Increasing the size of the queue 10x makes git-annex init 7% faster in a
repository with 86000 annexed files.
The memory use goes up, from 70876 kb to 85376 kb.
Avoids database querying overhead when the database is newly created.
In the large repository where git-annex init took 24 seconds, this sped it
up to 20.47 seconds, a speedup of around 15%.
Sponsored-by: Dartmouth College's DANDI project
When running eg git-annex get, for each file it has to read from and
write to the keys database. But it's reading exclusively from one table,
and writing to a different table. So, it is not necessary to flush the
write to the database before reading. This avoids writing the database
once per file, instead it will buffer 1000 changes before writing.
Benchmarking getting 1000 small files from a local origin,
git-annex get now takes 13.62s, down from 22.41s!
git-annex drop now takes 9.07s, down from 18.63s!
Wowowowowowowow!
(It would perhaps have been better if there were separate databases for
the two tables. At least it would have avoided this complexity. Ah well,
this is better than splitting the table in a annex.version upgrade.)
Sponsored-by: Dartmouth College's Datalad project
to track down a blocked indefinitely on MVar that seems to occur after
sqlite throws ErrorBusy but that I have not been able to reproduce when
I made commits synthetically throw ErrorBusy.
Sponsored-by: Dartmouth College's Datalad project
reconcileStaged populates the db, so scanAnnexedFiles does not need to
do it again. It still makes a pass over the HEAD tree, but populating
the db was most of the expensive part.
Benchmarking with 100,000 files, git-annex init now takes 40 seconds,
vs 37 seconds with the old, buggy version of this fix. It should be
possible to win those 3 precious seconds per 100k files back, in the
case when when annex.thin is not set, with improvements to reconcileStaged
that avoid needing this second pass.
Sponsored-by: Dartmouth College's Datalad project
Rather than first deleting and then inserting, upsert lets the key
associated with a file be updated in place.
Benchmarked with 100,000 files, and an empty keys database, running
reconcileStaged. It improved from 47 seconds to 34 seconds.
So this got reconcileStaged to be as fast as scanAssociatedFiles,
or faster -- scanAssociatedFiles benchmarks at 37 seconds.
(Also checked for other users of deleteWhere that could be sped up by
upsert. There are a couple, but they are not in performance critical
code paths, eg recordExportTreeCurrent is only run once per tree
export.)
I would have liked to rename FileKeyIndex to FileKeyUnique since it is
being used as a uniqueness constraint now, not just to get an index.
But, that gets converted into part of the SQL schema, and the name
is used by the upsert, so it can't be changed.
Sponsored-by: Dartmouth College's Datalad project
There seems to be no reason to check the time here. I think it was
inherited from code in Database.Fsck, which does have a reason to commit
every few minutes. Removing that syscall speeds up a git-annex init
in a repo with 100000 annexed files by about 3 seconds.
Sponsored-by: Dartmouth College's Datalad project
smudge: Fix a case where an unlocked annexed file that annex.largefiles
does not match could get its unchanged content checked into git, due to git
running the smudge filter unecessarily.
When the file has the same inodecache as an already annexed file,
we can assume that the user is not intending to change how it's stored in
git.
Note that checkunchangedgitfile already handled the inverse case, where the
file was added to git previously. That goes further and actually sha1
hashes the new file and checks if it's the same hash in the index.
It would be possible to generate a key for the file and see if it's the
same as the old key, however that could be considerably more expensive than
sha1 of a small file is, and it is not necessary for the case I have, at
least, where the file is not modified or touched, and so its inode will
match the cache.
git-annex add was changed, when adding a small file, to remove the inode
cache for it. This is necessary to keep the recipe in
doc/tips/largefiles.mdwn for converting from annex to git working.
It also avoids bugs/case_where_using_pathspec_with_git-commit_leaves_s.mdwn
which the earlier try at this change introduced.
This is untested because of rain, also I am operating from truncated
copiler error messages in a bug report that also doesn't mention what the
library version is. Still, it should work.
May break builds with old ghc, in particular DerivingStrategies is
I think fairly new? The pragmas could be ifdefed if necessary. Works with
ghc 8.6.5.
Adds a dependency on filepath-bytestring, an as yet unreleased fork of
filepath that operates on RawFilePath.
Git.Repo also changed to use RawFilePath for the path to the repo.
This does eliminate some RawFilePath -> FilePath -> RawFilePath
conversions. And filepath-bytestring's </> is probably faster.
But I don't expect a major performance improvement from this.
This is mostly groundwork for making Annex.Location use RawFilePath,
which will allow for a conversion-free pipleline.
