Thanks. I'll look into this next time it happens.
For now, I tried the create index commands again, and they worked in
a split-second. So, it must have been the lock contention, and not
the VACUUM'ng.
On Jul 21, 2006, at 6:45 AM, Tom Lane wrote:
> Claire McLister <mclister@zeesource.net> writes:
>> Yes, that could be the case. We have a python function that imports
>> CSV files, which can take a long time, and that may have been running
>> during that time. I didn't look at the pg_lock file. What should I be
>> looking for?
>
> A record with granted = false for the stuck process (joining pid to
> pg_stat_activity will help you determine which process goes with each
> record, or look in "ps" output). If you find one, look for a record
> for the same lock with granted = true and a conflicting lock type;
> that
> tells you which process is blocking the lock.
>
> regards, tom lane
>
> ---------------------------(end of
> broadcast)---------------------------
> TIP 5: don't forget to increase your free space map settings