Increase the minimum size of the management pool to 2 #90193
Merged
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
Make sure that individual `wait_for_completion=true` requests don't block concurrent stats and cat requests.
arteam
added
:Distributed/Distributed
|
Hi @arteam, I've created a changelog YAML for you. |
|
Pinging @elastic/es-distributed (Team:Distributed) |
DaveCTurner
previously requested changes
Sep 21, 2022
There was a problem hiding this comment.
I don't think this fixes the bug that #89564 exposes unfortunately - if there were two concurrent calls to the list-tasks API then they could both deadlock within TaskManager#waitForTaskCompletion even after this change.
Instead, I think we should make TaskManager#waitForTaskCompletion async rather than blocking the management thread as it does today..
henningandersen
approved these changes
Sep 21, 2022
I agree that is the ideal fix. However, if that cannot be done in time for 8.5.0 then increasing the minimum number of threads to 2 at least keeps ESS in the same situation as it was in in 8.4 and earlier. In 8.4 and earlier ESS told every node it had at least 2 processors to avoid strange lockups, and those strange lockups (undiagnosed at the time they were originally mentioned) were almost certainly the same problem as #89564. In 8.5 ESS is going to tell nodes with one processor that they have one processor, so the strange lockups will be seen by ESS users on small nodes unless something is done in Elasticsearch. So by all means fix this properly for 8.5.0 if there's time, but something has to be done for 8.5.0 even if that's a quick workaround. |
|
@elasticmachine run elasticsearch-ci/packaging-tests-windows-sample |
💔 Backport failedThe backport operation could not be completed due to the following error: You can use sqren/backport to manually backport by running |
|
The commit went to the master branch before the 8.5 branch was cut, so this change is included in 8.5 and doesn't need a backport. |
weizijun
added a commit
to weizijun/elasticsearch
that referenced
this pull request
Sep 22, 2022
* main: (186 commits) [DOCS] Add 8.5 release notes and fix links (elastic#90201) Mute DownsampleActionSingleNodeTests.testCannotRollupWhileOtherRollupInProgress (elastic#90198) Bump to version 8.6.0 Increase the minimum size of the management pool to 2 (elastic#90193) Speed `getIntLE` from `BytesReference` (elastic#90147) Restrict nodes for testClusterPrimariesActive1 (elastic#90191) Fix bug with `BigIntArray` serialization (elastic#90142) Synthetic _source: test _source filtering (elastic#90138) Modernize cardinality agg tests (elastic#90114) Mute failing test (elastic#90186) Move assertion in ES85BloomFilterPostingsFormat to fix test (elastic#90150) Restrict nodes for testClusterPrimariesActive2 (elastic#90184) Batch index delete cluster state updates (elastic#90033) Register stable plugins in ActionModule (elastic#90067) Mute failing test (elastic#90180) [HealthAPI] Disk: Use _ for diagnosis id (elastic#90179) [HealtAPI] Disk: use shorter help URLs (elastic#90178) Fixing disk health indicator unit tests (elastic#90175) Enable the health node and the disk health indicator elastic#84811 (elastic#90085) Add missing Disk Indicator health api IDs (elastic#90174) ...
weizijun
added a commit
to weizijun/elasticsearch
that referenced
this pull request
Sep 22, 2022
* main: (121 commits) [DOCS] Add 8.5 release notes and fix links (elastic#90201) Mute DownsampleActionSingleNodeTests.testCannotRollupWhileOtherRollupInProgress (elastic#90198) Bump to version 8.6.0 Increase the minimum size of the management pool to 2 (elastic#90193) Speed `getIntLE` from `BytesReference` (elastic#90147) Restrict nodes for testClusterPrimariesActive1 (elastic#90191) Fix bug with `BigIntArray` serialization (elastic#90142) Synthetic _source: test _source filtering (elastic#90138) Modernize cardinality agg tests (elastic#90114) Mute failing test (elastic#90186) Move assertion in ES85BloomFilterPostingsFormat to fix test (elastic#90150) Restrict nodes for testClusterPrimariesActive2 (elastic#90184) Batch index delete cluster state updates (elastic#90033) Register stable plugins in ActionModule (elastic#90067) Mute failing test (elastic#90180) [HealthAPI] Disk: Use _ for diagnosis id (elastic#90179) [HealtAPI] Disk: use shorter help URLs (elastic#90178) Fixing disk health indicator unit tests (elastic#90175) Enable the health node and the disk health indicator elastic#84811 (elastic#90085) Add missing Disk Indicator health api IDs (elastic#90174) ...
droberts195
added a commit
to droberts195/elasticsearch
that referenced
this pull request
Sep 22, 2022
Following elastic#90193 this should no longer fail.
droberts195
added a commit
that referenced
this pull request
Sep 23, 2022
Following #90193 this should no longer fail.
original-brownbear
added a commit
to original-brownbear/elasticsearch
that referenced
this pull request
Jan 4, 2023
Instead of blocking either a transport or management thread here (the latter will lead to a dead-lock if there's only a single management thread available, we should just queue up those sends and run them async via listener). This test is currently only working on `main` because of elastic#90193, this makes it work with a single management thread as well so it works on 7.x again. closes elastic#92629
original-brownbear
added a commit
that referenced
this pull request
Jan 4, 2023
Instead of blocking either a transport or management thread here (the latter will lead to a dead-lock if there's only a single management thread available, we should just queue up those sends and run them async via listener). This test is currently only working on `main` because of #90193, this makes it work with a single management thread as well so it works on 7.x again. closes #92629
original-brownbear
added a commit
to original-brownbear/elasticsearch
that referenced
this pull request
Jan 4, 2023
Instead of blocking either a transport or management thread here (the latter will lead to a dead-lock if there's only a single management thread available, we should just queue up those sends and run them async via listener). This test is currently only working on `main` because of elastic#90193, this makes it work with a single management thread as well so it works on 7.x again. closes elastic#92629
original-brownbear
added a commit
to original-brownbear/elasticsearch
that referenced
this pull request
Jan 4, 2023
Instead of blocking either a transport or management thread here (the latter will lead to a dead-lock if there's only a single management thread available, we should just queue up those sends and run them async via listener). This test is currently only working on `main` because of elastic#90193, this makes it work with a single management thread as well so it works on 7.x again. closes elastic#92629
elasticsearchmachine
pushed a commit
that referenced
this pull request
Jan 4, 2023
Instead of blocking either a transport or management thread here (the latter will lead to a dead-lock if there's only a single management thread available, we should just queue up those sends and run them async via listener). This test is currently only working on `main` because of #90193, this makes it work with a single management thread as well so it works on 7.x again. closes #92629
elasticsearchmachine
pushed a commit
that referenced
this pull request
Jan 4, 2023
Instead of blocking either a transport or management thread here (the latter will lead to a dead-lock if there's only a single management thread available, we should just queue up those sends and run them async via listener). This test is currently only working on `main` because of #90193, this makes it work with a single management thread as well so it works on 7.x again. closes #92629
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.
Make sure that individual
wait_for_completion=truerequests don't block concurrent stats and cat requests.See #89564