[
https://issues.apache.org/jira/browse/YUNIKORN-3147?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Hsien-Cheng(Ryan) Huang updated YUNIKORN-3147:
----------------------------------------------
Description:
in checking logs, using a deterministic log[0] to check.
But, the logs somehow goes in different order
err:
{code:java}
Entry 0: &{Message:Not enough queue quota LastOccurrence:2025-11-01
16:51:09.930115 +0800 CST m=+0.010604542 Count:1}
Entry 1: &{Message:Preemption max attempts exhausted LastOccurrence:2025-11-01
16:51:09.930114 +0800 CST m=+0.010603709 Count:1}
Entry 2: &{Message:Preemption does not help LastOccurrence:2025-11-01
16:51:09.930114 +0800 CST m=+0.010604251 Count:1}
{code}
normal:
{code:java}
Entry 0: &{Message:Preemption max attempts exhausted LastOccurrence:2025-11-01
16:52:48.659237 +0800 CST m=+0.011641251 Count:1}
Entry 1: &{Message:Preemption does not help LastOccurrence:2025-11-01
16:52:48.659237 +0800 CST m=+0.011641709 Count:1}
Entry 2: &{Message:Not enough queue quota LastOccurrence:2025-11-01
16:52:48.659238 +0800 CST m=+0.011641959 Count:1}
{code}
{code:java}
=== RUN TestTryAllocatePreemptQueue
2025-11-01T17:23:47.342+0800 INFO core.scheduler.queue
objects/queue.go:154 configured queue added to scheduler {"queueName":
"root"}
2025-11-01T17:23:47.343+0800 INFO core.scheduler.queue
objects/queue.go:1374 updating root queue max resources {"current max":
"nil resource", "new max": "map[first:20]"}
2025-11-01T17:23:47.343+0800 INFO core.scheduler.queue
objects/queue.go:154 configured queue added to scheduler {"queueName":
"root.parent"}
2025-11-01T17:23:47.343+0800 INFO core.scheduler.queue
objects/queue.go:154 configured queue added to scheduler {"queueName":
"root.parent.child1"}
2025-11-01T17:23:47.343+0800 INFO core.scheduler.queue
objects/queue.go:154 configured queue added to scheduler {"queueName":
"root.parent.child2"}
2025-11-01T17:23:47.343+0800 INFO core.scheduler.application
objects/application.go:163 Unknown gang scheduling style, using soft style
as default {"gang scheduling style": ""}
2025-11-01T17:23:47.343+0800 INFO core.scheduler.fsm
objects/application_state.go:147Application state transition {"appID":
"app-1", "source": "New", "destination": "Accepted", "event": "runApplication"}
2025-11-01T17:23:47.343+0800 INFO core.scheduler.application
objects/application.go:633 ask added successfully to application
{"appID": "app-1", "user": "testuser", "ask": "alloc1", "placeholder": false,
"pendingDelta": "map[first:5]"}
2025-11-01T17:23:47.343+0800 INFO core.scheduler.application
objects/application.go:633 ask added successfully to application
{"appID": "app-1", "user": "testuser", "ask": "alloc2", "placeholder": false,
"pendingDelta": "map[first:5]"}
2025-11-01T17:23:47.343+0800 INFO core.scheduler.application
objects/application.go:163 Unknown gang scheduling style, using soft style
as default {"gang scheduling style": ""}
2025-11-01T17:23:47.343+0800 INFO core.scheduler.fsm
objects/application_state.go:147Application state transition {"appID":
"app-2", "source": "New", "destination": "Accepted", "event": "runApplication"}
2025-11-01T17:23:47.343+0800 INFO core.scheduler.application
objects/application.go:633 ask added successfully to application
{"appID": "app-2", "user": "testuser", "ask": "alloc3", "placeholder": false,
"pendingDelta": "map[first:5]"}
2025-11-01T17:23:47.344+0800 INFO core.scheduler.ugm
ugm/manager.go:634 User tracker doesn't exists. Creating user tracker.
