tests/test-convert-mtn.t
author |
Gregory Szorc <gregory.szorc@gmail.com> |
|
Thu, 13 Oct 2016 12:50:27 +0200 |
changeset 30155 |
b7a966ce89ed |
parent 28683 |
d0210a35c81a
|
child 36050 |
233fb0b91a35 |
permissions |
-rw-r--r-- |
changelog: disable delta chains
This patch disables delta chains on changelogs. After this patch, new
entries on changelogs - including existing changelogs - will be stored
as the fulltext of that data (likely compressed). No delta computation
will be performed.
An overview of delta chains and data justifying this change follows.
Revlogs try to store entries as a delta against a previous entry (either
a parent revision in the case of generaldelta or the previous physical
revision when not using generaldelta). Most of the time this is the
correct thing to do: it frequently results in less CPU usage and smaller
storage.
Delta chains are most effective when the base revision being deltad
against is similar to the current data. This tends to occur naturally
for manifests and file data, since only small parts of each tend to
change with each revision. Changelogs, however, are a different story.
Changelog entries represent changesets/commits. And unless commits in a
repository are homogonous (same author, changing same files, similar
commit messages, etc), a delta from one entry to the next tends to be
relatively large compared to the size of the entry. This means that
delta chains tend to be short. How short? Here is the full vs delta
revision breakdown on some real world repos:
Repo % Full % Delta Max Length
hg 45.8 54.2 6
mozilla-central 42.4 57.6 8
mozilla-unified 42.5 57.5 17
pypy 46.1 53.9 6
python-zstandard 46.1 53.9 3
(I threw in python-zstandard as an example of a repo that is homogonous.
It contains a small Python project with changes all from the same
author.)
Contrast this with the manifest revlog for these repos, where 99+% of
revisions are deltas and delta chains run into the thousands.
So delta chains aren't as useful on changelogs. But even a short delta
chain may provide benefits. Let's measure that.
Delta chains may require less CPU to read revisions if the CPU time
spent reading smaller deltas is less than the CPU time used to
decompress larger individual entries. We can measure this via
`hg perfrevlog -c -d 1` to iterate a revlog to resolve each revision's
fulltext. Here are the results of that command on a repo using delta
chains in its changelog and on a repo without delta chains:
hg (forward)
! wall 0.407008 comb 0.410000 user 0.410000 sys 0.000000 (best of 25)
! wall 0.390061 comb 0.390000 user 0.390000 sys 0.000000 (best of 26)
hg (reverse)
! wall 0.515221 comb 0.520000 user 0.520000 sys 0.000000 (best of 19)
! wall 0.400018 comb 0.400000 user 0.390000 sys 0.010000 (best of 25)
mozilla-central (forward)
! wall 4.508296 comb 4.490000 user 4.490000 sys 0.000000 (best of 3)
! wall 4.370222 comb 4.370000 user 4.350000 sys 0.020000 (best of 3)
mozilla-central (reverse)
! wall 5.758995 comb 5.760000 user 5.720000 sys 0.040000 (best of 3)
! wall 4.346503 comb 4.340000 user 4.320000 sys 0.020000 (best of 3)
mozilla-unified (forward)
! wall 4.957088 comb 4.950000 user 4.940000 sys 0.010000 (best of 3)
! wall 4.660528 comb 4.650000 user 4.630000 sys 0.020000 (best of 3)
mozilla-unified (reverse)
! wall 6.119827 comb 6.110000 user 6.090000 sys 0.020000 (best of 3)
! wall 4.675136 comb 4.670000 user 4.670000 sys 0.000000 (best of 3)
pypy (forward)
! wall 1.231122 comb 1.240000 user 1.230000 sys 0.010000 (best of 8)
! wall 1.164896 comb 1.160000 user 1.160000 sys 0.000000 (best of 9)
pypy (reverse)
! wall 1.467049 comb 1.460000 user 1.460000 sys 0.000000 (best of 7)
! wall 1.160200 comb 1.170000 user 1.160000 sys 0.010000 (best of 9)
The data clearly shows that it takes less wall and CPU time to resolve
revisions when there are no delta chains in the changelogs, regardless
of the direction of traversal. Furthermore, not using a delta chain
means that fulltext resolution in reverse is as fast as iterating
forward. So not using delta chains on the changelog is a clear CPU win
for reading operations.
