[pgpool-general: 8207] Re: Cleanup of WAL archive

Jon SCHEWE jon.schewe at raytheon.com
Fri Jun 3 04:24:59 JST 2022

> Sorry for the late response.
> > I have 3 postgresql servers in a cluster behind pgpool. I'm using streaming replication. I'm finding that the WAL archive directory on the primary server grows without bound. I would expect the WAL files to be removed once the 2 replicas have consumed the files. How are others handling this?
> >
> > Doing more reading on this and looking at the scripts that are used by pgpool, it's not clear to me that I need to be archiving WAL files. If I'm using streaming replication, then the backup should always be up to date and due to the pg_basebackup call in the follow_master.sh that should keep everything updated, correct? Or are the archived WAL files there to handle the time between when the current primary server hands off to the standby server? If so, how are the archived WAL files supposed to be cleaned up?
> If you are using replication slot, you can disable "archive_mode".
> The sample scripts are just an example to help you easily setup a cluster.
> They may need to be customized to meet your specific requirements.
> If you wish to enable "archive_mode", you should periodically delete the old WALs.

If I'm using streaming replication, under what circumstances would I want "archive_mode" enabled?

Jon Schewe

More information about the pgpool-general mailing list