Cascading Standby WALs Question

From: Don Seiler <don(at)seiler(dot)us>
To: pgsql-admin <pgsql-admin(at)postgresql(dot)org>
Subject: Cascading Standby WALs Question
Date: 2017-09-06 21:07:25
Message-ID: CAHJZqBDDQMBZMJ9=z2RUzDDPBuXYZ+KBpz=oOvMi29gNkusOzA@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-admin

PostgreSQL 9.2 (I know it's EOL this month).

Current setup is one master and one standby. Standby does streaming
replication but the master also rsyncs WAL files to the standby DR site.
However the WAL file storage on the master is on our NetApp which is
mirrored to the standby DR site as well. The standby has
archive_cleanup_command configured and purges WAL files that it no longer
needs.

Looking to turn the one standby into a cascading standby. However I'd also
like to just use the NetApp mirror to make the WAL files available in the
event that streaming replication goes offline (and cut out the redundant
rsync traffic). My thinking is to do this:

- Mount NetApp mirror read-only on all DR standbys (both cascading and
downstream)
- Configure restore_command to point to this mount on all DR standbys
- Comment out the archive_cleanup_command since we don't want them to
delete WAL files anymore (the NetApp keeps 7 days of WAL files, managed by
a cleanup script on the primary).
- I don't believe archive_command is in play on 9.2 standbys/replicas,
but I assume I would want to disable that as well in this case?

Thanks,
Don.

--
Don Seiler
www.seiler.us

Browse pgsql-admin by date

  From Date Subject
Next Message Sunil N Shinde 2017-09-08 05:49:30 postgres streaming replication for HA
Previous Message Jerry Sievers 2017-09-06 15:37:51 Re: Problem with transaction ID wraparound