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

[ML] Fix rare ML daily maintenance test race condition (#64043) 

Depending on thread scheduling the ML daily maintenance
tests could do one more iteration than expected, causing
rare failures.

Fixes #64036
This commit is contained in:
David Roberts 2020-10-22 12:39:20 +01:00
parent fa8ad3abde
commit cb0c538b35
1 changed files with 10 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
x-pack/plugin/ml/src/test/java/org/elasticsearch/xpack/ml/MlDailyMaintenanceServiceTests.java
View File

@ -230,8 +230,16 @@ public class MlDailyMaintenanceServiceTests extends ESTestCase {
private MlDailyMaintenanceService createService(CountDownLatch latch, Client client) {
return new MlDailyMaintenanceService(Settings.EMPTY, threadPool, client, clusterService, mlAssignmentNotifier, () -> {
// We need to be careful that an unexpected iteration doesn't get squeezed in by the maintenance threadpool in
// between the latch getting counted down to zero and the main test thread stopping the maintenance service.
// This could happen if the main test thread happens to be waiting for a CPU for the whole 100ms after the
// latch counts down to zero.
if (latch.getCount() > 0) {
latch.countDown();
return TimeValue.timeValueMillis(100);
} else {
return TimeValue.timeValueHours(1);
}
});
}