Comment 2 for bug 281040
Revision history for this message
| Jonathan Lange (jml) wrote Re: Timeout updating branch record in puller | #2 |
© 2004
Canonical Ltd.
•
Terms of use
•
Data privacy
•
Contact Launchpad Support
•
Blog
•
Careers
•
System status
•
0e1f616
(Get the code!)
On IRC later:
<BjornT> stub: it could also be that the update statement is blocking on some block, couldn't it? afaik, there's a bug in the oops reporting, so that the last sql statement isn't included as sql, but as non-sql. considering that the time is pretty much exactly 25 seconds, it looks like it's the statement that is timing out, doesn't it?
<stub> BjornT, jml: I think you are right. The only thing that would block that particular statement (apart from me or a losa messing with the db schema) is another connection has updated that row and not committed.
<stub> Make an educated guess as to what other process could be opening a transaction on Branch that long and stop it doing that?
<stub> We don't have any tools to tell us unfortunately.
<jtv> stub: pg_stat_activity?
<stub> jtv: Doesn't help after the fact
<stub> jml: Also, not all of our branches are mirrors, so it makes sense to break out that metadata into a separate table.
<jml> stub: actually we do mirror all of our branches.
<jml> stub: (but it still makes sense to move that data to a separate table)