| From: | Laurenz Albe <laurenz(dot)albe(at)cybertec(dot)at> | 
|---|---|
| To: | Prince Pathria <prince(dot)pathria(at)goevive(dot)com>, pgsql-admin(at)postgresql(dot)org | 
| Subject: | Re: Archive clean up command in recovery.conf | 
| Date: | 2018-09-24 06:22:04 | 
| Message-ID: | 2779539bc6e5413b71064f92b44e2a7f6bdfcde0.camel@cybertec.at | 
| Views: | Whole Thread | Raw Message | Download mbox | Resend email | 
| Thread: | |
| Lists: | pgsql-admin | 
Prince Pathria wrote:
> I have a doubt, assume that I have 3 psql10 nodes (n1, n2, n3)
> and n1 is master ,n2+n3 are standbys. In n1 archive mode is on,
> and wal_logs are going to volume which is shared with n2 and n3.
> If archive_cleanup_command = 'pg_archivecleanup /path/to/nfs %r'
> is enabled on both n2 and n3.
> 
> Can a situation might arise that a wal segment required by n2
> is already cleaned up by n3?
Yes, that pg_archivecleanup is a bad idea, because the faster
standby might remove files that the slower one still needs.
> If yes, then how to mitigate it?
Use streaming replication and replication slots.
Yours,
Laurenz Albe
| From | Date | Subject | |
|---|---|---|---|
| Next Message | CONVERS Yann - DREAL Auvergne-Rhône-Alpes/CIDDAE/SIG | 2018-09-24 07:47:11 | Tr: Re: WAL segments held for replcation | 
| Previous Message | Laurenz Albe | 2018-09-24 06:19:53 | Re: WAL segments held for replcation |