update and open a todo about something I'm pondering

This commit is contained in:
Joey Hess 2022-10-12 15:53:56 -04:00
parent 6fbd337e34
commit d5cd1de280
No known key found for this signature in database
GPG key ID: DB12DB0FF05F8F38
2 changed files with 36 additions and 0 deletions

View file

@ -0,0 +1,16 @@
[[!comment format=mdwn
username="joey"
subject="""comment 26"""
date="2022-10-12T19:34:05Z"
content="""
I've avoided the excessive sqlite database writes. Which doubled the speed
of git-annex in some circumstances, wow!
@yoh see if it still happens once you upgrade to a git-annex
with [[!commit 6fbd337e34cd62881ed5e21616e9c00d7bf35378]].
It would be possible for git-annex to do its own locking around writes
to the sqlite database. That would surely avoid any problem that sqlite might
have that would cause ErrorBusy. I want to think some about
[[todo/withExclusiveLock_blocking_issue]] first.
"""]]

View file

@ -0,0 +1,20 @@
Some parts of git-annex use withExclusiveLock or otherwise wait for an
exclusive lock and hold it while performing an operation. Now consider what
happens if the git-annex process is suspended. Another git-annex process
that is running and that blocks on the same lock will stall forever, until
the git-annex process is resumed.
These time windows tend to be small, but may not always be.
Would it be better for the second git-annex process, rather than hanging
indefinitely, to try to take the lock a few times over a few seconds, and
then error out? The risk with doing that is, when 2 concurrent git-annex
processes are running and taking the locks repeatedly, one might get
unlucky, fail to take the lock, and error out, when waiting a little longer
would have succeeded, because the other process is not holding the lock all
the time.
Is there any better way git-annex could handle this? Is it a significant
problem at all? I don't think I've ever seen it happen, but I rarely ^Z
git-annex either. How do other programs handle this, if at all?
--[[Joey]]