Pods are not running and failing very often

praveen

Hello Team,

I have deployed the pubsub+ HA using helm chart in the GDCE environment (Google Distributed Cloud Edge) which is GKE cluster.

There are 3 nodes, hence 3 pods are suppose to run. but 2 pods are failing very often. and one pod is about to start, but not starting.

solace.PNG

One of the Pod log says: "ERROR: startup-broker.sh-Config-sync never reached state "Up"

Chart downloaded: https://solace.com/blog/deploy-solace-in-a-kubernetes-cluster-with-helm/

marc

Hi @praveen,

Are there health checks being performed against the containers? If so what is configured? And are the ports open for them to be performed? To me this looks like GKE is thinking the containers are unhealthy and restarting them every so often.

A lot of times the default setting is looking for a web server on a certain port…but I'm not sure what GKE does.

Note that if you have access to paid support I would recommend opening a ticket on this as they can dig into the details quickly with you :)

Hope that helps,

-Marc

praveen

Hi Marc,

Cluster is configured with CIS complaints. Is that a reason for the Pods restart?

If yes, can share the link to get the solace manifests that works for CIS.

Thanks,

Praveen