Details
-
Improvement
-
Resolution: Fixed
-
Major
-
4.0.0
-
Security Level: Public
-
400-1977
10 bucket, 20 gsi indexes
6 Node
8 Core X SSD nodes
Description
1. Create cluster, indexes
2. Load data on the buckets, stop the load
3. The indexes are all "ready", indexer and cbq-engine have no load on them
4. Reboot the cluster,
I noticed heavy CPU usage by indexer, projector ~ 90% on indexer nodes.
- select * from system:indexes – shows every index is online
- curl getIndexStatus – shows every Index is ready.
- Indexer activity ( screenshot attached) is heavy on the indexer nodes
The indexes are likely in the recovery mode - Currently it is not exposed to the user as "Recovery" mode, the UI, getIndexStats nothing shows the Recovery mode.
Might be a good idea to expose this as "Recovery mode" for Index/ Status.