This message was deleted.

iirc the purpose of that is to execute a one-shot command to update the config. It’s not meant to run forever, and its intended purpose is to do what it’s doing - update the config to sync with the configuration as managed by rancher. Why are you trying to override the configuration locally instead of managing it through rancher?

Managing through rancher would be preferred, I just cannot figure out how to update the server address that the agent is talking to

It comes from the rancher config, where you tell it what its address is.

Are you referring to the server-url in the global settings?

yep

So even when I update that, the agents don't start communicating to the new server -- is there another way for me to push the change down to the agent?

there used to be a tool to do that, but I don’t know that its supported any longer

Hmm. Is the server config supposed to be automatically sent down to existing agents when it is updated or no?

right

Looking at the doc, it seems like this command should do the trick?

kubectl annotate <http://clusters.management.cattle.io|clusters.management.cattle.io> <CLUSTER_ID> io.cattle.agent.force.deploy=true

you can do that if they are still connected to the server, yes. If they are disconnected because you changed the URL or cert out from under them and now they’re disconnected, then you need to follow the steps to manually redeploy the agents.

Bingo, that turned out to be the key, then I just had to re-run the agent registration command and it connected right up! Thanks so much for the help

classic x-y problem :)