Re: BUG #4232: CREATE INDEX CONCURRENTLY

Many thanks - we actually considered that at some point and I looked for
more specific info on this particular issue and I found about the new
pg_stat_clear_snapshot() function in 8.3 as well.=20
Now that we know that this issue was fixed in 8.3 already it's even more
incentive for us to plan for an upgrade.

Lawrence Cohan.

-----Original Message-----
From: Tom Lane [mailto:tgl@sss.pgh.pa.us]=20
Sent: Tuesday, June 10, 2008 4:58 PM
To: Lawrence Cohan
Cc: pgsql-bugs@postgresql.org
Subject: Re: [BUGS] BUG #4232: CREATE INDEX CONCURRENTLY=20

"Lawrence Cohan" <lawrencec@1shoppingcart.com> writes:
> We must run maintenance tasks like analyze, reindex and vacuum against
our
> PG databases however due to the fact that we are running a 24/7 system
that
> requires database access the reindex at the database level is way too
heavy
> and it is generating deadlocks. I created a job to CREATE INDEX
CONCURRENTLY
> on all user tables and DROP existing INDEX so we don't impact our
> production and now our application is getting errors (like the one
below)
> just because the OID for the index was changed. Is there anything we
could
> do to workaround this issue as so far the only option that clears it
is an
> IIS RESET.

Presumably the errors are coming from re-use of cached plans.  The only
really simple solution would be to upgrade to PG 8.3, which knows about
regenerating cached plans when needed.

            regards, tom lane
Attention:
The information contained in this message and or attachments is intended on=
ly for the person or entity to which it is addressed and may =0D
contain confidential and/or privileged material.  Any review, retransmissio=
n, dissemination or other use of, or taking of any action in =0D
reliance upon, this information by persons or entities other than the inten=
ded recipient is prohibited. If you received this in error, please =0D
contact the sender and delete the material from any system and destroy any =
copies.