[bitnami/redis]: hotfix, regenerate sentinel config at each boot-up

[rsecob]

Description of the change

After upgrading to 15.0.0 we realized there was an issue with sentinel. If the service is restarted within its running container, the sentinel config remained unchanged. The script is supposed to prune the known sentinels and replicas and regenerate them when fetching the available hosts on the network. This part of the script was not working as expected which caused this error:

>>> 'sentinel known-sentinel rt-collector-master 10.8.5.30 26379 ecac44470a643ce5d853d166ad8abc2135e4bdbf'
Duplicate runid for sentinel.

Our proposed solution is to override the existing configuration (if it exists) with the original configuration at each bootup. This ensures all replicas start with the expected configuration file.

Benefits

• Fixes CrashBackLoop state if the service is restarted in a running container.

Checklist

• Chart version bumped in Chart.yaml according to semver.
• Title of the PR starts with chart name (e.g. [bitnami/<name_of_the_chart>])

[yilmi]

Hi @rsecob,

Thanks for this contribution!!! I've been trying to reproduce your issue using our latest chart 15.0.3 without success.

Would you mind confirming that you also see this issue with the 15.0.3 release? And confirm the version of containers you are using?

Thanks,
Yass

[rsecob]

Thank you @yilmi for your review.

Yes I confirm we have the same issue. In order to see the issue happening, you need to log in into the container and kill the process.

Here is an example:

values.yaml

auth:
  enabled: false

sentinel:
  enabled: true

Deploy and let it dispatch the three workers

helm install -f values.yaml test bitnami/redis

Then log in into one sentinel container:

ktl exec -ti test-redis-node-2 -c sentinel -- bash

And kill the sentinel process with kill 1

[yilmi]

Thanks for the details, let me check that and get back to you quickly!

[yilmi]

lgtm, one note, I had to recreate the pods. Thanks again @rsecob for this contribution!