[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: svn commit: rev 5194 - trunk/subversion/libsvn_fs

From: Karl Fogel <kfogel_at_newton.ch.collab.net>
Date: 2003-03-04 04:32:13 CET

I ran five stress.pl's with 5194 subtracted, and got one instance of
the famous error:

   Committing:
   Sending wcstress.24092/trunk/bar1/foo1
   Sending wcstress.24092/trunk/bar2/foo1
   Sending wcstress.24092/trunk/bar2/foo2
   Sending wcstress.24092/trunk/foo2
   Transmitting file data .subversion/libsvn_client/commit.c:670: \
     (apr_err=160020)
   svn: Item already exists in filesystem
   svn: Commit failed (details follow):
   subversion/libsvn_fs/bdb/nodes-table.c:146: (apr_err=160020)
   svn: successor id `0.0.1z3' (for `0.0.1ym') already exists \
     in filesystem '/home/kfogel/src/stress/repostress/db'
   subversion/libsvn_fs/bdb/bdb-err.c:61: (apr_err=160030)
   svn: Berkeley DB deadlock error
   svn: Berkeley DB error while reading node revision for \
     filesystem /home/kfogel/src/stress/repostress/db:
   DB_LOCK_DEADLOCK: Locker killed to resolve a deadlock
   unexpected commit fail: exit status: 256
   
Of course, this doesn't tell us much, as far as 5194 is concerned. It
would be useful to do, say, 20 runs of five instances each, and see if
there's any statistically significant difference in how often this
particular error occurs.

I don't know how the error is occurring, in terms of specific code
paths.

-Karl

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Mar 4 05:08:03 2003

This is an archived mail posted to the Subversion Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.