repos.sync throws NoSuchChangeset

[csapuntz@…]

It seems to me that a sync function shouldn't throw.

Stacktrace:
  File "/usr/bin/bitten-master", line 7, in ?
    sys.exit(
  File "build/bdist.linux-i686/egg/bitten/master.py", line 405, in main
  File "build/bdist.linux-i686/egg/bitten/util/beep.py", line 150, in run
  File "build/bdist.linux-i686/egg/bitten/master.py", line 66, in _enqueue_builds
  File "build/bdist.linux-i686/egg/bitten/queue.py", line 162, in populate
  File "/usr/lib/python2.3/site-packages/trac/versioncontrol/cache.py", line 102, in sync
    self.youngest = self.repos.normalize_rev(self.youngest)
  File "/usr/lib/python2.3/site-packages/trac/versioncontrol/svn_fs.py", line 312, in normalize_rev
    raise NoSuchChangeset(rev)
trac.versioncontrol.api.NoSuchChangeset: No changeset 9408 in the repository

[Emmanuel Blot]

Have you recently upgraded to 0.10.4 ?

Try to run

trac-admin <project> resync

It may be a duplicate of #5259.

[csapuntz@…]

We have been running 0.10.4 for at least a week now with many successful checkins and many successful bitten builds (though we have been seeing a lot of bitten crashes)

It looks like a race of some sort to me. bitten is in a separate process than trac.

Replying to eblot:

Have you recently upgraded to 0.10.4 ?

Try to run

trac-admin <project> resync

It may be a duplicate of #5259.

[csapuntz@…]

BTW, I did a resync after we upgraded.

[Christian Boos]

Get the repository youngest_rev right before we call normalize_rev

[Christian Boos]

Can you please try attachment:fix-sync-race-0.10.5dev.patch​ and report back if the this fixes the problem?

[anonymous]

Yes, this seems to fix our problem. Bitten has been up for 9 hours. Previously it crashed every 1:30 hrs or so, esp when trac was being used by our developers and checkins were happening.

[Christian Boos]

Ok, thanks for testing this. Patch applied in r5611 (trunk) and r5612 (0.10-stable).