cockroach icon indicating copy to clipboard operation
cockroach copied to clipboard

sqlliveness: add timeouts to heartbeats

Open aadityasondhi opened this issue 2 years ago • 1 comments

Previously, sqlliveness heartbeat operations could block on the transactions that were involved. This change introduces some timeouts of the length of the heartbeat during the create and refresh operations.

Resolves #85541

Release note: None

Release justification: low-risk bugfix to existing functionality

aadityasondhi avatar Sep 07 '22 20:09 aadityasondhi

This change is Reviewable

cockroach-teamcity avatar Sep 07 '22 20:09 cockroach-teamcity

bors r=ajwerner

aadityasondhi avatar Sep 23 '22 17:09 aadityasondhi

Build succeeded:

craig[bot] avatar Sep 23 '22 18:09 craig[bot]

Thank you!

Can we get a 22.1 backport? CC serverless is running 22.1. Major upgrade can take some time in practice. The lack of these HBs has led to inactionable alerts hitting the pagers of our SREs.

joshimhoff avatar Sep 26 '22 12:09 joshimhoff

Why do major upgrades lead to unavailability? I feel like I was missing some context on the motivation here.

ajwerner avatar Sep 26 '22 14:09 ajwerner

Ah sorry for confusion!

Major upgrades don't lead to unavailability. The issue fixed by this PR has nothing to do with major upgrades. CC serverless is simply running 22.1, and I expect it'll take some time to move to 22.2, based on past experience with serverless major upgrades. As a result, I'd love it if we backport this change into 22.1, not just 22.2, so as to avoid reoccurrence of the issue fixed by this PR.

joshimhoff avatar Sep 26 '22 14:09 joshimhoff