honeymoose
/
OpenSearch
mirror of https://github.com/honeymoose/OpenSearch.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

[ML] Handle failed datafeed in MlDistributedFailureIT (#52631) (#52789)

This commit is contained in:
David Kyle 2020-02-26 08:18:37 +00:00 committed by GitHub
parent 05f1cd74a6
commit 37be695d5c
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 10 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
x-pack/plugin/ml/src/test/java/org/elasticsearch/xpack/ml/integration/MlDistributedFailureIT.java
View File

@ -245,20 +245,27 @@ public class MlDistributedFailureIT extends BaseMlIntegTestCase {
// stopping.
PersistentTasksCustomMetaData tasks = clusterService().state().getMetaData().custom(PersistentTasksCustomMetaData.TYPE);
PersistentTasksCustomMetaData.PersistentTask<?> task = MlTasks.getDatafeedTask(datafeedId, tasks);
// It is possible that the datafeed has already detected the job failure and
// terminated itself. In this happens there is no persistent task to stop
assumeFalse("The datafeed task is null most likely because the datafeed detected the job had failed. " +
"This is expected to happen extremely rarely but the test cannot continue in these circumstances.", task == null);
UpdatePersistentTaskStatusAction.Request updatePersistentTaskStatusRequest =
new UpdatePersistentTaskStatusAction.Request(task.getId(), task.getAllocationId(), DatafeedState.STOPPING);
new UpdatePersistentTaskStatusAction.Request(task.getId(), task.getAllocationId(), DatafeedState.STOPPING);
PersistentTaskResponse updatePersistentTaskStatusResponse =
client().execute(UpdatePersistentTaskStatusAction.INSTANCE, updatePersistentTaskStatusRequest).actionGet();
client().execute(UpdatePersistentTaskStatusAction.INSTANCE, updatePersistentTaskStatusRequest).actionGet();
assertNotNull(updatePersistentTaskStatusResponse.getTask());
// Confirm the datafeed state is now stopping - this may take a while to update in cluster state
assertBusy(() -> {
GetDatafeedsStatsAction.Request datafeedStatsRequest = new GetDatafeedsStatsAction.Request(datafeedId);
GetDatafeedsStatsAction.Response datafeedStatsResponse =
client().execute(GetDatafeedsStatsAction.INSTANCE, datafeedStatsRequest).actionGet();
client().execute(GetDatafeedsStatsAction.INSTANCE, datafeedStatsRequest).actionGet();
assertEquals(DatafeedState.STOPPING, datafeedStatsResponse.getResponse().results().get(0).getDatafeedState());
});
// Stop the node running the failed job/stopping datafeed
ensureGreen(); // replicas must be assigned, otherwise we could lose a whole index
internalCluster().stopRandomNode(settings -> jobNode.getName().equals(settings.get("node.name")));