* Deal with clock skew, both forwards and backwards, when logging
information to the git-annex branch.
* GIT_ANNEX_VECTOR_CLOCK can now be set to a fixed value (eg 1)
rather than needing to be advanced each time a new change is made.
* Misuse of GIT_ANNEX_VECTOR_CLOCK will no longer confuse git-annex.
When changing a file in the git-annex branch, the vector clock to use is now
determined by first looking at the current time (or GIT_ANNEX_VECTOR_CLOCK
when set), and comparing it to the newest vector clock already in use in
that file. If a newer time stamp was already in use, advance it forward by
a second instead.
When the clock is set to a time in the past, this avoids logging with
an old timestamp, which would risk that log line later being ignored in favor
of "newer" line that is really not newer.
When a log entry has been made with a clock that was set far ahead in the
future, this avoids newer information being logged with an older timestamp
and so being ignored in favor of that future-timestamped information.
Once all clocks get fixed, this will result in the vector clocks being
incremented, until finally enough time has passed that time gets back ahead
of the vector clock value, and then it will return to usual operation.
(This latter situation is not ideal, but it seems the best that can be done.
The issue with it is, since all writers will be incrementing the last
vector clock they saw, there's no way to tell when one writer made a write
significantly later in time than another, so the earlier write might
arbitrarily be picked when merging. This problem is why git-annex uses
timestamps in the first place, rather than pure vector clocks.)
Advancing forward by 1 second is somewhat arbitrary. setDead
advances a timestamp by just 1 picosecond, and the vector clock could
too. But then it would interfere with setDead, which wants to be
overrulled by any change. So it could use 2 picoseconds or something,
but that seems weird. It could just as well advance it forward by a
minute or whatever, but then it would be harder for real time to catch
up with the vector clock when forward clock slew had happened.
A complication is that many log files contain several different peices of
information, and it may be best to only use vector clocks for the same peice
of information. For example, a key's location log file contains
InfoPresent/InfoMissing for each UUID, and it only looks at the vector
clocks for the UUID that is being changed, and not other UUIDs.
Although exactly where the dividing line is can be hard to determine.
Consider metadata logs, where a field "tag" can have multiple values set
at different times. Should it advance forward past the last tag?
Probably. What about when a different field is set, should it look at
the clocks of other fields? Perhaps not, but currently it does, and
this does not seems like it will cause any problems.
Another one I'm not entirely sure about is the export log, which is
keyed by (fromuuid, touuid). So if multiple repos are exporting to the
same remote, different vector clocks can be used for that remote.
It looks like that's probably ok, because it does not try to determine
what order things occurred when there was an export conflict.
Sponsored-by: Jochen Bartl on Patreon
14683da9eb caused a test suite failure.
When the content of a key is not present, a LinkAnnexFailed is returned,
but replaceFile then tried to move the file into place, and since it was
not written, that crashed.
Sponsored-by: Boyd Stephen Smith Jr. on Patreon
git-annex get when run as the first git-annex command in a new repo did not
populate unlocked files. (Reversion in version 8.20210621)
I am not entirely happy with this, because I don't understand how
428c91606b caused the problem in the first
place, and I don't fully understand how skipping calling scanAnnexedFiles
during autoinit avoids the problem.
Kept the explicit call to scanAnnexedFiles during git-annex init,
so that when reconcileStaged is expensive, it can be made to run then,
rather than at some later point when the information is needed.
Sponsored-by: Brock Spratlen on Patreon
The pass was needed to populate files when annex.thin was set,
but in commit 73e0cbbb19,
reconcileStaged started to do that. So, this second pass is not needed
any longer.
This avoids it calling enteringStage VerifyStage when it's used in
places that only fall back to verification rarely, and which might be
called while in TransferStage and be going to perform a transfer after
the verification.
This is a result of an audit of every use of getInodeCaches,
to find places that misbehave when the annex object is not in the inode
cache, despite pointer files for the same key being in the inode cache.
Unfortunately, that is the case for objects that were in v7 repos that
upgraded to v8. Added a note about this gotcha to getInodeCaches.
Database.Keys.reconcileStaged, then annex.thin is set, would fail to
populate pointer files in this situation. Changed it to check if the
annex object is unmodified the same way inAnnex does, falling back to a
checksum if the inode cache is not recorded.
Sponsored-by: Dartmouth College's Datalad project
The goal is that Database.Keys be able to use it; it can't use
Annex.Content.Presence due to an import loop.
Several other things also needed to be moved to Annex.Verify as a
conseqence.
Some uses of linkFromAnnex are inside replaceWorkTreeFile, which was
already safe, but others use it directly on the work tree file, which
was race-prone. Eg, if the work tree file was first removed, then
linkFromAnnex called to populate it, the user could have re-written it in
the interim.
This came to light during an audit of all calls of addInodeCaches,
looking for such races. All the other uses of it seem ok.
Sponsored-by: Brett Eisenberg on Patreon
In Annex.Content, the object file was statted after pointer files were
populated. But if annex.thin is set, once the pointer files are
populated, the object file can potentially be modified via the hard
link. So, it was possible, though seemingly very unlikely, for the inode
of the modified object file to be cached.
