fleetlock Wait for pod evictions when draining, up to 60 seconds.

The drainer attempts to evict pods but doesn't wait at all (?) for evictions to complete which does not allow time for graceful termination of pods that need it.

Sep 26 '24 01:09 vadimberezniker

Is there any appetite for this change? Draining nodes w/o waiting at all for pods to be evicted isn't ideal behavior. This can lead to ungraceful pod termination if the host is rebooted after the drain.

Oct 08 '24 18:10 vadimberezniker

This would solve a huge headache of ours as fleetlock shuts down our nodes before our database pods gracefully shuts down, forcing an ugly failover.

@dghubble, sorry for the ping but is this feature something that can be considered?

Sep 08 '25 12:09 wolffberg

FWIW we've been running with this change in production for almost a year now

Sep 08 '25 23:09 vadimberezniker

Thanks @vadimberezniker, we made a fork with the commits from this PR and are very happy with the results 🎉

It really should be merged in the official provider but a fix is a fix I guess.

Oct 23 '25 08:10 wolffberg