{"user": "testuser"}
2025-11-01T17:23:47.344+0800 INFO core.scheduler.ugm
ugm/manager.go:254 Group tracker set for user application {"group": "",
"user": "testuser", "application": "app-1", "queue path": "root.parent.child1"}
2025-11-01T17:23:47.344+0800 INFO core.scheduler.node
objects/node.go:429 node allocation processed {"appID": "app-1",
"allocationKey": "alloc1", "allocatedResource": "map[first:5]", "placeholder":
false, "targetNode": "node1"}
2025-11-01T17:23:47.344+0800 INFO core.scheduler.fsm
objects/application_state.go:147Application state transition {"appID":
"app-1", "source": "Accepted", "destination": "Running", "event":
"runApplication"}
2025-11-01T17:23:47.344+0800 INFO core.scheduler.node
objects/node.go:429 node allocation processed {"appID": "app-1",
"allocationKey": "alloc2", "allocatedResource": "map[first:5]", "placeholder":
false, "targetNode": "node1"}
2025-11-01T17:23:47.344+0800 INFO core.scheduler.ugm
ugm/manager.go:254 Group tracker set for user application {"group": "",
"user": "testuser", "application": "app-2", "queue path": "root.parent.child2"}
application_test.go:2246: assertion failed: Preemption does not help
(log[0].Message string) != Preemption max attempts exhausted
(common.PreemptionMaxAttemptsExhausted string)
--- FAIL: TestTryAllocatePreemptQueue (0.01s)
FAIL
FAIL github.com/apache/yunikorn-core/pkg/scheduler/objects 0.317s
FAIL
{code}
was:
in checking logs, using a deterministic log[0] to check.
But, the logs somehow goes in different order
err:
{code:java}
Entry 0: &{Message:Not enough queue quota LastOccurrence:2025-11-01
16:51:09.930115 +0800 CST m=+0.010604542 Count:1}
Entry 1: &{Message:Preemption max attempts exhausted LastOccurrence:2025-11-01
16:51:09.930114 +0800 CST m=+0.010603709 Count:1}
Entry 2: &{Message:Preemption does not help LastOccurrence:2025-11-01
16:51:09.930114 +0800 CST m=+0.010604251 Count:1}
{code}
normal:
{code:java}
Entry 0: &{Message:Preemption max attempts exhausted LastOccurrence:2025-11-01
16:52:48.659237 +0800 CST m=+0.011641251 Count:1}
Entry 1: &{Message:Preemption does not help LastOccurrence:2025-11-01
16:52:48.659237 +0800 CST m=+0.011641709 Count:1}
Entry 2: &{Message:Not enough queue quota LastOccurrence:2025-11-01
16:52:48.659238 +0800 CST m=+0.011641959 Count:1}
{code}
> Flaky test on ./pkg/scheduler/objects -run ^TestTryAllocatePreemptQueue$
> ------------------------------------------------------------------------
>
> Key: YUNIKORN-3147
> URL: https://issues.apache.org/jira/browse/YUNIKORN-3147
> Project: Apache YuniKorn
> Issue Type: Test
> Reporter: Hsien-Cheng(Ryan) Huang
> Assignee: Hsien-Cheng(Ryan) Huang
> Priority: Minor
>
> in checking logs, using a deterministic log[0] to check.
> But, the logs somehow goes in different order
> err:
> {code:java}
> Entry 0: &{Message:Not enough queue quota LastOccurrence:2025-11-01
> 16:51:09.930115 +0800 CST m=+0.010604542 Count:1}
> Entry 1: &{Message:Preemption max attempts exhausted
> LastOccurrence:2025-11-01 16:51:09.930114 +0800 CST m=+0.010603709 Count:1}
> Entry 2: &{Message:Preemption does not help LastOccurrence:2025-11-01
> 16:51:09.930114 +0800 CST m=+0.010604251 Count:1}
> {code}
> normal:
> {code:java}
> Entry 0: &{Message:Preemption max attempts exhausted
> LastOccurrence:2025-11-01 16:52:48.659237 +0800 CST m=+0.011641251 Count:1}
> Entry 1: &{Message:Preemption does not help LastOccurrence:2025-11-01
> 16:52:48.659237 +0800 CST m=+0.011641709 Count:1}
> Entry 2: &{Message:Not enough queue quota LastOccurrence:2025-11-01
> 16:52:48.659238 +0800 CST m=+0.011641959 Count:1}
> {code}
> {code:java}
> === RUN TestTryAllocatePreemptQueue
> 2025-11-01T17:23:47.342+0800 INFO core.scheduler.queue
> objects/queue.go:154 configured queue added to scheduler {"queueName":
> "root"}
