This message was deleted.
# harvester
a
This message was deleted.
r
Hi @quaint-alarm-7893, may I ask where did you see those OOM errors? Also, if there’s a support bundle, we can look further into it 🙏
q
i got them off the dmsg logs from one of the nodes.
i dont have a support bundle though 😞 i've been dealing w/ stability issues all day too (from a power outage) so not sure it would be helpful w/ all the other stuff i'm dealing with too
r
yeah, it’s definitely better than nothing. You can follow the step here to generate one if you have time 🙂 wish you bring them back to normal soon
here's the bundle. the issue was harvester-01 node
👍 1
r
it was fluentd eating up memory
you mentioned you just experienced a power outage. is it possible that lots of logs flow in and the output place is still not ready?
q
i have nothing custom setup w/ it. so whatever it is, its all out-of-the-box stuff.
all of the alert / monitoring is all default still.
and i get that ruby error periodically.
but i didnt most-recently get it after a recent reboot. (power outage was this morning, like 12hrs ago, but i rebooted again about 2hrs ago. that's when i logged the ruby error)
but again, i see it from time to time. is there something i can/should to do prevent it from blowing up like that? (oom)
r
If you’re seeing the issue happen from time to time, maybe it’s better to raise the memory limit of fluentd Pod. Here’s the steps about it https://docs.harvesterhci.io/v1.1/logging/harvester-logging#configuring-log-resources
q
Thanks.
106 Views