Fix: ILMDownsampleDisruptionIT.testILMDownsampleRollingRestart #137423
Conversation
elasticsearchmachine
added
needs:triage
gmarouli
added
the
:StorageEngine/Downsampling
|
Pinging @elastic/es-storage-engine (Team:StorageEngine) |
elasticsearchmachine
removed
the
needs:triage
| // We assert that ILM to successfully completed the phase | ||
| assertBusy(() -> { |
There was a problem hiding this comment.
Could we do this with ClusterServiceUtils#addTemporaryStateListener rather than polling any API? We can check the API response afterwards ofc but every assertBusy adds a little extra idle time to tests and in total that's a pretty big contributor to overall test runtime.
...c/internalClusterTest/java/org/elasticsearch/xpack/downsample/ILMDownsampleDisruptionIT.java
Outdated
Show resolved
Hide resolved
| }, 60, TimeUnit.SECONDS); | ||
| // We assert that ILM successfully completed the phase | ||
| logger.info("Waiting for ILM to complete the phase for index [{}]", targetIndex); | ||
| ClusterServiceUtils.addTemporaryStateListener(clusterState -> { |
There was a problem hiding this comment.
This just adds (and returns) the listener - you have to safeAwait on it too :)
There was a problem hiding this comment.
You could also use awaitClusterState instead :)
...c/internalClusterTest/java/org/elasticsearch/xpack/downsample/ILMDownsampleDisruptionIT.java
Outdated
Show resolved
Hide resolved
| ensureGreen(targetIndex); | ||
| // We assert that ILM successfully completed the phase | ||
| logger.info("Waiting for ILM to complete the phase for index [{}]", targetIndex); | ||
| awaitClusterState(clusterState -> { |
There was a problem hiding this comment.
Ohh yes I'd completely forgotten we had an awaitClusterState to do this 👍
There was a problem hiding this comment.
Me too, credits to @nielsbauman
There was a problem hiding this comment.
I have less of an excuse tho 😇
$ git annotate -- test/framework/src/main/java/org/elasticsearch/test/ClusterServiceUtils.java | grep awaitClusterState
1337476bc9efc (David Turner 2025-08-26 12:54:45 +0100 241) public static void awaitClusterState(Predicate<ClusterState> statePredicate, ClusterService clusterService) {
…ticsearch/xpack/downsample/ILMDownsampleDisruptionIT.java Co-authored-by: David Turner <david.turner@elastic.co>
gmarouli
added
the
auto-merge-without-approval
4 participants
This PR stabilises
testILMDownsampleRollingRestartand closes #136585.This test started failing more often after #135834 because this change increased the chance of manifesting #137422.
To reduce this chance we wait for ILM to finish the phase before we end the test.
Closes #136585