Release semaphore immediately after startup process exit #6463
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
pan3793
changed the title
|
The idea makes sense, and a similar mechanism was applied to batch submission too #6028 |
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #6463 +/- ##
======================================
Coverage 0.00% 0.00%
======================================
Files 675 675
Lines 41641 41643 +2
Branches 5685 5685
======================================
- Misses 41641 41643 +2 ☔ View full report in Codecov by Sentry. |
cxzl25
approved these changes
Jun 13, 2024
pan3793
reviewed
Jun 13, 2024
kyuubi-server/src/main/scala/org/apache/kyuubi/engine/EngineRef.scala
Outdated
Show resolved
Hide resolved
pan3793
pushed a commit
that referenced
this pull request
Jun 13, 2024
# 🔍 Description ## Issue References 🔗 The concurrency limit for the engine startup process is mainly used to avoid overload on the machine(or container) of the Kyuubi server, the current implementation holds startupProcessSemaphore until the session is established successfully. While for Spark on YARN cluster mode, some YARN queue resource insufficiency may block the subsequent Spark application submissions to other queues, significantly affecting the Kyuubi server's resource utilization. ## Describe Your Solution 🔧 We should immediately release the `startupProcessSemaphore` after the engine startup process exits (i.e., after the `spark-submit` process exits) as the load has already disappeared. ## Types of changes 🔖 - [x] Bugfix (non-breaking change which fixes an issue) - [ ] New feature (non-breaking change which adds functionality) - [ ] Breaking change (fix or feature that would cause existing functionality to change) ## Test Plan 🧪 I tested it on a cluster of 50 kyuubi Servers, and kyuubi server resource utilization increased by 70% --- # Checklist 📝 - [ ] This patch was not authored or co-authored using [Generative Tooling](https://www.apache.org/legal/generative-tooling.html) **Be nice. Be informative.** Closes #6463 from ic4y/master-p003. Closes #6463 f7de68c [ic4y] Improve code quality d8b0248 [ic4y] [Improve][EngineRef] Optimize Engine Startup Concurrency Limit Authored-by: ic4y <ic4y@apache.org> Signed-off-by: Cheng Pan <chengpan@apache.org> (cherry picked from commit 95ed748) Signed-off-by: Cheng Pan <chengpan@apache.org>
pan3793
pushed a commit
that referenced
this pull request
Jun 13, 2024
The concurrency limit for the engine startup process is mainly used to avoid overload on the machine(or container) of the Kyuubi server, the current implementation holds startupProcessSemaphore until the session is established successfully. While for Spark on YARN cluster mode, some YARN queue resource insufficiency may block the subsequent Spark application submissions to other queues, significantly affecting the Kyuubi server's resource utilization. We should immediately release the `startupProcessSemaphore` after the engine startup process exits (i.e., after the `spark-submit` process exits) as the load has already disappeared. - [x] Bugfix (non-breaking change which fixes an issue) - [ ] New feature (non-breaking change which adds functionality) - [ ] Breaking change (fix or feature that would cause existing functionality to change) I tested it on a cluster of 50 kyuubi Servers, and kyuubi server resource utilization increased by 70% --- - [ ] This patch was not authored or co-authored using [Generative Tooling](https://www.apache.org/legal/generative-tooling.html) **Be nice. Be informative.** Closes #6463 from ic4y/master-p003. Closes #6463 f7de68c [ic4y] Improve code quality d8b0248 [ic4y] [Improve][EngineRef] Optimize Engine Startup Concurrency Limit Authored-by: ic4y <ic4y@apache.org> Signed-off-by: Cheng Pan <chengpan@apache.org> (cherry picked from commit 95ed748) Signed-off-by: Cheng Pan <chengpan@apache.org>
|
Thanks, merged to master/1.9.2/1.8.3 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
🔍 Description
Issue References 🔗
The concurrency limit for the engine startup process is mainly used to avoid overload on the machine(or container) of the Kyuubi server, the current implementation holds startupProcessSemaphore until the session is established successfully. While for Spark on YARN cluster mode, some YARN queue resource insufficiency may block the subsequent Spark application submissions to other queues, significantly affecting the Kyuubi server's resource utilization.
Describe Your Solution 🔧
We should immediately release the
startupProcessSemaphoreafter the engine startup process exits (i.e., after thespark-submitprocess exits) as the load has already disappeared.Types of changes 🔖
Test Plan 🧪
I tested it on a cluster of 50 kyuubi Servers, and kyuubi server resource utilization increased by 70%
Checklist 📝
Be nice. Be informative.