Closed ifalex closed 1 week ago
Looking through the structure of the service, observed that the issue happens due to entrypoint who tries to modify /etc/redis/sentinel.conf based on env variables that are obtained through Helm. A proposal to mitigate this issue would be to move the creation of sentinel.conf through k8s config management native mechanism which is configMap. That way the entrypoint script would not be needed anymore allowing the read-only file system capability avaiable.
What version of redis operator are you using?
redis-operator version: redis-operator version: redis-operator:0.16.0 (built image from tag)
Does this issue reproduce with the latest release? Yes
What operating system and processor architecture are you using (
kubectl version
)?kubectl version
OutputClient Version: v1.29.3 Kustomize Version: v5.0.4-0.20230601165947-6ce0bf390ce3 Server Version: v1.28.8+k3s1
What did you do?
What did you expect to see? As the file becomes readable only I expect here to be allowed to set an emptyDir from configuration as it can be performed for the other crds provided. Not allowing this results in bellow error CrashLoopBackOff
What did you see instead?
Running sentinel without TLS mode ACL_MODE is not true, skipping ACL file modification Starting sentinel service ..... 1:X 06 Jun 2024 06:26:21.394 # Sentinel config file /etc/redis/sentinel.conf is not writable: Read-only file system. Exiting... /usr/bin/entrypoint-sentinel.sh: line 38: /etc/redis/sentinel.conf: Read-only file system /usr/bin/entrypoint-sentinel.sh: line 18: /etc/redis/sentinel.conf: Read-only file system /usr/bin/entrypoint-sentinel.sh: line 76: /etc/redis/sentinel.conf: Read-only file system