Bearing in mind that these indexes are really uniqueness constraints
that just happen to also make sqlite generate indexes.
In Database.ContentIndentifier, the ContentIndentifiersKeyRemoteCidIndex
is fine as a uniqueness constraint because it contains all rows from the
table. The ContentIndentifiersCidRemoteIndex is also ok because there
can only be one key for a given (cid, uuid) combination.
In Database.Export, the new ExportTreeFileKeyIndex is the same pair as
the old ExportTreeKeyFileIndex (previously ExportTreeIndex). And
in Database.Keys.SQL, the new InodeCacheKeyIndex is the same pair as the
old KeyInodeCacheIndex.
This is a non-backwards compatable change, so not suitable for merging
w/o a annex.version bump and transition code. Not yet tested.
This improves performance of git-annex benchmark --databases
across the board by 10-25%, since eg Key roundtrips as a ByteString.
(serializeKey' produces a lazy ByteString, so there is still a
copy involved in converting it to a strict ByteString. It may be faster
to switch to using bytestring-strict-builder.)
FilePath and Key are both stored as blobs. This avoids mojibake in some
situations. It would be possible to use varchar instead, if persistent
could avoid converting that to Text, but it seems there is no good
way to do so. See doc/todo/sqlite_database_improvements.mdwn
Eliminated some ugly artifacts of using Read/Show serialization;
constructors and quoted strings are no longer stored in sqlite.
Renamed SRef to SSha to reflect that it is only ever a git sha,
not a ref name. Since it is limited to the characters in a sha,
it is not affected by mojibake, so still uses String.
The only good thing about it is it does not require a major version bump
to improve the database. That will need to happen at some point though.
Potentially very very slow in a large repository.
Ugly use of raw sql.
fsck --incremental/--more: Fix bug that prevented the incremental fsck
information from being updated every 5 minutes as it was supposed to be; it
was only updated after 1000 files were checked, which may be more files
that are possible to fsck in a given fsck time window.
Thanks to Peter Simons for help with analysis of this bug.
Auditing for other cases of the same mistake, the keys db also had it
backwards. This seems unlikely to really have been a problem;
it would need associated files updates etc to be coming in slowly for some
reason and then be interrupted to cause any problem.
IIRC the design of the keys db assumes that any interruped
operation will be restarted, and so it can lose any buffered database
updates safely.
This does not change the overall license of the git-annex program, which
was already AGPL due to a number of sources files being AGPL already.
Legally speaking, I'm adding a new license under which these files are
now available; I already released their current contents under the GPL
license. Now they're dual licensed GPL and AGPL. However, I intend
for all my future changes to these files to only be released under the
AGPL license, and I won't be tracking the dual licensing status, so I'm
simply changing the license statement to say it's AGPL.
(In some cases, others wrote parts of the code of a file and released it
under the GPL; but in all cases I have contributed a significant portion
of the code in each file and it's that code that is getting the AGPL
license; the GPL license of other contributors allows combining with
AGPL code.)
The keys database handle needs to be closed after merging, because the
smudge filter, in another process, updates the database. Old cached info
can be read for a while from the open database handle; closing it ensures
that the info written by the smudge filter is available.
This is pretty horribly ad-hoc, and it's especially nasty that the
transferrer closes the database every time.
This is a mostly backwards compatable change. I broke backwards
compatability in the case where a filename starts with double-quote.
That seems likely to be very rare, and v6 unlocked files are a new feature
anyway, and fsck needs to fix missing associated file mappings anyway. So,
I decided that is good enough.
The encoding used is to just show the String when it contains a problem
character. While that adds some overhead to addAssociatedFile and
removeAssociatedFile, those are not called very often. This approach has
minimal decode overhead, because most filenames won't be encoded that way,
and it only has to look for the leading double-quote to skip the expensive
read. So, getAssociatedFiles remains fast.
I did consider using ByteString instead, but getting a FilePath converted
with all chars intact, even surrigates, is difficult, and it looks like
instance PersistField ByteString uses Text, which I don't trust for problem
encoded data. It would probably be slower too, and it would make the
database less easy to inspect manually.
This lets readonly repos be used. If a repo is readonly, we can ignore the
keys database, because nothing that we can do will change the state of the
repo anyway.
The benchmark shows that the database access is quite fast indeed!
And, it scales linearly to the number of keys, with one exception,
getAssociatedKey.
Based on this benchmark, I don't think I need worry about optimising
for cases where all files are locked and the database is mostly empty.
