Description
See also MB-18912 for details about forcing vbucket allocation on indexer queue.
It is expected that system should eventually get out of this situation, however in watson 4.6.0-3556, the queue went up to 235M Items. I stopped the test, and for 48+ hours queue is still constant, (see screen)
2016-12-11T16:41:21.737-08:00 [Warn] Indexer::MutationQueue Max Wait Period for Node Alloc Expired 300030. Forcing Alloc. Bucket default Vbucket 773
|
2016-12-11T16:46:22.127-08:00 [Warn] Indexer::MutationQueue Max Wait Period for Node Alloc Expired 300030. Forcing Alloc. Bucket default Vbucket 773
|
2016-12-11T16:51:22.817-08:00 [Warn] Indexer::MutationQueue Max Wait Period for Node Alloc Expired 300030. Forcing Alloc. Bucket default Vbucket 795
|
2016-12-11T16:56:23.328-08:00 [Warn] Indexer::MutationQueue Max Wait Period for Node Alloc Expired 300030. Forcing Alloc. Bucket default Vbucket 795
|
The result of this is that queries are not returning items, and delete queries do not delete items, which in longevity test leads to running out of disk space.
Cluster is currently active in this state: 172.23.105.60:8091
Attachments
Issue Links
- blocks
-
MB-20290 4.6.0 Point Release
- Closed