Hello
Running Cluster in non-sharding mode and getting always unclean shutdown.
Seems terminationGracePeriodSeconds needs to be set in order delay pod shutdown.
Thnaks
John
PS: why dont you provide all the kubernetes params by default ?
Hello
Running Cluster in non-sharding mode and getting always unclean shutdown.
Seems terminationGracePeriodSeconds needs to be set in order delay pod shutdown.
Thnaks
John
PS: why dont you provide all the kubernetes params by default ?
Hello @jamoser ,
seems right now you cannot set the grace period with CR.
as for
PS: why dont you provide all the kubernetes params by default ?
well, it is tricky.
termnationGracePeriodSeconds
is a container property, but there are others which should land under StatefulSet or Pod or something else. x-kubernetes-preserve-unknown-fields: true
, but not sure yet how does it work. Will check.Will you be interested in sending a quick PR for fixing it?
Any guide how to do a PR ?
Hey @jamoser!
Short answer - just submit a PR at GitHub - percona/percona-server-mongodb-operator: Percona Operator for MongoDB
Sorry - not having access to JIRA nor github.
termnationGracePeriodSeconds
is an attribute of Statefulset of the original MongoDB distro. I assume Percona Replicaset is also based on a Statefulset … so a pass through is required.
And this param is quite important - otherwise a restart/reschedule of the pods will cause a crash of the mongodb which can lead to a very long recovery. Even worse it can totally crash the cluster if liveness probe is not configured properly.
Regards
John
Bump … would be quite important to make this work.
I added it into our backlog: [K8SPSMDB-927] Pass termnationGracePeriodSeconds to RS containers - Percona JIRA
If you still want to submit PR, it has high chances to be included in the next release
I can only offer beer …
Unanswered | Unsolved | Solved
MySQL, InnoDB, MariaDB and MongoDB are trademarks of their respective owners.
Copyright © 2006 - 2024 Percona LLC. All rights reserved.