Details
-
Bug
-
Resolution: Fixed
-
Major
-
4.0.0
-
Security Level: Public
-
mac
-
Untriaged
-
No
Description
The test that I did was for indexer settings was:
0. Change "Stable Snapshot Interval:" value to 2 mins (120000 ms)
1. Load bucket with 10K docs
2. Create index i1 on one of the fields and wait until it is active.
3. curl -u Administrator:asdasd localhost:9102/stats shows that
"default:i1:items_count":10000,"default:i1:num_docs_indexed":10000
4. Create 1M mutations
5. While mutations being indexed, I see stats -
"default:i1:items_count":10000,"default:i1:num_docs_indexed":241019,
6. After waiting for more than two mins I see that though all 1M new docs are indexed, the committed items still show as 10000 -
"default:i1:items_count":10000,"default:i1:num_docs_indexed":1010000
This will be the state until new mutations are created. Though scan returns all new mutations, those docs are indexed in memory. If the indexer crashes at this point, the indexing of 1M docs will happen all over again.
Attachments
Issue Links
- mentioned in
-
Page Loading...
For Gerrit Dashboard: MB-15052 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
61392,3 | MB-15052 Force Commit Every Stable Snapshot Interval | unstable | indexing | Status: MERGED | +2 | +1 |