An example of a user-visible operation showing this speed-up is revset
evaluation. Here are results for
`hg perfrevset 'author(gps) or author(mpm)'`:
hg
! wall 1.655506 comb 1.660000 user 1.650000 sys 0.010000 (best of 6)
! wall 1.612723 comb 1.610000 user 1.600000 sys 0.010000 (best of 7)
mozilla-central
! wall 17.629826 comb 17.640000 user 17.600000 sys 0.040000 (best of 3)
! wall 17.311033 comb 17.300000 user 17.260000 sys 0.040000 (best of 3)
What about 00changelog.i size?
Repo Delta Chains No Delta Chains
hg 7,033,250 6,976,771
mozilla-central 82,978,748 81,574,623
mozilla-unified 88,112,349 86,702,162
pypy 20,740,699 20,659,741
The data shows that removing delta chains from the changelog makes the
changelog smaller.
Delta chains are also used during changegroup generation. This
operation essentially converts a series of revisions to one large
delta chain. And changegroup generation is smart: if the delta in
the revlog matches what the changegroup is emitting, it will reuse
the delta instead of recalculating it. We can measure the impact
removing changelog delta chains has on changegroup generation via
`hg perfchangegroupchangelog`:
hg
! wall 1.589245 comb 1.590000 user 1.590000 sys 0.000000 (best of 7)
! wall 1.788060 comb 1.790000 user 1.790000 sys 0.000000 (best of 6)
mozilla-central
! wall 17.382585 comb 17.380000 user 17.340000 sys 0.040000 (best of 3)
! wall 20.161357 comb 20.160000 user 20.120000 sys 0.040000 (best of 3)
mozilla-unified
! wall 18.722839 comb 18.720000 user 18.680000 sys 0.040000 (best of 3)
! wall 21.168075 comb 21.170000 user 21.130000 sys 0.040000 (best of 3)
pypy
! wall 4.828317 comb 4.830000 user 4.820000 sys 0.010000 (best of 3)
! wall 5.415455 comb 5.420000 user 5.410000 sys 0.010000 (best of 3)
The data shows eliminating delta chains makes the changelog part of
changegroup generation slower. This is expected since we now have to
compute deltas for revisions where we could recycle the delta before.
It is worth putting this regression into context of overall changegroup
times. Here is the rough total CPU time spent in changegroup generation
for various repos while using delta chains on the changelog:
Repo CPU Time (s) CPU Time w/ compression
hg 4.50 7.05
mozilla-central 111.1 222.0
pypy 28.68 75.5
Before compression, removing delta chains from the changegroup adds
~4.4% overhead to hg changegroup generation, 1.3% to mozilla-central,
and 2.0% to pypy. When you factor in zlib compression, these percentages
are roughly divided by 2.
While the increased CPU usage for changegroup generation is unfortunate,
I think it is acceptable because the percentage is small, server
operators (those likely impacted most by this) have other mechanisms
to mitigate CPU consumption (namely reducing zlib compression level and
pre-generated clone bundles), and because there is room to optimize this
in the future. For example, we could use the nullid as the base revision,
effectively encoding the full revision for each entry in the changegroup.
When doing this, `hg perfchangegroupchangelog` nearly halves:
mozilla-unified
! wall 21.168075 comb 21.170000 user 21.130000 sys 0.040000 (best of 3)
! wall 11.196461 comb 11.200000 user 11.190000 sys 0.010000 (best of 3)
This looks very promising as a future optimization opportunity.
It's worth that the changes in test-acl.t to the changegroup part size.
This is because revision 6 in the changegroup had a delta chain of
length 2 before and after this patch the base revision is nullrev.
When the base revision is nullrev, cg2packer.deltaparent() hardcodes
the *previous* revision from the changegroup as the delta parent.
This caused the delta in the changegroup to switch base revisions,
the delta to change, and the size to change accordingly. While the
size increased in this case, I think sizes will remain the same
on average, as the delta base for changelog revisions doesn't matter
too much (as this patch shows). So, I don't consider this a regression.
#require mtn
Monotone directory is called .monotone on *nix and monotone
on Windows.
