Peter Bacsko created YUNIKORN-2980:
--------------------------------------
Summary: DaemonSet preemption: don't flood the logs if victim
selection fails
Key: YUNIKORN-2980
URL: https://issues.apache.org/jira/browse/YUNIKORN-2980
Project: Apache YuniKorn
Issue Type: Sub-task
Components: core - scheduler
Reporter: Peter Bacsko
Assignee: Peter Bacsko
If we can't find a proper victim for a DaemonSet pod, we constaly print the
following to the console in every scheduling cycle:
{noformat}
log.Log(log.SchedApplication).Info("Triggering preemption process for daemon
set ask",
zap.String("ds allocation key", ask.GetAllocationKey()))
[...]
log.Log(log.SchedApplication).Warn("Problem in finding the victims for
preempting resources to meet required ask requirements",
zap.String("ds allocation key", ask.GetAllocationKey()),
zap.String("node id", reserve.nodeID))
{noformat}
Since we attempt to schedule in every 100ms, this is logged at least 10 times
in every second.
Suggestion: do the preemption silently and only log if it has succeeded, just
like regular preemption. Otherwise, use {{Allocation.LogAllocationFailure()}}
which we already do for a variety of things.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