> 2025-11-01T17:23:47.343+0800 INFO core.scheduler.queue
> objects/queue.go:1374 updating root queue max resources {"current
> max": "nil resource", "new max": "map[first:20]"}
> 2025-11-01T17:23:47.343+0800 INFO core.scheduler.queue
> objects/queue.go:154 configured queue added to scheduler {"queueName":
> "root.parent"}
> 2025-11-01T17:23:47.343+0800 INFO core.scheduler.queue
> objects/queue.go:154 configured queue added to scheduler {"queueName":
> "root.parent.child1"}
> 2025-11-01T17:23:47.343+0800 INFO core.scheduler.queue
> objects/queue.go:154 configured queue added to scheduler {"queueName":
> "root.parent.child2"}
> 2025-11-01T17:23:47.343+0800 INFO core.scheduler.application
> objects/application.go:163 Unknown gang scheduling style, using soft
> style as default {"gang scheduling style": ""}
> 2025-11-01T17:23:47.343+0800 INFO core.scheduler.fsm
> objects/application_state.go:147Application state transition {"appID":
> "app-1", "source": "New", "destination": "Accepted", "event":
> "runApplication"}
> 2025-11-01T17:23:47.343+0800 INFO core.scheduler.application
> objects/application.go:633 ask added successfully to application
> {"appID": "app-1", "user": "testuser", "ask": "alloc1", "placeholder": false,
> "pendingDelta": "map[first:5]"}
> 2025-11-01T17:23:47.343+0800 INFO core.scheduler.application
> objects/application.go:633 ask added successfully to application
> {"appID": "app-1", "user": "testuser", "ask": "alloc2", "placeholder": false,
> "pendingDelta": "map[first:5]"}
> 2025-11-01T17:23:47.343+0800 INFO core.scheduler.application
> objects/application.go:163 Unknown gang scheduling style, using soft
> style as default {"gang scheduling style": ""}
> 2025-11-01T17:23:47.343+0800 INFO core.scheduler.fsm
> objects/application_state.go:147Application state transition {"appID":
> "app-2", "source": "New", "destination": "Accepted", "event":
> "runApplication"}
> 2025-11-01T17:23:47.343+0800 INFO core.scheduler.application
> objects/application.go:633 ask added successfully to application
> {"appID": "app-2", "user": "testuser", "ask": "alloc3", "placeholder": false,
> "pendingDelta": "map[first:5]"}
> 2025-11-01T17:23:47.344+0800 INFO core.scheduler.ugm
> ugm/manager.go:634 User tracker doesn't exists. Creating user tracker.
> {"user": "testuser"}
> 2025-11-01T17:23:47.344+0800 INFO core.scheduler.ugm
> ugm/manager.go:254 Group tracker set for user application {"group": "",
> "user": "testuser", "application": "app-1", "queue path":
> "root.parent.child1"}
> 2025-11-01T17:23:47.344+0800 INFO core.scheduler.node
> objects/node.go:429 node allocation processed {"appID": "app-1",
> "allocationKey": "alloc1", "allocatedResource": "map[first:5]",
> "placeholder": false, "targetNode": "node1"}
> 2025-11-01T17:23:47.344+0800 INFO core.scheduler.fsm
> objects/application_state.go:147Application state transition {"appID":
> "app-1", "source": "Accepted", "destination": "Running", "event":
> "runApplication"}
> 2025-11-01T17:23:47.344+0800 INFO core.scheduler.node
> objects/node.go:429 node allocation processed {"appID": "app-1",
> "allocationKey": "alloc2", "allocatedResource": "map[first:5]",
> "placeholder": false, "targetNode": "node1"}
> 2025-11-01T17:23:47.344+0800 INFO core.scheduler.ugm
> ugm/manager.go:254 Group tracker set for user application {"group": "",
> "user": "testuser", "application": "app-2", "queue path":
> "root.parent.child2"}
> application_test.go:2246: assertion failed: Preemption does not help
> (log[0].Message string) != Preemption max attempts exhausted
> (common.PreemptionMaxAttemptsExhausted string)
> --- FAIL: TestTryAllocatePreemptQueue (0.01s)
> FAIL
> FAIL github.com/apache/yunikorn-core/pkg/scheduler/objects 0.317s
> FAIL
> {code}
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