#if windows
$ mtndir=monotone
#else
$ mtndir=.monotone
#endif
$ echo "[extensions]" >> $HGRCPATH
$ echo "convert=" >> $HGRCPATH
Windows version of monotone home
$ APPDATA=$HOME; export APPDATA
tedious monotone keys configuration
The /dev/null redirection is necessary under Windows, or
it complains about home directory permissions
$ mtn --quiet genkey test@selenic.com 1>/dev/null 2>&1 <<EOF
> passphrase
> passphrase
> EOF
$ cat >> $HOME/$mtndir/monotonerc <<EOF
> function get_passphrase(keypair_id)
> return "passphrase"
> end
> EOF
create monotone repository
$ mtn db init --db=repo.mtn
$ mtn --db=repo.mtn --branch=com.selenic.test setup workingdir
$ cd workingdir
$ echo a > a
$ mkdir dir
$ echo b > dir/b
$ echo d > dir/d
$ $PYTHON -c 'file("bin", "wb").write("a\\x00b")'
$ echo c > c
$ mtn add a dir/b dir/d c bin
mtn: adding 'a' to workspace manifest
mtn: adding 'bin' to workspace manifest
mtn: adding 'c' to workspace manifest
mtn: adding 'dir' to workspace manifest
mtn: adding 'dir/b' to workspace manifest
mtn: adding 'dir/d' to workspace manifest
$ mtn ci -m initialize
mtn: beginning commit on branch 'com.selenic.test'
mtn: committed revision 0f6e5e4f2e7d2a8ef312408f57618abf026afd90
update monotone working directory
$ mtn mv a dir/a
mtn: skipping 'dir', already accounted for in workspace
mtn: renaming 'a' to 'dir/a' in workspace manifest
$ echo a >> dir/a
$ echo b >> dir/b
$ mtn drop c
mtn: dropping 'c' from workspace manifest
$ $PYTHON -c 'file("bin", "wb").write("b\\x00c")'
$ mtn ci -m update1
mtn: beginning commit on branch 'com.selenic.test'
mtn: committed revision 51d0a982464573a2a2cf5ee2c9219c652aaebeff
$ cd ..
convert once
$ hg convert -s mtn repo.mtn
assuming destination repo.mtn-hg
initializing destination repo.mtn-hg repository
scanning source...
sorting...
converting...
1 initialize
0 update1
$ cd workingdir
$ echo e > e
$ mtn add e
mtn: adding 'e' to workspace manifest
$ mtn drop dir/b
mtn: dropping 'dir/b' from workspace manifest
$ mtn mv bin bin2
mtn: renaming 'bin' to 'bin2' in workspace manifest
$ mtn ci -m 'update2 "with" quotes'
mtn: beginning commit on branch 'com.selenic.test'
mtn: committed revision ebe58335d85d8cb176b6d0a12be04f5314b998da
test directory move
$ mkdir -p dir1/subdir1
$ mkdir -p dir1/subdir2_other
$ echo file1 > dir1/subdir1/file1
$ echo file2 > dir1/subdir2_other/file1
$ mtn add dir1/subdir1/file1 dir1/subdir2_other/file1
mtn: adding 'dir1' to workspace manifest
mtn: adding 'dir1/subdir1' to workspace manifest
mtn: adding 'dir1/subdir1/file1' to workspace manifest
mtn: adding 'dir1/subdir2_other' to workspace manifest
mtn: adding 'dir1/subdir2_other/file1' to workspace manifest
$ mtn ci -m createdir1
mtn: beginning commit on branch 'com.selenic.test'
mtn: committed revision a8d62bc04fee4d2936d28e98bbcc81686dd74306
$ mtn rename dir1/subdir1 dir1/subdir2
mtn: skipping 'dir1', already accounted for in workspace
mtn: renaming 'dir1/subdir1' to 'dir1/subdir2' in workspace manifest
$ mtn ci -m movedir1
mtn: beginning commit on branch 'com.selenic.test'
mtn: committed revision 2c3d241bbbfe538b1b51d910f5676407e3f4d3a6
test subdirectory move
$ mtn mv dir dir2
mtn: renaming 'dir' to 'dir2' in workspace manifest
$ echo newfile > dir2/newfile
$ mtn drop dir2/d
mtn: dropping 'dir2/d' from workspace manifest
$ mtn add dir2/newfile
mtn: adding 'dir2/newfile' to workspace manifest
$ mtn ci -m movedir
mtn: beginning commit on branch 'com.selenic.test'
mtn: committed revision fdb5a02dae8bfce3a79b3393680af471016e1b4c
Test directory removal with empty directory
$ mkdir dir2/dir
$ mkdir dir2/dir/subdir
$ echo f > dir2/dir/subdir/f
$ mkdir dir2/dir/emptydir
$ mtn add --quiet -R dir2/dir