In those cases, database access will be misses, and according to this
benchmark, should add only 50 milliseconds to runtime.
(NB: There may be some overhead to getting the database opened and locking
the handle that this benchmark doesn't see.)
joey@darkstar:~/src/git-annex>./git-annex benchmark
setting up database with 1000
setting up database with 10000
benchmarking keys database/getAssociatedFiles from 1000 (hit)
time 62.77 μs (62.70 μs .. 62.85 μs)
1.000 R² (1.000 R² .. 1.000 R²)
mean 62.81 μs (62.76 μs .. 62.88 μs)
std dev 201.6 ns (157.5 ns .. 259.5 ns)
benchmarking keys database/getAssociatedFiles from 1000 (miss)
time 50.02 μs (49.97 μs .. 50.07 μs)
1.000 R² (1.000 R² .. 1.000 R²)
mean 50.09 μs (50.04 μs .. 50.17 μs)
std dev 206.7 ns (133.8 ns .. 295.3 ns)
benchmarking keys database/getAssociatedKey from 1000 (hit)
time 211.2 μs (210.5 μs .. 212.3 μs)
1.000 R² (0.999 R² .. 1.000 R²)
mean 211.0 μs (210.7 μs .. 212.0 μs)
std dev 1.685 μs (334.4 ns .. 3.517 μs)
benchmarking keys database/getAssociatedKey from 1000 (miss)
time 173.5 μs (172.7 μs .. 174.2 μs)
1.000 R² (0.999 R² .. 1.000 R²)
mean 173.7 μs (173.0 μs .. 175.5 μs)
std dev 3.833 μs (1.858 μs .. 6.617 μs)
variance introduced by outliers: 16% (moderately inflated)
benchmarking keys database/getAssociatedFiles from 10000 (hit)
time 64.01 μs (63.84 μs .. 64.18 μs)
1.000 R² (1.000 R² .. 1.000 R²)
mean 64.85 μs (64.34 μs .. 66.02 μs)
std dev 2.433 μs (547.6 ns .. 4.652 μs)
variance introduced by outliers: 40% (moderately inflated)
benchmarking keys database/getAssociatedFiles from 10000 (miss)
time 50.33 μs (50.28 μs .. 50.39 μs)
1.000 R² (1.000 R² .. 1.000 R²)
mean 50.32 μs (50.26 μs .. 50.38 μs)
std dev 202.7 ns (167.6 ns .. 252.0 ns)
benchmarking keys database/getAssociatedKey from 10000 (hit)
time 1.142 ms (1.139 ms .. 1.146 ms)
1.000 R² (1.000 R² .. 1.000 R²)
mean 1.142 ms (1.140 ms .. 1.144 ms)
std dev 7.142 μs (4.994 μs .. 10.98 μs)
benchmarking keys database/getAssociatedKey from 10000 (miss)
time 1.094 ms (1.092 ms .. 1.096 ms)
1.000 R² (1.000 R² .. 1.000 R²)
mean 1.095 ms (1.095 ms .. 1.097 ms)
std dev 4.277 μs (2.591 μs .. 7.228 μs)
Writes are optimised by queueing up multiple writes when possible.
The queue is flushed after the Annex monad action finishes. That makes it
happen on program termination, and also whenever a nested Annex monad action
finishes.
Reads are optimised by checking once (per AnnexState) if the database
exists. If the database doesn't exist yet, all reads return mempty.
Reads also cause queued writes to be flushed, so reads will always be
consistent with writes (as long as they're made inside the same Annex monad).
A future optimisation path would be to determine when that's not necessary,
which is probably most of the time, and avoid flushing unncessarily.
Design notes for this commit:
- separate reads from writes
- reuse a handle which is left open until program
exit or until the MVar goes out of scope (and autoclosed then)
- writes are queued
- queue is flushed periodically
- immediate queue flush before any read
- auto-flush queue when database handle is garbage collected
- flush queue on exit from Annex monad
(Note that this may happen repeatedly for a single database connection;
or a connection may be reused for multiple Annex monad actions,
possibly even concurrent ones.)
- if database does not exist (or is empty) the handle
is not opened by reads; reads instead return empty results
- writes open the handle if it was not open previously
Fsck can use the queue for efficiency since it is write-heavy, and only
reads a value before writing it. But, the queue is not suited to the Keys
database.
Renamed the db to keys, since it is various info about a Keys.
Dropping a key will update its pointer files, as long as their content can
be verified to be unmodified. This falls back to checksum verification, but
I want it to use an InodeCache of the key, for speed. But, I have not made
anything populate that cache yet.
