Re: archived WALL files question

[Scott Mead <scott.lists@xxxxxxxxxxxxxxxx>]

On Thu, Apr 15, 2010 at 11:31 AM, Renato Oliveira <renato.oliveira@xxxxxxxxxxx> wrote:

Dear all,

 

I was reading again the documentation... “The archive command should generally de designed to refuse to overwrite any pre-existing archive file.”

This means it will keep writing logs to the folder specified forever, and without an intervention, the media will run out of space.

 

What do you guys do with regards to this situation, for example:

How to you clean up the old archived logs?

 

For example:

you archive your log files from your main Postgres server to a folder /mnt/pitr for example

You set your standby to pick the logs from /mnt/pitr, then it archives each log as it comes.

/mnt/pitr will fill up very quickly and run out of space if we don’t have a process to DELETE/ARCHIEVE older logs.

 

I guess the process which picks up the logs for the standby server, needs to take care of the logs, by deleting the older ones or by archiving them permanently?

Depends on what it is you're trying to accomplish:

*) PITR slave server constantly applying logs

   If all you want is a server to constantly apply the logs and you don't care about them afterwards, look into the '%r' macro in pg_standby.  It will automatically archive files for you -- Of course, your standby instance needs to have write access to the /mnt/pitr folder to delete from.

   If you are using the archive_command to copy files in the /mnt/pitr directory, and then doing a cron based copy to a backup server, have your cronjob delete files from the primary after it is confirmed that the logs got shipped safely to the backup.

*) Backup retention time

   If you're trying to keep logs around so that you can do a point in time recovery with old backups, you want to figure your retention times and determine your RTO (http://en.wikipedia.org/wiki/Recovery_time_objective).

     If you need to be able to recovery to any point in time for the past 1 week with a low RTO, then you want to keep that week's worth of logs uncompressed and available.  Anything beyond that, use a cron job to compress the logs (they usually compress pretty well based on your data).  

     Basically, you need to keep all the low-RTO required logs around so that you can quickly get at them.  If you don't have any low RTO requirements and you just want to keep a few weeks worth of data around, I would recommend that you add a few lines of code to the end of your backup job to compress (or you could delete if you don't want them) all the logs prior to the backup that you are taking.

Hope this helps

--Scott

 

 

How do you guys deal with this problem?

 

Thank you very much in advance

 

Best regards

 

Renato

 

 

 

Renato Oliveira
Systems Administrator
e-mail: renato.oliveira@xxxxxxxxxxx

 

Tel: +44 (0)1763 260811
Fax: +44 (0)1763 262410
www.grant.co.uk

 

Grant Instruments (Cambridge) Ltd
 
Company registered in England, registration number 658133
 
Registered office address:
29 Station Road,
Shepreth,
CAMBS SG8 6GB
UK

 

 

 

 

 

P Please consider the environment before printing this email

CONFIDENTIALITY: The information in this e-mail and any attachments is confidential. It is intended only for the named recipients(s). If you are not the named recipient please notify the sender immediately and do not disclose the contents to another person or take copies.

 

VIRUSES: The contents of this e-mail or attachment(s) may contain viruses which could damage your own computer system. Whilst Grant Instruments (Cambridge) Ltd has taken every reasonable precaution to minimise this risk, we cannot accept liability for any damage which you sustain as a result of software viruses. You should therefore carry out your own virus checks before opening the attachment(s).

 

OpenXML: For information about the OpenXML file format in use within Grant Instruments please visit our website