Details
-
Bug
-
Resolution: Fixed
-
Major
-
1.8.0, 1.8.1
-
Security Level: Public
Description
From some customer clusters, we observed the sqlite-level errors that prevent the flusher from persisting dirty items into disk. We even saw that this symptom happens across all the buckets on a single node at the same time. To resolve this issue, we should re-establish the connections to the sqlite upon the sqlite-level failures.
Attachments
For Gerrit Dashboard: MB-5401 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
16508,2 | MB-5401 Reopen database upon failures from the database layer | branch-181 | ep-engine | Status: MERGED | +2 | +1 |
16674,1 | MB-5401 Reopen database upon failures from the database layer | branch-18 | ep-engine | Status: MERGED | +2 | +1 |
16678,1 | Merge branch 'branch-18' | master | ep-engine | Status: ABANDONED | 0 | 0 |
16679,1 | Merge branch 'branch-18' | master | ep-engine | Status: ABANDONED | 0 | 0 |
16680,1 | Merge branch 'branch-18' | master | ep-engine | Status: ABANDONED | 0 | 0 |
16681,1 | Merge branch 'branch-18' | master | ep-engine | Status: MERGED | +2 | +1 |