Re: Index space growing even after cleanup via autovacuum in Postgres 9.2

From: Tirthankar Barari <tbarari(at)verizon(dot)com>
To: Francisco Olarte <folarte(at)peoplecall(dot)com>, Tirthankar Barari <tbarari(at)verizon(dot)com>
Cc: pgsql-general <pgsql-general(at)postgresql(dot)org>
Subject: Re: Index space growing even after cleanup via autovacuum in Postgres 9.2
Date: 2014-01-13 16:26:38
Message-ID: 52D413BE.5080904@terremark.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On 01/10/2014 07:06 AM, Francisco Olarte wrote:
> Hi:
>
> On Thu, Jan 9, 2014 at 7:50 PM, Tirthankar Barari <tbarari(at)verizon(dot)com> wrote:
>> We have a table where we insert about 10 million rows everyday. We keep 14
>> day's worth of entries (i.e. 140 mil). A scheduled task wakes up every day
>> and deletes all entries past the 14 day window (i.e. deletes entries from
>> the 15th day in the past).
> Not related to your vacuum problem, but if your pattern is something
> like deleting everything inserted 15 days ago you may want to think of
> using partitioning or simple inheritance. Make the scheduled task
> create a new partition/child table, redirect insertions to it, drop
> the oldest partition.
>
> Also, if the pattern is insert-only all the time, delete whole day, it
> may be easier to exclude the table from autovacuum and make the
> scheduled task vacuum the table after deleting.
>
> Francisco Olarte.

Thanks for your input. The rentention window was supposed to be variable and dynamically changeable. So, partitioning is our last resort. Will try the vacuum after delete instead of autovacuum.

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Abraham, Danny 2014-01-13 16:43:14 Re: PG 924, Windows 2012, error code 487
Previous Message Adrian Klaver 2014-01-13 15:35:31 Re: PG 924, Windows 2012, error code 487