$ mtn ci -m emptydir
mtn: beginning commit on branch 'com.selenic.test'
mtn: committed revision 8bbf76d717001d24964e4604739fdcd0f539fc88
$ mtn drop -R dir2/dir
mtn: dropping 'dir2/dir/subdir/f' from workspace manifest
mtn: dropping 'dir2/dir/subdir' from workspace manifest
mtn: dropping 'dir2/dir/emptydir' from workspace manifest
mtn: dropping 'dir2/dir' from workspace manifest
$ mtn ci -m dropdirectory
mtn: beginning commit on branch 'com.selenic.test'
mtn: committed revision 2323d4bc324e6c82628dc04d47a9fd32ad24e322
test directory and file move
$ mkdir -p dir3/d1
$ echo a > dir3/a
$ mtn add dir3/a dir3/d1
mtn: adding 'dir3' to workspace manifest
mtn: adding 'dir3/a' to workspace manifest
mtn: adding 'dir3/d1' to workspace manifest
$ mtn ci -m dirfilemove
mtn: beginning commit on branch 'com.selenic.test'
mtn: committed revision 47b192f720faa622f48c68d1eb075b26d405aa8b
$ mtn mv dir3/a dir3/d1/a
mtn: skipping 'dir3/d1', already accounted for in workspace
mtn: renaming 'dir3/a' to 'dir3/d1/a' in workspace manifest
$ mtn mv dir3/d1 dir3/d2
mtn: skipping 'dir3', already accounted for in workspace
mtn: renaming 'dir3/d1' to 'dir3/d2' in workspace manifest
$ mtn ci -m dirfilemove2
mtn: beginning commit on branch 'com.selenic.test'
mtn: committed revision 8b543a400d3ee7f6d4bb1835b9b9e3747c8cb632
test directory move into another directory move
$ mkdir dir4
$ mkdir dir5
$ echo a > dir4/a
$ mtn add dir4/a dir5
mtn: adding 'dir4' to workspace manifest
mtn: adding 'dir4/a' to workspace manifest
mtn: adding 'dir5' to workspace manifest
$ mtn ci -m dirdirmove
mtn: beginning commit on branch 'com.selenic.test'
mtn: committed revision 466e0b2afc7a55aa2b4ab2f57cb240bb6cd66fc7
$ mtn mv dir5 dir6
mtn: renaming 'dir5' to 'dir6' in workspace manifest
$ mtn mv dir4 dir6/dir4
mtn: skipping 'dir6', already accounted for in workspace
mtn: renaming 'dir4' to 'dir6/dir4' in workspace manifest
$ mtn ci -m dirdirmove2
mtn: beginning commit on branch 'com.selenic.test'
mtn: committed revision 3d1f77ebad0c23a5d14911be3b670f990991b749
test diverging directory moves
$ mkdir -p dir7/dir9/dir8
$ echo a > dir7/dir9/dir8/a
$ echo b > dir7/dir9/b
$ echo c > dir7/c
$ mtn add -R dir7
mtn: adding 'dir7' to workspace manifest
mtn: adding 'dir7/c' to workspace manifest
mtn: adding 'dir7/dir9' to workspace manifest
mtn: adding 'dir7/dir9/b' to workspace manifest
mtn: adding 'dir7/dir9/dir8' to workspace manifest
mtn: adding 'dir7/dir9/dir8/a' to workspace manifest
$ mtn ci -m divergentdirmove
mtn: beginning commit on branch 'com.selenic.test'
mtn: committed revision 08a08511f18b428d840199b062de90d0396bc2ed
$ mtn mv dir7 dir7-2
mtn: renaming 'dir7' to 'dir7-2' in workspace manifest
$ mtn mv dir7-2/dir9 dir9-2
mtn: renaming 'dir7-2/dir9' to 'dir9-2' in workspace manifest
$ mtn mv dir9-2/dir8 dir8-2
mtn: renaming 'dir9-2/dir8' to 'dir8-2' in workspace manifest
$ mtn ci -m divergentdirmove2
mtn: beginning commit on branch 'com.selenic.test'
mtn: committed revision 4a736634505795f17786fffdf2c9cbf5b11df6f6
test large file support (> 32kB)
>>> fp = file('large-file', 'wb')
>>> for x in xrange(10000): fp.write('%d\n' % x)
>>> fp.close()
$ md5sum.py large-file
5d6de8a95c3b6bf9e0ffb808ba5299c1 large-file
$ mtn add large-file
mtn: adding 'large-file' to workspace manifest
$ mtn ci -m largefile
mtn: beginning commit on branch 'com.selenic.test'
mtn: committed revision f0a20fecd10dc4392d18fe69a03f1f4919d3387b
test suspending (closing a branch)
$ mtn suspend f0a20fecd10dc4392d18fe69a03f1f4919d3387b 2> /dev/null
$ cd ..
convert incrementally
$ hg convert -s mtn repo.mtn
assuming destination repo.mtn-hg
scanning source...
sorting...
converting...
