-
Problem report
-
Resolution: Fixed
-
Critical
-
6.0.13, 6.2.7, 6.4.0beta6
-
None
-
Sprint 98 (Mar 2023), Sprint 99 (Apr 2023)
-
1
Steps to reproduce:
- use HA with several nodes
- try to upgrade with all nodes automatically deployed (human error, by Ansible, some Kubernetes issues or not enough constraints in PCS (Corosync/Pacemaker), etc...)
Result:
- get duplicates or broken schema
Expected:
- Nodes should agree on upgrade leader if more than 1 node is started and upgrade is required.
- If upgrade already failed on some step - second node should not try to do something.
- part of
-
ZBXNEXT-7657 Add HA quorum and killswitch for database consistency protection
-
- Open
-
-
ZBX-22309 HA maintenance like upgrade procedure is not described in the documentation
-
- Closed
-
