On 06/30/2015 07:37 PM, Alvaro Herrera wrote:
> Jeff Janes wrote:
>
>> I've gotten the LWLock deadlock again. User backend 24841 holds the
>> WALInsertLocks 7 and is blocked attempting to acquire 6 . So it seems to
>> be violating the lock ordering rules (although I don't see that rule
>> spelled out in xlog.c)
>
> Hmm, interesting -- pg_stat_statement is trying to re-do an operation
> that involves updating a GIN index, while WAL-logging for the original
> update is still ongoing, it seems.
I don't think pg_stat_statement has anything to do with this. You can
see from the backtrace that pg_stat_statement is enabled, as the call
went through the pgss_ExecutorRun executor hook, but that's all.
- Heikki