Hi, my executor pods are getting oomkilled. But not sure whether java application within executor pod is OOM killed or POD itself is oomkilled. I tried to add GC to executor extra java options and heap dump also, but whenever pod dies no heap dump is seen. So could you please help to identify whether pod is having issue or application. One more thing the executor pods are removed so I don't have any chance to fetch the logs. But such constant restarts of executor pods is making the cluster unstable. Could you please help?
This is all stuff you could do trivially on Mesos years ago with an ounce of Spot market tooling. :)
Mesos isn't dead. It's just taking a break.
Very useful session!
Thank you very much! A load of knowledge transported, and I think I have much to research now...
Great session!!
You need to supply overhead memory explicitly unless YARN case
Hi, my executor pods are getting oomkilled. But not sure whether java application within executor pod is OOM killed or POD itself is oomkilled. I tried to add GC to executor extra java options and heap dump also, but whenever pod dies no heap dump is seen. So could you please help to identify whether pod is having issue or application. One more thing the executor pods are removed so I don't have any chance to fetch the logs. But such constant restarts of executor pods is making the cluster unstable. Could you please help?
Jill in