This is an archive of the discontinued Mercurial Phabricator instance.

Differential D10725

revlog: update data file record before index rename
ClosedPublic

Authored by joerg.sonnenberger on May 17 2021, 10:26 PM.

Details

Reviewers
indygreg
valentin.gatienbaron
Group Reviewers
hg-reviewers
Commits
rHG46b828b85eb7: revlog: update data file record before index rename
rHG25f42db82ac8: revlog: update data file record before index rename
rHGd92310d45ecb: revlog: update data file record before index rename
Summary

When migrating from inline to non-inline data storage, the data file is
recorded initially as zero sized so that it is removed on failure. But
the record has to be updated before the index is renamed, otherwise
data is lost on rollback.

Diff Detail

Repository
rHG Mercurial
Lint
Automatic diff as part of commit; lint not applicable.
Unit
Automatic diff as part of commit; unit tests not applicable.

Event Timeline

valentin.gatienbaron accepted this revision.May 18 2021, 10:44 PM
valentin.gatienbaron added a subscriber: valentin.gatienbaron.
valentin.gatienbaron added inline comments.
tests/test-transaction-rollback-on-revlog-split.t
7–18

The change here in the next commit should probably be folded into this commit.

joerg.sonnenberger added a commit: rHGd92310d45ecb: revlog: update data file record before index rename.May 19 2021, 1:58 PM
This revision was not accepted when it landed; it landed in state Needs Review.
Closed by commit rHGd92310d45ecb: revlog: update data file record before index rename (authored by joerg.sonnenberger). · Explain Why
This revision was automatically updated to reflect the committed changes.

Revision Contents
Changeset List

Diff 28125

mercurial/revlog.py

e = self.index.entry_binary(i) e = self.index.entry_binary(i)
if i == 0 and self._docket is None: if i == 0 and self._docket is None:
header = self._format_flags | self._format_version header = self._format_flags | self._format_version
header = self.index.pack_header(header) header = self.index.pack_header(header)
e = header + e e = header + e
fp.write(e) fp.write(e)
if self._docket is not None: if self._docket is not None:
self._docket.index_end = fp.tell() self._docket.index_end = fp.tell()
# There is a small transactional race here. If the rename of
# the index fails, we should remove the datafile. It is more
# important to ensure that the data file is not truncated
# when the index is replaced as otherwise data is lost.
tr.replace(self._datafile, self.start(trindex))
# the temp file replace the real index when we exit the context # the temp file replace the real index when we exit the context
# manager # manager
tr.replace(self._indexfile, trindex * self.index.entry_size) tr.replace(self._indexfile, trindex * self.index.entry_size)
nodemaputil.setup_persistent_nodemap(tr, self) nodemaputil.setup_persistent_nodemap(tr, self)
self._chunkclear() self._chunkclear()
if existing_handles: if existing_handles:

tests/test-transaction-rollback-on-revlog-split.t

Test correctness of revlog inline -> non-inline transition Test correctness of revlog inline -> non-inline transition
---------------------------------------------------------- ----------------------------------------------------------
Helper extension to intercept renames.
$ cat > $TESTTMP/intercept_rename.py << EOF
> import os
> import sys
> from mercurial import extensions, util
>
> def extsetup(ui):
> def close(orig, *args, **kwargs):
> path = args[0]._atomictempfile__name
> if path.endswith(b'/.hg/store/data/file.i'):
> os._exit(80)
> return orig(*args, **kwargs)
> extensions.wrapfunction(util.atomictempfile, 'close', close)
> EOF
valentin.gatienbaronUnsubmitted
Not Done

The change here in the next commit should probably be folded into this commit.

valentin.gatienbaron: The change here in the next commit should probably be folded into this commit.
Test offset computation to correctly factor in the index entries themselve. Test offset computation to correctly factor in the index entries themselve.
Also test that the new data size has the correct size if the transaction is aborted
after the index has been replaced.
Test repo has one small, one moderate and one big change. The clone has Test repo has one small, one moderate and one big change. The clone has
the small and moderate change and will transition to non-inline storage when the small and moderate change and will transition to non-inline storage when
adding the big change. adding the big change.
$ hg init troffset-computation --config format.revlog-compression=none $ hg init troffset-computation --config format.revlog-compression=none
$ cd troffset-computation $ cd troffset-computation
$ printf '% 20d' '1' > file $ printf '% 20d' '1' > file
$ hg commit -Aqm_ $ hg commit -Aqm_
$ printf '% 1024d' '1' > file $ printf '% 1024d' '1' > file
$ hg commit -Aqm_ $ hg commit -Aqm_
$ dd if=/dev/zero of=file bs=1k count=128 > /dev/null 2>&1 $ dd if=/dev/zero of=file bs=1k count=128 > /dev/null 2>&1
$ hg commit -Aqm_ $ hg commit -Aqm_
$ cd .. $ cd ..
$ hg clone -r 1 troffset-computation troffset-computation-copy --config format.revlog-compression=none -q $ hg clone -r 1 troffset-computation troffset-computation-copy --config format.revlog-compression=none -q
$ cd troffset-computation-copy $ cd troffset-computation-copy
Reference size:
$ f -s .hg/store/data/file*
.hg/store/data/file.i: size=1174
$ cat > .hg/hgrc <<EOF $ cat > .hg/hgrc <<EOF
> [hooks] > [hooks]
> pretxnchangegroup = python:$TESTDIR/helper-killhook.py:killme > pretxnchangegroup = python:$TESTDIR/helper-killhook.py:killme
> EOF > EOF
$ hg pull ../troffset-computation $ hg pull ../troffset-computation
pulling from ../troffset-computation pulling from ../troffset-computation
[80] [80]
$ cat .hg/store/journal | tr -s '\000' ' ' | grep data/file | tail -1
The first file.i entry should match the size above.
The first file.d entry is the temporary record during the split,
the second entry after the split happened. The sum of the second file.d
and the second file.i entry should match the first file.i entry.
$ cat .hg/store/journal | tr -s '\000' ' ' | grep data/file
data/file.i 1174
data/file.d 0
data/file.d 1046
data/file.i 128 data/file.i 128
$ cd ..
Now retry the same but intercept the rename of the index and check that
the journal does not contain the new index size. This demonstrates the edge case
where the data file is left as garbage.
$ hg clone -r 1 troffset-computation troffset-computation-copy2 --config format.revlog-compression=none -q
$ cd troffset-computation-copy2
$ cat > .hg/hgrc <<EOF
> [extensions]
> intercept_rename = $TESTTMP/intercept_rename.py
> [hooks]
> pretxnchangegroup = python:$TESTDIR/helper-killhook.py:killme
> EOF
$ hg pull ../troffset-computation
pulling from ../troffset-computation
[80]
$ cat .hg/store/journal | tr -s '\000' ' ' | grep data/file
data/file.i 1174
data/file.d 0
data/file.d 1046
$ cd ..