12 update2 "with" quotes
11 createdir1
10 movedir1
9 movedir
8 emptydir
7 dropdirectory
6 dirfilemove
5 dirfilemove2
4 dirdirmove
3 dirdirmove2
2 divergentdirmove
1 divergentdirmove2
0 largefile
$ glog()
> {
> hg log -G --template '{rev} "{desc|firstline}" files: {files}\n' "$@"
> }
$ cd repo.mtn-hg
$ hg up -C
12 files updated, 0 files merged, 0 files removed, 0 files unresolved
no open descendant heads on branch "com.selenic.test", updating to a closed head
(committing will reopen branch "com.selenic.test")
$ glog
@ 14 "largefile" files: large-file
|
o 13 "divergentdirmove2" files: dir7-2/c dir7/c dir7/dir9/b dir7/dir9/dir8/a dir8-2/a dir9-2/b
|
o 12 "divergentdirmove" files: dir7/c dir7/dir9/b dir7/dir9/dir8/a
|
o 11 "dirdirmove2" files: dir4/a dir6/dir4/a
|
o 10 "dirdirmove" files: dir4/a
|
o 9 "dirfilemove2" files: dir3/a dir3/d2/a
|
o 8 "dirfilemove" files: dir3/a
|
o 7 "dropdirectory" files: dir2/dir/subdir/f
|
o 6 "emptydir" files: dir2/dir/subdir/f
|
o 5 "movedir" files: dir/a dir/d dir2/a dir2/newfile
|
o 4 "movedir1" files: dir1/subdir1/file1 dir1/subdir2/file1
|
o 3 "createdir1" files: dir1/subdir1/file1 dir1/subdir2_other/file1
|
o 2 "update2 "with" quotes" files: bin bin2 dir/b e
|
o 1 "update1" files: a bin c dir/a dir/b
|
o 0 "initialize" files: a bin c dir/b dir/d
manifest
$ hg manifest
bin2
dir1/subdir2/file1
dir1/subdir2_other/file1
dir2/a
dir2/newfile
dir3/d2/a
dir6/dir4/a
dir7-2/c
dir8-2/a
dir9-2/b
e
large-file
contents
$ cat dir2/a
a
a
$ test -d dir2/dir && echo 'removed dir2/dir is still there!'
[1]
file move
$ hg log -v -C -r 1 | grep copies
copies: dir/a (a)
check directory move
$ hg manifest -r 4
bin2
dir/a
dir/d
dir1/subdir2/file1
dir1/subdir2_other/file1
e
$ test -d dir1/subdir2 || echo 'new dir1/subdir2 does not exist!'
$ test -d dir1/subdir1 && echo 'renamed dir1/subdir1 is still there!'
[1]
$ hg log -v -C -r 4 | grep copies
copies: dir1/subdir2/file1 (dir1/subdir1/file1)
check file remove with directory move
$ hg manifest -r 5
bin2
dir1/subdir2/file1
dir1/subdir2_other/file1
dir2/a
dir2/newfile
e
check file move with directory move
$ hg manifest -r 9
bin2
dir1/subdir2/file1
dir1/subdir2_other/file1
dir2/a
dir2/newfile
dir3/d2/a
e
check file directory directory move
$ hg manifest -r 11
bin2
dir1/subdir2/file1
dir1/subdir2_other/file1
dir2/a
dir2/newfile
dir3/d2/a
dir6/dir4/a
e
check divergent directory moves
$ hg manifest -r 13
bin2
dir1/subdir2/file1
dir1/subdir2_other/file1
dir2/a
dir2/newfile
dir3/d2/a
dir6/dir4/a
dir7-2/c
dir8-2/a
dir9-2/b
e
test large file support (> 32kB)
$ md5sum.py large-file
5d6de8a95c3b6bf9e0ffb808ba5299c1 large-file
check branch closing
$ hg branches -a
$ hg branches -c
com.selenic.test 14:* (closed) (glob)