Re: control the number of clog files and xlog files

From: Duan Ligong <duanlg(at)nec-as(dot)nec(dot)com(dot)cn>
To: Alvaro Herrera <alvherre(at)commandprompt(dot)com>
Cc: pgsql-performance(at)postgresql(dot)org
Subject: Re: control the number of clog files and xlog files
Date: 2008-08-27 11:27:53
Views: Raw Message | Whole Thread | Download mbox
Lists: pgsql-performance

Alvaro, Thanks for your answer.
It would be very helpful.

> > Would you like to be so kind as to answer the following questions:
> >
> > - Is there any way to control the number of clog files and xlog files?
> > I encounter an issue that there are too many clog files under the
> > pg_clog/ directory which occupy more space than I can endure..
> pg_clog files are controlled by tuple freezing, which is done by vacuum,
> and it depends on the autovacuum_min_freeze_age parameter and
> vacuum_freeze_min_age. Please read

So can we reduce the number of clog by increasing the
autovacuum_min_freeze_age parameter and vacuum_freeze_min_age

> and

> > - What determines the number of clog files? what determines the
> > number of xlog files?
> The number of xlog files will depend on checkpoints. You need to
> restrict checkpoint_segments to control this. Note that this can have a
> serious performance impact.
> > - I understand pg_xlog is used to record WAL. but what is pg_clog
> > is used to? Is it used to record some meta-information on the xlog?
> clog is the "commit log", i.e. it records transactions that have been
> committed and those that have been aborted. You cannot delete files
> unless you want to corrupt your database.

Could you explain how the clog files work roughly?
(What is inside of the clog files? when and how the new clog files
are created? when and in what case the old files are deleted
or rotated? how does postgresql regard a file is old enough to be
deleted? Does Vacuum will definitely cause deleting of old files
and creating of new clog files?)

> > - What effect does Deleting the clog and xlogfiles bring about?
> > Will it cause Postgresql abnormal stopping?
> Your data will be corrupt. It may continue to work for a while, and
> suddenly stop working at a future time.

I encoutered a scenario that there are many files and some of them
are as old as one month ago. Does all these files including the
old files are still useful for postgresql? and when will they deleted
or rotated? Or should they be deleted and maintained by external

Best regards

> --
> Alvaro Herrera
> The PostgreSQL Company - Command Prompt, Inc.
> --
> Sent via pgsql-performance mailing list (pgsql-performance(at)postgresql(dot)org)
> To make changes to your subscription:

Duan Ligong
: 8-0086-22-354

In response to


Browse pgsql-performance by date

  From Date Subject
Next Message Bill Moran 2008-08-27 13:01:23 Re: select on 22 GB table causes "An I/O error occured while sending to the backend." exception
Previous Message Florian Weimer 2008-08-27 07:21:35 Re: select on 22 GB table causes "An I/O error occured while sending to the backend." exception