Re: [Admin] Reclaim Space from 4 TB Database

From: Rui DeSousa <rui(at)crazybean(dot)net>
To: Rahul Saha <rahul(dot)blooming(at)gmail(dot)com>
Cc: "pgsql-admin(at)postgresql(dot)org" <pgsql-admin(at)postgresql(dot)org>
Subject: Re: [Admin] Reclaim Space from 4 TB Database
Date: 2018-07-18 15:30:51
Message-ID: 70718D12-F9B1-48D1-9370-CF3720E07A05@crazybean.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin

> On Jul 18, 2018, at 6:17 AM, Rahul Saha <rahul(dot)blooming(at)gmail(dot)com> wrote:
>
> Hi Experts,
>
> Greetings !!
>
> Background - PG version 9.4
> Linux - Rhel
>
> I have a scenario in production environment where the DB size is close to 4 TB and couple of tables are 1.5TB each.
>
> Now, as the table are highly transactional therefore there is lot of bloat in tables. While I run Vaccum Full in weekend when there is no incoming connection, vacuum full never completes in 2 days.
>
> Firstly, I wanted to know what are the factor I should consider to speed up Vacuum Full process other than investing memory.
>
> Secondly are there any other option to reclaim space. (Pg_repck is a option). I am looking for more option if available.
>
> Thank you everyone.
> --
> Thanks & Regards
> Rahul Saha
>
> +91 8806351116
>
> Sent from my iPhone

If the table are highly transactional and you have auto vacuum enabled then the tables really should normalize. I would suggest looking at your auto vacuum settings and being more aggressive about them. Also consider upgrading to the latest version of Postgres as much work has been done to improve auto vacuum.

In response to

Browse pgsql-admin by date

  From Date Subject
Next Message Naveen Kumar 2018-07-18 15:45:52 ERROR: unexpected chunk number 452 (expected 0) for toast value 94674063 in pg_toast_56980977
Previous Message bala jayaram 2018-07-18 14:31:07 logical decoding cannot be used while in recovery 9.5