Restarting postgres master instance

Mads Holden's Avatar

Mads Holden

05 Sep, 2014 01:25 PM

My Postgres farm (12787) currently has one master and one slave, and is functioning correctly on AWS.
I recently received an email from Amazon stating that the instance running as master in my farm needs to be
rebooted for maintentance. What is the recommended way of rebooting a master in a postgres farm, eliminating
or at least minimizing downtime?

  1. 1 Posted by Dmitry Korsakov on 05 Sep, 2014 04:08 PM

    Dmitry Korsakov's Avatar

    Hi Mads.

    Rebooting the master node using Scalr CP should be sufficient.

    However, if something bad happens to the master, one of your slaves will be automatically promoted to master. And to ensure that promotion won't fail it is a good practice to have a fresh (taken less then an hour before promotion) data bundle.

    So, create fresh data bundle, reboot the master and let us know how it goes.

  2. 2 Posted by Mads Holden on 11 Sep, 2014 09:03 AM

    Mads Holden's Avatar

    Rebooting the instance worked fine for the database, but Amazon did not consider it rebooted, even if I did a hard reboot. What worked for me was using the AWS api to stop and start the instance. The database was of course down when the instance was, but came back with no problems.

  3. marc closed this discussion on 22 Jul, 2016 08:54 PM.

Comments are currently closed for this discussion. You can start a new one.

Keyboard shortcuts

Generic

? Show this help
ESC Blurs the current field

Comment Form

r Focus the comment reply box
^ + ↩ Submit the comment

You can use Command ⌘ instead of Control ^ on Mac