Command.Fix and Command.Fsck had similar problems, statting the work
tree files after they were in place. Changed them to stat the temp file
that gets moved into place. This does rely on .git/annex being on the
same filesystem. If it's not, the cached inode will not be the same as
the one that the temp file gets moved to. Result will be that git-annex
will later need to do an expensive verification of the content of the
worktree files. Note that the cross-filesystem move of the temp file
already is a larger amount of extra work, so this seems acceptable.
Sponsored-by: Luke Shumaker on Patreon
Fix bug that caused some transfers to incorrectly fail with "content
changed while it was being sent", when the content was not changed.
While I don't know how to reproduce the problem that several people
reported, it is presumably due to the inode cache somehow being stale.
So check isUnmodified', and if it's not modified, include the file's
current inode cache in the set to accept, when checking for modification
after the transfer.
That seems like the right thing to do for another reason: The failure
says the file changed while it was being sent, but if the object file was
changed before the transfer started, that's wrong. So it needs to check
before allowing the transfer at all if the file is modified.
(Other calls to sameInodeCache or elemInodeCaches, when operating on inode
caches from the database, could also be problimatic if the inode cache is
somehow getting stale. This does not address such problems.)
Sponsored-by: Dartmouth College's Datalad project
* sync: When --quiet is used, run git commit, push, and pull without
their ususual output.
* merge: When --quiet is used, run git merge without its usual output.
This might also make --quiet work better for some other commands
that make commits, like git-annex adjust.
Sponsored-by: Kevin Mueller on Patreon
To avoid mistakes like commit 0ccbed4f6f,
be explicit about the two variants of this.
Incidentially avoids a small amount of overhead in calling reverse.
Sponsored-by: Shae Erisson on Patreon
Fix a bug that prevented getting content from a repository that started out
as a bare repository, or had annex.crippledfilesystem set, and was
converted to a non-bare repository.
This unfortunately means that inAnnex check gets slowed down by a stat call
in normal repos when the content is not present. Oh well, such is the cost
of backwards compatability with old mistakes.
Sponsored-by: Mark Reidenbach on Patreon
init: Fix misbehavior when core.sharedRepository = group that caused it to
enter an adjusted branch. (Reversion in version 8.20210630)
Commit 4b1b9d7a83 made init call
freezeContent in case there was a hook that could prevent writing in
situations where perms don't. But with the above git config, freezeContent
does not prevent write at all. So init needs to do what freezeContent does
with a non-shared git config.
Or init could check for that config, and skip the probing, since it
won't actually be preventing write to any files. But that would make init
too aware if details of Annex.Perms, and also would break if the git config
were changed after init.
Sponsored-by: Dartmouth College's Datalad project
Dropping an object with drop --unused or dropunused will mark it as
dead, preventing fsck --all from complaining about it after it's been
dropped from all repositories.
If another repository still has a copy, it won't be treated as dead
until it's also dropped from there.
The drop has to use --unused, can't be --key or something else, because
this indicates that the user has recently ran git-annex unused. If it
checked the unused log on every drop, bad things would happen when the
unused log was out of date, eg a file used to be unused but then got
re-added. Marking such a file as dead could be confusing. When the user
uses --unused/dropunused, they must consider the unused information to be
up-to-date.
The particular workflow this enables is:
git annex add foo
git annex unannex foo
git annex unused
git annex drop --unused / dropunused
git annex fsck --all # no warnings
The docs for git-annex unannex say to use git-annex unused and dropunused,
so the user should be pointed in this direction when they want to undo an
accidental add.
Sponsored-by: Brock Spratlen on Patreon
Transfers from or to a local git repo could fail without a reason being
given, if the content failed to verify, or if the object file's stat
changed while it was being copied. Now display messages in these cases.
Sponsored-by: Jack Hill on Patreon
When stall detection is enabled, and a transfer is in progress,
it would display a doubled message:
(transfer already in progress, or unable to take transfer lock) (transfer already in progress, or unable to take transfer lock)
That happened because the forward retry decider had a start size of 0,
and an end size of whatever amount of the object the other process had
downloaded. So it incorrectly thought that the transferrer process had
made progress, when it had in fact immediately given up with that
message.
Instead, use the reported value from the progress meter. If a remote
does not report progress, this will mean it doesn't forward retry, in a
situation where it used to. But most remotes do report progress, and any
remote that does not can be fixed to, by using watchFileSize when
downloading. Also, some remotes might preallocate the temp file (eg
bittorrent), so relying on statting its size at this level to get
progress is dubious.
The same change was made to Annex/Transfer.hs, although only
Annex/TransferrerPool.hs needed to be changed to avoid the duplicate
message.
(An alternate fix would have been to start the retry decider with the
size of the object file before downloading begins, rather than 0.)
Sponsored-by: Brett Eisenberg on Patreon
Freeze first sets the file perms, and then runs
freezecontent-command. Thaw runs thawcontent-command before
restoring file permissions. This is in case the freeze command
prevents changing file perms, as eg setting a file immutable does.
