We are seeing an error in our Couchbase Operator Kubernetes cluster that once it happens we seem to lose all functionality within our Couchbase cluster and the only recoverable path is to completely destroy and rebuild the couchbase cluster.
Client-side error-report for user “Administrator” on node ‘email@example.com’:
User-Agent:Mozilla/5.0 (Macintosh; Intel Mac OS X 10_14_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/74.0.3729.169 Safari/537.36
Failed to get bucket insights.
Try refreshing the bucket insights.
I assume you’re referring to upgrading your GKE nodes? Or is it the couchbase cluster? If GKE, then I’m also assuming you’re running operator 1.2 - otherwise this operation is not supported.
Aside from versioning/environment, it’s not clear to me what’s going on, so I suggest running cbopinfo and sharing more info about what versions you’re upgrading between so that we can take a further look -> https://docs.couchbase.com/operator/1.2/cbopinfo.html#examples
So, there must be more info here and you’re not losing “all functionality”. At least the logging is working! Is there more impact from an app perspective? Does cbc pillowfight or sdk-doctor connect to the cluster or is it just the UI?