This message was deleted.
# k3sa
adamant-kite-43734
06/17/2024, 11:29 PMThis message was deleted.
h
hundreds-evening-84071
06/18/2024, 1:08 AMwhat about memory resource ?
have you just tried to run k3s-killall.sh and restart k3s services ?
b
brainy-actor-53929
06/18/2024, 2:39 AMMemory seems quite high as well. Yeah, I've restarted the process and the memory/cpu shoots up again soon after. I'll try the killall.sh tomorrow, haven't seen that before.
I am pretty sure I had resource limits on my pods but maybe they weren't right.
h
hundreds-evening-84071
06/18/2024, 12:23 PMwhat version of k3s? We had similar issue but it has been several version ago (thinking 1.19 - 1.21?)
b
brainy-actor-53929
06/18/2024, 12:54 PM1.29.5
brainy-actor-53929
06/18/2024, 1:11 PMSo I deployed kube-prometheus-stack and it is using the control plane node for that workload, but it's not the containers ramming the resources it seems to be k3s-server itself.
But resource limits should keep it from OOMing right?
h
hundreds-evening-84071
06/18/2024, 1:14 PMthat is the exact same issue we used to have with k3s
even with resource limits set the nodes used to OOM and hang
b
brainy-actor-53929
06/18/2024, 1:15 PMinteresting. maybe it is Prometheus causing problems. I'll try to clean things up and make sure the limits are actually in place and report back
h
hundreds-evening-84071
06/18/2024, 1:16 PMgood luck
b
brainy-actor-53929
06/18/2024, 1:16 PMthanks!!
brainy-actor-53929
06/19/2024, 12:35 AMI think my resource limits weren't set correctly on some pods, but even on a fresh install I see k3s-server pushing 66% cpu, sometimes spiking to 80% and 25% mem. The pi clone I'm using has 4 cores and 2gb ram, it's a libre computer le-potato.