-
Type:
Bug
-
Resolution: Unresolved
-
Priority:
Medium
-
None
-
Affects Version/s: None
-
Component/s: Repository - Branches / Tags
-
None
-
13
-
Severity 3 - Minor
-
10
When the TagIndexer gets backed up indexing repositories, it can trigger significant amounts of logging:
2017-08-07 06:26:14,416 WARN [hz.hazelcast.event-2] c.a.b.internal.tag.idx.TagIndexer [repository:105570] Could not schedule background tag indexing; the task queue is full 2017-08-07 06:26:14,578 WARN [hz.hazelcast.event-2] c.a.b.internal.tag.idx.TagIndexer [repository:112924] Could not schedule background tag indexing; the task queue is full 2017-08-07 06:26:14,609 WARN [hz.hazelcast.event-2] c.a.b.internal.tag.idx.TagIndexer [repository:36201] Could not schedule background tag indexing; the task queue is full 2017-08-07 06:26:14,627 WARN [hz.hazelcast.event-2] c.a.b.internal.tag.idx.TagIndexer [repository:77257] Could not schedule background tag indexing; the task queue is full 2017-08-07 06:26:14,876 WARN [hz.hazelcast.event-2] c.a.b.internal.tag.idx.TagIndexer [repository:114088] Could not schedule background tag indexing; the task queue is full 2017-08-07 06:26:14,883 WARN [hz.hazelcast.event-2] c.a.b.internal.tag.idx.TagIndexer [repository:106903] Could not schedule background tag indexing; the task queue is full ...
At a minimum, scheduling should apply a cooldown between logging rejected tasks, rather than logging every task. If possible, however, it would be better if tag indexing could use a more efficient mechanism for queuing and running indexing.