requires file first, as it is the first to be modified. util.copyfile(srcrepo.vfs.join('requires'), backupvfs.join('requires')) # We install an arbitrary requirement that clients must not support # as a mechanism to lock out new clients during the data swap. This is # better than allowing a client to continue while the repository is in # an inconsistent state. ui.write(_('marking source repository as being upgraded; clients will be ' 'unable to read from repository\n')) scmutil.writerequires(srcrepo.vfs, srcrepo.requirements | {'upgradeinprogress'}) ui.write(_('starting in-place swap of repository data\n')) ui.write(_('replaced files will be backed up at %s\n') % backuppath) # Now swap in the new store directory. Doing it as a rename should make # the operation nearly instantaneous and atomic (at least in well-behaved # environments). ui.write(_('replacing store...\n')) tstart = util.timer() util.rename(srcrepo.spath, backupvfs.join('store')) util.rename(dstrepo.spath, srcrepo.spath) elapsed = util.timer() - tstart ui.write(_('store replacement complete; repository was inconsistent for ' '%0.1fs\n') % elapsed) # We first write the requirements file. Any new requirements will lock # out legacy clients. ui.write(_('finalizing requirements file and making repository readable ' 'again\n')) scmutil.writerequires(srcrepo.vfs, requirements) # The lock file from the old store won't be removed because nothing has a # reference to its new location. So clean it up manually. Alternatively, we # could update srcrepo.svfs and other variables to point to the new # location. This is simpler. backupvfs.unlink('store/lock') return backuppath def upgraderepo(ui, repo, run=False, optimize=None): """Upgrade a repository in place.""" optimize = set(optimize or []) repo = repo.unfiltered() # Ensure the repository can be upgraded. missingreqs = requiredsourcerequirements(repo) - repo.requirements if missingreqs: raise error.Abort(_('cannot upgrade repository; requirement ' 'missing: %s') % _(', ').join(sorted(missingreqs))) blockedreqs = blocksourcerequirements(repo) & repo.requirements if blockedreqs: raise error.Abort(_('cannot upgrade repository; unsupported source ' 'requirement: %s') % _(', ').join(sorted(blockedreqs))) # FUTURE there is potentially a need to control the wanted requirements via # command arguments or via an extension hook point. newreqs = localrepo.newreporequirements( repo.ui, localrepo.defaultcreateopts(repo.ui)) newreqs.update(preservedrequirements(repo)) noremovereqs = (repo.requirements - newreqs - supportremovedrequirements(repo)) if noremovereqs: raise error.Abort(_('cannot upgrade repository; requirement would be ' 'removed: %s') % _(', ').join(sorted(noremovereqs))) noaddreqs = (newreqs - repo.requirements - allowednewrequirements(repo)) if noaddreqs: raise error.Abort(_('cannot upgrade repository; do not support adding ' 'requirement: %s') % _(', ').join(sorted(noaddreqs))) unsupportedreqs = newreqs - supporteddestrequirements(repo) if unsupportedreqs: raise error.Abort(_('cannot upgrade repository; do not support ' 'destination requirement: %s') % _(', ').join(sorted(unsupportedreqs))) # Find and validate all improvements that can be made. alloptimizations = findoptimizations(repo) # Apply and Validate arguments. optimizations = [] for o in alloptimizations: if o.name in optimize: optimizations.append(o) optimize.discard(o.name) if optimize: # anything left is unknown raise error.Abort(_('unknown optimization action requested: %s') % ', '.join(sorted(optimize)), hint=_('run without arguments to see valid ' 'optimizations')) deficiencies = finddeficiencies(repo) actions = determineactions(repo, deficiencies, repo.requirements, newreqs) actions.extend(o for o in sorted(optimizations) # determineactions could have added optimisation if o not in actions) def printrequirements(): ui.write(_('requirements\n')) ui.write(_(' preserved: %s\n') % _(', ').join(sorted(newreqs & repo.requirements))) if repo.requirements - newreqs: ui.write(_(' removed: %s\n') % _(', ').join(sorted(repo.requirements - newreqs))) if newreqs - repo.requirements: ui.write(_(' added: %s\n') % _(', ').join(sorted(newreqs - repo.requirements))) ui.write('\n') def printupgradeactions(): for a in actions: ui.write('%s\n %s\n\n' % (a.name, a.upgrademessage)) if not run: fromconfig = [] onlydefault = [] for d in deficiencies: if d.fromconfig(repo): fromconfig.append(d) elif d.default: onlydefault.append(d) if fromconfig or onlydefault: if fromconfig: ui.write(_('repository lacks features recommended by ' 'current config options:\n\n')) for i in fromconfig: ui.write('%s\n %s\n\n' % (i.name, i.description)) if onlydefault: ui.write(_('repository lacks features used by the default ' 'config options:\n\n')) for i in onlydefault: ui.write('%s\n %s\n\n' % (i.name, i.description)) ui.write('\n') else: ui.write(_('(no feature deficiencies found in existing ' 'repository)\n')) ui.write(_('performing an upgrade with "--run" will make the following ' 'changes:\n\n')) printrequirements() printupgradeactions() unusedoptimize = [i for i in alloptimizations if i not in actions] if unusedoptimize: ui.write(_('additional optimizations are available by specifying ' '"--optimize <name>":\n\n')) for i in unusedoptimize: ui.write(_('%s\n %s\n\n') % (i.name, i.description)) return # Else we're in the run=true case. ui.write(_('upgrade will perform the following actions:\n\n')) printrequirements() printupgradeactions() upgradeactions = [a.name for a in actions] ui.write(_('beginning upgrade...\n')) with repo.wlock(), repo.lock(): ui.write(_('repository locked and read-only\n')) # Our strategy for upgrading the repository is to create a new, # temporary repository, write data to it, then do a swap of the # data. There are less heavyweight ways to do this, but it is easier # to create a new repo object than to instantiate all the components # (like the store) separately. tmppath = pycompat.mkdtemp(prefix='upgrade.', dir=repo.path) backuppath = None try: ui.write(_('creating temporary repository to stage migrated ' 'data: %s\n') % tmppath) # clone ui without using ui.copy because repo.ui is protected repoui = repo.ui.__class__(repo.ui) dstrepo = hg.repository(repoui, path=tmppath, create=True) with dstrepo.wlock(), dstrepo.lock(): backuppath = _upgraderepo(ui, repo, dstrepo, newreqs, upgradeactions) finally: ui.write(_('removing temporary repository %s\n') % tmppath) repo.vfs.rmtree(tmppath, forcibly=True) if backuppath: ui.warn(_('copy of old repository backed up at %s\n') % backuppath) ui.warn(_('the old repository will not be deleted; remove ' 'it to free up disk space once the upgraded ' 'repository is verified\n'))