Re: Query on Postgres locks

Yes and hence my original question- why would the queries and the locks waiting to be granted disappear when there is no timeout?

Sent from my Android phone using Symantec TouchDown (www.symantec.com)

-----Original Message-----
From: Jan Harasym [jan(dot)harasym(at)massive(dot)se]
Received: Thursday, 24 Sep 2015, 20:31
To: Porwal, Utkarsh [utkarsh(dot)porwal(at)emc(dot)com]; Wei Shan [weishan(dot)ang(at)gmail(dot)com]
CC: pgsql-admin(at)postgresql(dot)org [pgsql-admin(at)postgresql(dot)org]
Subject: RE: [ADMIN] Query on Postgres locks

For many things “0” is “disabled”

So there is no timeout, it will wait forever.

From: pgsql-admin-owner(at)postgresql(dot)org [mailto:pgsql-admin-owner(at)postgresql(dot)org] On Behalf Of Porwal, Utkarsh
Sent: September 24, 2015 12:37 PM
To: Wei Shan
Cc: pgsql-admin(at)postgresql(dot)org
Subject: Re: [ADMIN] Query on Postgres locks

Hi Wei shan,

For me the statement_timeout is defaulted to zero

statement_timeout | 0 | ms

Thanks and Regards,
Utkarsh Porwal

From: Wei Shan [mailto:weishan(dot)ang(at)gmail(dot)com]
Sent: Thursday, September 24, 2015 3:49 PM
To: Porwal, Utkarsh
Cc: pgsql-admin(at)postgresql(dot)org<mailto:pgsql-admin(at)postgresql(dot)org>
Subject: Re: [ADMIN] Query on Postgres locks

Hi,

If you look at pg_settings table, you can see the default value for statement_timeout. Only from 9.3, there's a parameter called lock_timeout.

Cheers.

On 24 September 2015 at 17:37, Porwal, Utkarsh <utkarsh(dot)porwal(at)emc(dot)com<mailto:utkarsh(dot)porwal(at)emc(dot)com>> wrote:
The postgres version is –

psql (9.0.7)

From: pgsql-admin-owner(at)postgresql(dot)org<mailto:pgsql-admin-owner(at)postgresql(dot)org> [mailto:pgsql-admin-owner(at)postgresql(dot)org<mailto:pgsql-admin-owner(at)postgresql(dot)org>] On Behalf Of Porwal, Utkarsh
Sent: Thursday, September 24, 2015 3:04 PM
To: pgsql-admin(at)postgresql(dot)org<mailto:pgsql-admin(at)postgresql(dot)org>
Subject: [ADMIN] Query on Postgres locks

Dear All,

I have a perl script which will attempt to drop/recreate some intermediate tables. This works by trying to acquire an access exclusive lock on the intermediate tables.

When there is a backup running through pg_dump, I see an entry in pg_locks/pg_stat_activity table for the above which is waiting for pg_dump to finish, which is expected.

Since backup on huge environment takes around 8 hrs to complete, I noticed that the entries waiting for locks are removed after certain amount of time which I couldn’t record even though the perl script is still running.
Do you guys know if the lock waiting will ultimately timeout and when? I don’t have any statement_timeout specified.

Any pointers on this behavior?

Regards,
Utkarsh

--
Regards,
Ang Wei Shan