Also, changing file perms tends to mess up previously set ACLs.
git-annex init's probe for crippled filesystem uses them, so if file perms
don't work, but freezecontent-command manages to prevent write to a file,
it won't treat the filesystem as crippled.
When the the filesystem has been probed as crippled, the hooks are not
used, because there seems to be no point then; git-annex won't be relying
on locking annex objects down. Also, this avoids them being run when the
file perms have not been changed, in case they somehow rely on
git-annex's setting of the file perms in order to work.
Sponsored-by: Dartmouth College's Datalad project
Eg, before with a .gitattributes like:
*.2 annex.numcopies=2
*.1 annex.numcopies=1
And foo.1 and foo.2 having the same content and key, git-annex drop foo.1 foo.2
would succeed, leaving just 1 copy, despite foo.2 needing 2 copies.
It dropped foo.1 first and then skipped foo.2 since its content was gone.
Now that the keys database includes locked files, this longstanding wart
can be fixed.
Sponsored-by: Noam Kremen on Patreon
Most of this is just refactoring. But, handleDropsFrom
did not verify that associated files from the keys db were still
accurate, and has now been fixed to.
A minor improvement to this would be to avoid calling catKeyFile
twice on the same file, when getting the numcopies and mincopies value,
in the common case where the same file has the highest value for both.
But, it avoids checking every associated file, so it will scale well to
lots of dups already.
Sponsored-by: Kevin Mueller on Patreon
This was an old problem when the files were being added unlocked,
so the changelog mentions that being fixed. However, recently it's also
affected locked files.
The fix for locked files is kind of stupidly simple. moveAnnex already
handles populating unlocked files, and only does it when the object file
was not already present. So remove the redundant populateUnlockedFiles
call. (That call was added all the way back in
cfaac52b88, and has always been
unncessary.)
Sponsored-by: Dartmouth College's Datalad project
moveAnnex only gets to that check if the object file was not present
before. So in the case where dup files are being added repeatedly,
it will only run the first time, and so there's no significant speedup
from doing it; all it avoids is a single sqlite lookup. Since MVar
accesses do have overhead, it's better to optimise for the common case,
where unlocked files are supported.
removeAnnex is less clear cut, but I think mostly is skipped running on
keys when the object has already been dropped, so similar reasoning
applies.
This will mostly just avoid a DB lookup, so things get marginally
faster. But in cases where there are many files using the same key, it
can be a more significant speedup.
Added overhead is one MVar lookup per call, which should be small
enough, since this happens after transferring or ingesting a file,
which is always a lot more work than that. It would be nice, though,
to move getGitConfig to AnnexRead, which there is an open todo about.
That seems very unlikely to happen, but still, it's possible it could.
And with the recent addition of locked files to the keys db, this could
be called by places that did not call it before, so it seems even more
important it's correct.
Adds an extra stat of the file, and is potentially racy, but both
problems are fixed by the unix-2.8.0 path. I have not tested that path
builds because that package is not yet released and it would be difficult
to install it since it's tightly tied to a ghc version.
Clear visible progress bar first.
Removed showSideActionAfter because it can't be used in reconcileStaged
(import loop). Instead, it counts the number of files it
processes and displays it after it's seen a sufficient to know it's
taking a while.
Sponsored-by: Dartmouth College's Datalad project
This makes git checkout and git merge hooks do the work to catch up with
changes that they made to the tree. Rather than doing it at some later
point when the user is not thinking about that past operation.
Sponsored-by: Dartmouth College's Datalad project
Following commit c941ab6f5b, this avoids
the second, redundant scan when annex.thin is not set.
The benchmark now runs in 35.5 seconds, down from 40 seconds.
Note that the inode cache of the annex object has to be passed to
addInodeCaches now, because it might not already be in the inode caches,
unlike previously.
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
This reverts commit 0f10f208a7.
The implementation of this turns out to be unsafe; it can lead to a keys
db deadlock. scanAnnexedFiles injects a call to inAnnex into
reconcileStaged, but inAnnex sometimes needs to read from the keys db,
which will try to re-open it when it's in the process of being opened.
The exclusive lock of gitAnnexKeysDbLock will then deadlock.
This needs to be done in some other way...
reconcileStaged was doing a redundant scan to scannAnnexedFiles.
It would probably make sense to move the body of scannAnnexedFiles
into reconcileStaged, the separation does not really serve any purpose.
Sponsored-by: Dartmouth College's Datalad project
Avoids users thinking this scan is a big deal, when it's not in the
majority of repos.
showSideActionAfter has some ugly caveats, since it has to display in
the background of another action. I could not see a better way to do it
and it works fine in this particular case. It also doesn't really belong
in Annex.Concurrent, but cannot go in Messages due to an import loop.
Sponsored-by: Dartmouth College's Datalad project
Streaming through git this way speeds it up by around 25%. This is
similar to the optimisations of seeking annexed files.
Sponsored-by: Dartmouth College's Datalad project
