Details
-
Bug
-
Resolution: Fixed
-
Critical
-
7.1.3
-
Untriaged
-
0
-
Unknown
Description
Huge index was created and it was creating memory pressure on the node as plasma was unable to clear up memory due to minimum memory requirements and it was slowing down the flush. User deletes the index to reduce the memory but the index deletion is stuck on flush which is slow.
2023-05-03T09:07:27.840+00:00 [Info] clustMgrAgent::OnIndexDelete Notification Received for Drop IndexId 18246201642360272851 &{0}
|
2023-05-03T09:07:27.840+00:00 [Info] Indexer::handleDropIndex - IndexInstId 18246201642360272851
|
|
2023-05-03T09:07:27.919+00:00 [Info] ClustMgr::handleIndexMap, updated instances:
|
InstanceId: 18246201642360272851 Name: report-user-actions-max-last-modified-with-actions Keyspace: user-data-bucket:_default:_default State: INDEX_STATE_DELETED Stream: INIT_STREAM RState: RebalActive Version: 0 ReplicaId: 1
|
|
2023-05-03T09:07:34.836+00:00 [Warn] Timekeeper::flushMonitor Waiting for flush to finish for 1200 seconds. Stream INIT_STREAM KeyspaceId user-data-bucket.
|
Attachments
Issue Links
- is cloned by
-
MB-58456 [BP 7.2.3] Decouple DeleteIndex with flush so that memory is reclaimed
- Closed