Use workaround to fix high CPU usage by LinkedTransferQueue #3756
Conversation
This file contains 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
This is a workaround for JDK-8301341 by using the Java 11 LinkedTransferQueue with the QueueingThreadPoolExecutor in the Core and jUPnP. Signed-off-by: Wouter Born <[email protected]>
If you grant me owner rights on your repo, I can transfer it to the openHAB org. |
|
I can only add collaborators so I asked to transfer it to you so you can transfer it to openHAB and give me back permissions. 🙂 |
|
Done: https://github.com/openhab/openhab-base-fixes |
|
Great! Are the GAV parameters OK? If so, I can update the POM for the repo change and add a GitHub release. |
Fine with me. |
|
I've build the artifacts and added them to these releases:
So if you can upload those we can see if the build succeeds. |
kaikreuzer
added
rebuild
|
Thanks, I have uploaded both to Artifactory. |
wborn
added a commit
to wborn/openhab-addons
that referenced
this pull request
Aug 13, 2023
Related to openhab/openhab-core#3756 Signed-off-by: Wouter Born <[email protected]>
wborn
added a commit
to wborn/openhab-distro
that referenced
this pull request
Aug 13, 2023
Related to openhab/openhab-core#3756 Signed-off-by: Wouter Born <[email protected]>
|
It seems to build pretty well except for #3257 acting up again. |
kaikreuzer
pushed a commit
to openhab/openhab-addons
that referenced
this pull request
Aug 14, 2023
* Resolve runbundles for workaround Related to openhab/openhab-core#3756 Signed-off-by: Wouter Born <[email protected]>
kaikreuzer
pushed a commit
to openhab/openhab-distro
that referenced
this pull request
Aug 14, 2023
Related to openhab/openhab-core#3756 Signed-off-by: Wouter Born <[email protected]>
wborn
added a commit
that referenced
this pull request
Aug 14, 2023
This is a workaround for JDK-8301341 by using the Java 11 LinkedTransferQueue with the QueueingThreadPoolExecutor in the Core and jUPnP. Signed-off-by: Wouter Born <[email protected]>
wborn
added a commit
to openhab/openhab-addons
that referenced
this pull request
Aug 14, 2023
* Resolve runbundles for workaround Related to openhab/openhab-core#3756 Signed-off-by: Wouter Born <[email protected]>
|
This pull request has been mentioned on openHAB Community. There might be relevant details there: https://community.openhab.org/t/openhab-4-1-milestone-discussion/149502/32 |
austvik
pushed a commit
to austvik/openhab-addons
that referenced
this pull request
Mar 27, 2024
* Resolve runbundles for workaround Related to openhab/openhab-core#3756 Signed-off-by: Wouter Born <[email protected]> Signed-off-by: Jørgen Austvik <[email protected]>
minwoox
added a commit
to minwoox/centraldogma
that referenced
this pull request
Apr 22, 2024
…Executor Motivation: The use of `LinkedTransferQueue` in JDK 17 has a bug causing a thread to consume all of a CPU's usage: https://bugs.openjdk.org/browse/JDK-8301341. Although this bug has been fixed, the patch has not been backported to JDK 17. Since copying the code from JDK 11 poses licensing issues (openhab/openhab-core#3756), we need a different solution. Switching to `LinkedBlockingQueue` offers a feasible workaround. While `LinkedTransferQueue` is typically efficient under high multi-thread contention due to its use of CAS, our current usage does not exhibit significant contention as the queue tasks' size remains close to 0. Modifications: - Replaced the use of `LinkedTransferQueue` with `LinkedBlockingQueue` in `ThreadPoolExecutor`. Result: - You no longer see an excessive CPU usage by a single thread.
minwoox
added a commit
to line/centraldogma
that referenced
this pull request
Apr 23, 2024
…Executor (#946) Motivation: The use of `LinkedTransferQueue` in JDK 17 has a bug causing a thread to consume all of a CPU's usage: https://bugs.openjdk.org/browse/JDK-8301341. Although this bug has been fixed, the patch has not been backported to JDK 17. Since copying the code from JDK 11 poses licensing issues (openhab/openhab-core#3756), we need a different solution. Switching to `LinkedBlockingQueue` offers a feasible workaround. While `LinkedTransferQueue` is typically efficient under high multi-thread contention due to its use of CAS, our current usage does not exhibit significant contention as the queue tasks' size remains close to 0. Modifications: - Replaced the use of `LinkedTransferQueue` with `LinkedBlockingQueue` in `ThreadPoolExecutor`. Result: - You no longer see an excessive CPU usage by a single thread.
|
This pull request has been mentioned on openHAB Community. There might be relevant details there: https://community.openhab.org/t/jupnp-library-version-2-7-1-oh1-in-my-bundle-list/156221/9 |
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.
This is a workaround for JDK-8301341 by using the Java 11
LinkedTransferQueuewith theQueueingThreadPoolExecutorin the Core and jUPnP.Using the Java 11
LinkedTransferQueueis probably the most stable option because it did not cause these issues with OH3.To be able to use the same artifact with jUPnP, jUPnP also needs to be build with Java 11 or newer (it's currently still compatible with Java 8).
I put the Java 11
LinkedTransferQueuein a separate repo because it has a GPLv2+CE license.Perhaps the repo can also be moved to the openHAB organization?
Because Java and OpenJDK are trademarks, it is probably best to avoid these in repo and package names etc.
Let me know if you agree with the proposed artifact GAV parameters.
After that I can build the artifacts and they can be added to Artifactory.
Closes #3755