This message was deleted.
# k3sa
adamant-kite-43734
11/27/2023, 12:33 PMThis message was deleted.
😲 1
b
brave-afternoon-4801
11/27/2023, 12:43 PMThis is honestly very concerning that it was this easy to destabilize the cluster
brave-afternoon-4801
11/27/2023, 12:44 PMI unfortunately don't have time to look into this, but can firmly say, I am now deleting an entire cluster because of this.
c
creamy-pencil-82913
11/27/2023, 5:42 PMTry to bind what?
creamy-pencil-82913
11/27/2023, 5:42 PMThat's the supervisor and apiserver port. It's pretty important. Don't put other things on that port!
b
brave-afternoon-4801
11/27/2023, 5:43 PMI am gonna have to try to recreate it, but it happened when fiddling with the rancher-monitoring chart and one of the subcharts set with hostNetwork=true
brave-afternoon-4801
11/27/2023, 5:43 PMlol i agree, tis important 😛
but i'm very surprised that it kicked off a re-keying
brave-afternoon-4801
11/27/2023, 5:43 PMor at least that is my current assumption of what happened
brave-afternoon-4801
11/27/2023, 5:45 PMIf I was deving with a distro that didn't require a forked version of vagrant-libvirt, I would publish all these cases, but my current setup is difficult to reproduce
c
creamy-pencil-82913
11/27/2023, 5:48 PMIt didn't rekey anything. It sounds like you just ended up with something else occupying the port in place of k3s
b
brave-afternoon-4801
11/27/2023, 5:48 PMwell, but I removed that thing within 3 minutes and rebooted the nodes
brave-afternoon-4801
11/27/2023, 5:49 PMwithin about an hour they had all lost communication with each other
brave-afternoon-4801
11/27/2023, 5:49 PMwell, i shouldn't speculate with a fuzzy memory. I'll try to reproduce soon!
🎯 1