Details
-
Bug
-
Resolution: Fixed
-
Critical
-
Cheshire-Cat
-
1
-
Yes
Description
After a quorum failover, ns_server should increment the "next collection ID" and the manifest UID by 0x1000 to minimize the probability that collections created after quorum failover will conflict with those created before.
Attachments
Issue Links
- blocks
-
MB-45435 XDCR - handle quorum failover
- Closed
For Gerrit Dashboard: MB-45711 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
153083,2 | MB-45711 decouple manifest uid from next_uid, because in case of quorum | master | ns_server | Status: MERGED | +2 | +1 |
153084,7 | MB-45711 Increment next colleciton, scope and manifest IDs by 0x1000 during ... quorum failover | master | ns_server | Status: MERGED | +2 | +1 |
153085,9 | MB-45711 each node records last seen manifest uid, next_scope_uid, next_coll_uid for each bucket | master | ns_server | Status: MERGED | +2 | +1 |
153086,9 | MB-45711 limit the advancement of manifest, scope, collection uid to | master | ns_server | Status: MERGED | +2 | +1 |
153520,6 | MB-45711 disallow collections creation furing unfinished failover | master | ns_server | Status: MERGED | +2 | +1 |