You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Describe the bug
Redis Sentinel keeps old replicas around after a replica pod is deleted (and respawns). If a new Redis pod uses one of these ip addresses it becomes a slave of the wrong master and loses its own data.
This will contain the old dead slave with "flags": "s_down,slave" as well as the new one and others.
It is difficult to get another Redis pod to come up with the same IP but eventually it happens and that Redis becomes a replica of the wrong master and loses its data.
Expected behavior
The dead replicas should disappear from the replicas list. The workaround is to do this manually:
Which chart:
Bitname/redis 15.3.2
Describe the bug
Redis Sentinel keeps old replicas around after a replica pod is deleted (and respawns). If a new Redis pod uses one of these ip addresses it becomes a slave of the wrong master and loses its own data.
Our Redis's are deployed as follows:
To Reproduce
Deploy Redis as above. Kill one of the slave nodes. Check the list of replicas with one of the Sentinels:
This will contain the old dead slave with "flags": "s_down,slave" as well as the new one and others.
It is difficult to get another Redis pod to come up with the same IP but eventually it happens and that Redis becomes a replica of the wrong master and loses its data.
Expected behavior
The dead replicas should disappear from the replicas list. The workaround is to do this manually:
Version of Helm and Kubernetes:
helm version
:kubectl version
:The text was updated successfully, but these errors were encountered: