Setting Max Unavailable to 1 still keeps new failed containers

Rancher 2.x
1

I have the Max Unavailable setting set to 1, yet there are still many Failed containers that persist

Screenshot%20from%202018-09-13%2001-12-27
Screenshot%20from%202018-09-13%2001-12-271251Ɨ505 36.2 KB

Screenshot%20from%202018-09-13%2001-12-12
Screenshot%20from%202018-09-13%2001-12-121386Ɨ1155 150 KB

2

Max Unavailable has to do with how many pods are allowed to be unavailable during an upgrade, so that you maintain a minimum amount of available pods during the upgrade. It has nothing to do with the garbage collection of old/failed pods.

3

Is there a way to tell it to clean up the Failed pods?

4

https://kubernetes.io/docs/concepts/cluster-administration/kubelet-garbage-collection/

1 Like
5

Ah I see, maximum-dead-containers is default to -1

Iā€™m guessing the thought was if there was something important in the dead containers to keep?