One more thought -- I keep coming back to the fact that when we turn on logging in the JDBC driver on the client side,
theproblem does not occur. The only possible reason I can see for this having any affect on the problem is the small
delayintroduced by the synchronous logging. Since this is only showing up on commit of a database transaction which
followsclose on the heels of a rollback on the same connection, is there any chance that there is some very small
"settlingtime" needed for a rollback, and we're sometimes getting in ahead of this?
-Kevin
>>> Tom Lane <tgl@sss.pgh.pa.us> 09/12/05 5:39 PM >>>
"Kevin Grittner" <Kevin.Grittner@wicourts.gov> writes:
> org.postgresql.util.PSQLException: ERROR: canceling query due to user request
The only possible trigger of that message is a SIGINT sent to the backend.
Now the backend will SIGINT itself if a statement timeout expires, so one
possibility is that you have statement_timeout set and it's getting
exceeded. Otherwise you need to be looking for external causes.
regards, tom lane
