Deployment Architecture

Validating Cluster Bundle incorrectly labeling bundle as not requiring a restart

Communicator

I just stood up a brand new multi-site cluster on Splunk 7.2.4 and while I was pushing out the very first of my configurations I validated the bundle just to see if there were any new features and I noticed that the validation said that a restart was not needed, which I knew to be untrue because I was pushing out my volume definitions.

Sure enough, when I pushed it the cluster did a rolling restart.

Why would the validation process incorrectly flag a bundle as not needing a restart? This could have been a mess if it had been a live environment and I want to make sure I know how to account for it in the future.

0 Karma

Motivator

This one has bitten me previously as well, and I'm guessing you are using the CLI.

In order to check for restart, you have to explicitly tell Splunk to do so, which is frustrating.

splunk validate cluster-bundle --check-restart
0 Karma