Redis Replication/Sentinel cluster #1035
Replies: 1 comment 4 replies
-
|
Do you mean that the sentinel will monitor an outdated master IP of the Redis pod? |
Beta Was this translation helpful? Give feedback.
-
|
I guess. If I restart the RedisReplication pods several times, then check the RedisSentinel logs it is trying to monitor/connect to the IP that had been assigned the master RedisReplication pod. Currently the only workaround I know is to manually run the command to tell Sentinel what the new IP is. |
Beta Was this translation helpful? Give feedback.
-
|
That is indeed the issue. It's hard to identify the sentinel cluster as the newest pod of the replication cluster because they are in separate custom resources. I opened an issue at #1003 to address this. |
Beta Was this translation helpful? Give feedback.
-
|
In your ticket are you suggesting to add Sentinel as a sidecar to the RedisReplication pods? |
Beta Was this translation helpful? Give feedback.
-
|
No, they run in different workloads but share the same custom resource definition. |
Beta Was this translation helpful? Give feedback.
-
Has anyone been able to successfully run a Redis cluster using RedisReplication and RedisSentinel? I can get the initial deployment working just fine, but seems something goes wrong as the pods are restarted. At some point the Sentinels cache the IP of the master node even after they have restarted and gotten new pods, which leaves Sentinel failing trying to use an IP currently not assigned to any pod. In other cases I have seen it also get into a state where it thinks the port is 0.0.0.0
I am using redis-operator: v0.18.0 and redis: v7.2.3
Beta Was this translation helpful? Give feedback.
All reactions