Vlastimil Havranek <vlastimil.havranek@xitee.com> writes:
> Here is the log with the settings zou have suggested:
[ extracted... ]
> database_A SID:496f4f4d.454 16:12:27.953 LOG: duration: 0.000 ms
> parse <unnamed>: PREPARE TRANSACTION
> '48801_AOiMyUwnZ6MjBA==_Q3ZoYXZyYW5la2pkYmNEU0xpdmU='
> database_A SID:496f4f4d.454 16:12:27.953 LOG: duration: 0.000 ms bind
> <unnamed>: PREPARE TRANSACTION
> '48801_AOiMyUwnZ6MjBA==_Q3ZoYXZyYW5la2pkYmNEU0xpdmU='
> database_A SID:496f4f4d.454 16:12:27.953 LOG: execute <unnamed>:
> PREPARE TRANSACTION '48801_AOiMyUwnZ6MjBA==_Q3ZoYXZyYW5la2pkYmNEU0xpdmU='
> database_A SID:496f4f4d.454 16:12:27.953 LOG: duration: 0.000 ms
> database_A SID:496f4f4d.454 16:12:28.000 LOG: duration: 0.000 ms
> parse <unnamed>: COMMIT PREPARED
> '48801_AOiMyUwnZ6MjBA==_Q3ZoYXZyYW5la2pkYmNEU0xpdmU='
> database_A SID:496f4f4d.454 16:12:28.000 LOG: duration: 0.000 ms bind
> <unnamed>: COMMIT PREPARED
> '48801_AOiMyUwnZ6MjBA==_Q3ZoYXZyYW5la2pkYmNEU0xpdmU='
> database_A SID:496f4f4d.454 16:12:28.000 LOG: execute <unnamed>:
> COMMIT PREPARED '48801_AOiMyUwnZ6MjBA==_Q3ZoYXZyYW5la2pkYmNEU0xpdmU='
> database_A SID:496f4f4d.454 16:12:28.000 ERROR: prepared transaction
> with identifier "48801_AOiMyUwnZ6MjBA==_Q3ZoYXZyYW5la2pkYmNEU0xpdmU="
> does not exist
Well, this clearly shows the same GID being committed twice --- and by
the same session, so my early thoughts about a race condition are out.
So either Weblogic is broken, or there's a bug in the JDBC driver that
somehow causes this, or we've misread the spec and should allow this.
regards, tom lane