Ccloud kafka topic consume STOPS working (Windows 10)

Wondering if anyone has seen this behavior and how to resolve it…

Using Confluent Cloud , with local (windows 10) CLI install.

Been through the demos, all works well, for a while. After producing and consuming a good number of messages, I left the command windows open. Coming back later that afternoon, I found that attempts to connect to either produce or consume resulted in:

an apparent timeout after 2 minutes when connecting to consume (the time was consistent), the message was “Error - can’t connect to cluster”

no progress when trying to produce (no error message)

All other commands appeared to still work. I could list the cluster, the api keys, etc… just couldn’t produce/consume… as if the api-key to the SR was no longer valid, but there were no obvious errors, other than the eventual “Can’t connect”).

The first time it happened, opening a new command window resolved the issue. Since then, even that does not work.

C:\Users>ccloud version
ccloud - Confluent Cloud CLI

Version: v1.39.1
Git Ref: b9ca3f7e
Build Date: 2021-09-07T23:37:20Z
Go Version: go1.16.3 (windows/amd64)
Development: false

I believe rebooting will resolve the problem, but for how long? I can’t even say what period of inactivity is required to cause the problem to happen.

Any advice/feedback welcomed.

OK - rebooting did not resolve the issue - apparently it’s now wedged w.r.t. CLI.

Only recent change was upgrade from 1.37.1 to 1.39.1

Note that this problem happened with 1.37.1 (before the upgrade). Since the upgrade, it appears wedged permanently.

Nothing I have done gets me to a point where I can re-enter the SR API-KEY and secret.

For the record, I left everything (CCloud-related) logged out over the weekend, and was able to run both a consumer and producer successfully this morning. I did not have to re-enter any keys … it just worked. No indication as to what may have been wrong, but I do not consider the situation resolved.

Could you please provide me with the pkc? I will be happy to look at your cluster to see if there were any issues during the time.

Sorry, Brendan - did not see your post till just now. Was on a chat with Dan C @ confluent. He could see the issue, and could not access from an EC2 instance in the same region (eu-west-1). We tore down and recreated API key and it cleared for him, but not for me, so we are at the point that this appears to be on our side, either networking or (more likely) the VPN client we use.

1 Like