Nimble's Peer Persistence and failures fixups

[RimvydasLT]

Hi,

So lets imagine that I have two Nimbles and peer persistence configured. Lets imagine that GL and witness went down. Docs are saying that BGL will stop any IO activity. So my questions:

1. Is it possible somehow to bring volumes up on BGL in such scenario. And if the answer is yes, what steps must be done to achieve this.

2. Lets imagine that we've brought the volumes somehow up on BGL. I'm interested, what will happen if the witness and old GL will show up

[Thomas_Lam_HK]

Please login to InfoSIght and search for "KB-000372" under Knowledge Base Article, it contains all the recovery procedures for both questions.

[RimvydasLT]

Hi,

Sorry, but this article is talking about different thing and not what I'm asking about. Please, read my post one more time with more attention.

[Sheldon Smith]

From what I read, it IS what you're looking for.

"Scenario #2: Group Leader Unreachable - ASO not enabled"
covers what you are asking in your first question.

"Scenario #2: Recovery Steps"
is how to recover once the witness and GL are back up.

[RimvydasLT]

Hi,

Maybe my initial post was not clear enough then:( I was talking about scenario with ASO enabled. Or must I understand that and in this case I must follow steps mentioned in scenario 2 and nothing more?

[Sheldon Smith]

Your initial post was with ASO disabled -- witness down. Follow "Scenario #2: Recovery Steps".

As always, when in doubt, contact Nimble Support.