Re: pg_xlog and the WAL files

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: "Achakzai, Omar" <omar(dot)achakzai(at)logicacmg(dot)com>
Cc: pgsql-admin(at)postgresql(dot)org
Subject: Re: pg_xlog and the WAL files
Date: 2007-02-16 15:26:26
Message-ID: 27605.1171639586@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin

"Achakzai, Omar" <omar(dot)achakzai(at)logicacmg(dot)com> writes:
> 1. What happens when the file system on which the pg_xlog directly resides,=
> completely fills up? Does PostgreSQL wait until free space is available ag=
> ain?=20

I believe it'll PANIC and shut down. You won't lose any committed
transactions ... but transactions in progress are out of luck.

> 2. Our PostgreSQL database (version 8.1.4.) generates almost 20 GB of WAL f=
> iles per hour! What are the best practices related to administering large a=
> mounts of WAL files? Would it be wise to compress the WAL files?=20

What are your checkpoint settings? Reducing checkpoint frequency is one
way to cut the WAL volume.

regards, tom lane

In response to

Browse pgsql-admin by date

  From Date Subject
Next Message Ezequias Rodrigues da Rocha 2007-02-16 18:59:11 Debug (is it PostgreSQL?)
Previous Message Andy Shellam (Mailing Lists) 2007-02-16 15:07:21 Re: WAL files backup