[SPARK-41388][K8S] getReusablePVCs should ignore recently created PVCs in the previous batch
#38912
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…s in the previous batch
dongjoon-hyun
changed the title
getReusablePVCs should ignore recently created PVCs in the previous batch
dongjoon-hyun
commented
Dec 5, 2022
| .thenReturn(Seq(persistentVolumeClaim("pvc-0", "gp2", "200Gi")).asJava) | ||
| val pvc = persistentVolumeClaim("pvc-0", "gp2", "200Gi") | ||
| pvc.getMetadata | ||
| .setCreationTimestamp(Instant.now().minus(podAllocationDelay + 1, MILLIS).toString) |
Member
Author
There was a problem hiding this comment.
This is a previous test case for SPARK-35416: Support PersistentVolumeClaim Reuse.
Since our test framework does't fill CreationTimestamp, this PR added it properly.
Member
Author
|
Could you review this when you have some time, @viirya ? |
Member
Author
|
Thank you so much! |
Member
Author
|
All tests passed. Merged to master/3.3/3.2. |
dongjoon-hyun
added a commit
that referenced
this pull request
Dec 5, 2022
…VCs in the previous batch This PR aims to prevent `getReusablePVCs` from choosing recently created PVCs in the very previous batch by excluding newly created PVCs whose creation time is within `spark.kubernetes.allocation.batch.delay`. In case of slow K8s control plane situation where `spark.kubernetes.allocation.batch.delay` is too short relatively or `spark.kubernetes.executor.enablePollingWithResourceVersion=true` is used, `onNewSnapshots` may not return the full list of executor pods created by the previous batch. This sometimes makes Spark driver think the PVCs in the previous batch are reusable for the next batch. No. Pass the CIs with the newly created test case. Closes #38912 from dongjoon-hyun/SPARK-41388. Authored-by: Dongjoon Hyun <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]> (cherry picked from commit e234cd8) Signed-off-by: Dongjoon Hyun <[email protected]>
dongjoon-hyun
added a commit
that referenced
this pull request
Dec 5, 2022
…VCs in the previous batch This PR aims to prevent `getReusablePVCs` from choosing recently created PVCs in the very previous batch by excluding newly created PVCs whose creation time is within `spark.kubernetes.allocation.batch.delay`. In case of slow K8s control plane situation where `spark.kubernetes.allocation.batch.delay` is too short relatively or `spark.kubernetes.executor.enablePollingWithResourceVersion=true` is used, `onNewSnapshots` may not return the full list of executor pods created by the previous batch. This sometimes makes Spark driver think the PVCs in the previous batch are reusable for the next batch. No. Pass the CIs with the newly created test case. Closes #38912 from dongjoon-hyun/SPARK-41388. Authored-by: Dongjoon Hyun <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]> (cherry picked from commit e234cd8) Signed-off-by: Dongjoon Hyun <[email protected]> (cherry picked from commit 651f5da) Signed-off-by: Dongjoon Hyun <[email protected]>
beliefer
pushed a commit
to beliefer/spark
that referenced
this pull request
Dec 18, 2022
…VCs in the previous batch ### What changes were proposed in this pull request? This PR aims to prevent `getReusablePVCs` from choosing recently created PVCs in the very previous batch by excluding newly created PVCs whose creation time is within `spark.kubernetes.allocation.batch.delay`. ### Why are the changes needed? In case of slow K8s control plane situation where `spark.kubernetes.allocation.batch.delay` is too short relatively or `spark.kubernetes.executor.enablePollingWithResourceVersion=true` is used, `onNewSnapshots` may not return the full list of executor pods created by the previous batch. This sometimes makes Spark driver think the PVCs in the previous batch are reusable for the next batch. ### Does this PR introduce _any_ user-facing change? No. ### How was this patch tested? Pass the CIs with the newly created test case. Closes apache#38912 from dongjoon-hyun/SPARK-41388. Authored-by: Dongjoon Hyun <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]>
sunchao
pushed a commit
to sunchao/spark
that referenced
this pull request
Jun 2, 2023
…VCs in the previous batch ### What changes were proposed in this pull request? This PR aims to prevent `getReusablePVCs` from choosing recently created PVCs in the very previous batch by excluding newly created PVCs whose creation time is within `spark.kubernetes.allocation.batch.delay`. ### Why are the changes needed? In case of slow K8s control plane situation where `spark.kubernetes.allocation.batch.delay` is too short relatively or `spark.kubernetes.executor.enablePollingWithResourceVersion=true` is used, `onNewSnapshots` may not return the full list of executor pods created by the previous batch. This sometimes makes Spark driver think the PVCs in the previous batch are reusable for the next batch. ### Does this PR introduce _any_ user-facing change? No. ### How was this patch tested? Pass the CIs with the newly created test case. Closes apache#38912 from dongjoon-hyun/SPARK-41388. Authored-by: Dongjoon Hyun <[email protected]> Signed-off-by: Dongjoon Hyun <[email protected]> (cherry picked from commit e234cd8) Signed-off-by: Dongjoon Hyun <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What changes were proposed in this pull request?
This PR aims to prevent
getReusablePVCsfrom choosing recently created PVCs in the very previous batch by excluding newly created PVCs whose creation time is withinspark.kubernetes.allocation.batch.delay.Why are the changes needed?
In case of slow K8s control plane situation where
spark.kubernetes.allocation.batch.delayis too short relatively orspark.kubernetes.executor.enablePollingWithResourceVersion=trueis used,onNewSnapshotsmay not return the full list of executor pods created by the previous batch. This sometimes makes Spark driver think the PVCs in the previous batch are reusable for the next batch.Does this PR introduce any user-facing change?
No.
How was this patch tested?
Pass the CIs with the newly created test case.