[
https://issues.apache.org/jira/browse/YUNIKORN-3243?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Shubham Mishra updated YUNIKORN-3243:
-------------------------------------
Description:
When two sibling queues under the same parent have vastly different guaranteed
resources (e.g., 3600:1 ratio), the fair-share queue sorting in {{TryAllocate}}
causes the smaller queue to be completely starved — its {{app.tryAllocate()}}
is never called. This has two consequences:
# {*}Scheduling starvation{*}: The smaller queue's asks are never evaluated
for allocation, even when nodes have capacity.
# {*}Autoscaler blindness{*}: Because {{SetSchedulingAttempted(true)}} is only
set inside
{{app.[tryAllocate|https://github.com/apache/yunikorn-core/blob/master/pkg/scheduler/objects/application.go#L1029]()}}
(line 1035 of {{{}application.go{}}}), the starved queue's asks never get this
flag. {{inspectOutstandingRequests}} skips them, so the cluster autoscaler
(e.g., Karpenter) is never notified that capacity is needed.
The second issue is the more critical one — even if scheduling is delayed, the
autoscaler should be able to provision nodes in parallel. But with the current
design, the autoscaler signal is gated on queue visitation.
h3.
was:
When two sibling queues under the same parent have vastly different guaranteed
resources (e.g., 3600:1 ratio), the fair-share queue sorting in {{TryAllocate}}
causes the smaller queue to be completely starved — its {{app.tryAllocate()}}
is never called. This has two consequences:
# {*}Scheduling starvation{*}: The smaller queue's asks are never evaluated
for allocation, even when nodes have capacity.
# {*}Autoscaler blindness{*}: Because {{SetSchedulingAttempted(true)}} is only
set inside {{app.tryAllocate()}} (line 1035 of {{{}application.go{}}}), the
starved queue's asks never get this flag. {{inspectOutstandingRequests}} skips
them, so the cluster autoscaler (e.g., Karpenter) is never notified that
capacity is needed.
The second issue is the more critical one — even if scheduling is delayed, the
autoscaler should be able to provision nodes in parallel. But with the current
design, the autoscaler signal is gated on queue visitation.
h3.
> Fair-share queue sorting causes starvation of sibling queues with asymmetric
> guaranteed resources
> -------------------------------------------------------------------------------------------------
>
> Key: YUNIKORN-3243
> URL: https://issues.apache.org/jira/browse/YUNIKORN-3243
> Project: Apache YuniKorn
> Issue Type: Bug
> Reporter: Shubham Mishra
> Assignee: Shubham Mishra
> Priority: Major
>
> When two sibling queues under the same parent have vastly different
> guaranteed resources (e.g., 3600:1 ratio), the fair-share queue sorting in
> {{TryAllocate}} causes the smaller queue to be completely starved — its
> {{app.tryAllocate()}} is never called. This has two consequences:
> # {*}Scheduling starvation{*}: The smaller queue's asks are never evaluated
> for allocation, even when nodes have capacity.
> # {*}Autoscaler blindness{*}: Because {{SetSchedulingAttempted(true)}} is
> only set inside
> {{app.[tryAllocate|https://github.com/apache/yunikorn-core/blob/master/pkg/scheduler/objects/application.go#L1029]()}}
> (line 1035 of {{{}application.go{}}}), the starved queue's asks never get
> this flag. {{inspectOutstandingRequests}} skips them, so the cluster
> autoscaler (e.g., Karpenter) is never notified that capacity is needed.
> The second issue is the more critical one — even if scheduling is delayed,
> the autoscaler should be able to provision nodes in parallel. But with the
> current design, the autoscaler signal is gated on queue visitation.
> h3.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
