This message was deleted.
# generala
adamant-kite-43734
03/12/2024, 9:16 PMThis message was deleted.
a
abundant-hair-58573
03/12/2024, 9:31 PMI bumped one of my controlplanes up to 64G RAM and the kube-apiserver consumed all of it until the box became unresponsive
h
hundreds-evening-84071
03/12/2024, 9:58 PMWhich distribution? K3s, RKE2, RKE? And what kubernetes version?
I've noticed OOM before on K3s (on RHEL) on older release prior to ver 1.24.
hundreds-evening-84071
03/12/2024, 10:00 PMyou may want to deploy monitoring and look at grafana dashboard to see what exactly is causing OOM
a
abundant-hair-58573
03/12/2024, 10:00 PMRKE1 v1.19. Ancient I know, this is our legacy cluster that we're decomissioning in a month or two, but we still have lots of production folks on it. Increaseing all 5 control planes to 64G seems to have gotten things moving. I can at least get into the cluster now
abundant-hair-58573
03/12/2024, 10:01 PMI was watching top on a node right before it crash and saw kube-apiserver grab 97% of the memory lol