-
|
Hi guys, seems like something broke my setup when creating a fresh cluster on v1.12.7 on Hetzner hybrid (vSwitch). The following updated config doesn't work either on 1.12.x nor 1.13.x. Has anyone had this working? I tried everything but can't make it work. The worker nodes/masters on Cloud cannot reach out to the dedicated machines on vSwitch. The vSwitch itself is fine because I tested it with Ubuntu. Does anyone have any idea what's up? So, the dedicated machine joins K8s as a worker node, and it's ready, but it times out when trying to exec or see logs of a pod running on it. 100% connection issue. No firewalls on anything. Full machineconfig. The old way of managing network has the same issue on 1.12.x
The weirdest thing is that when applying the machineconfig to the dedicated node, Another detail is that when I run this command 3 times in a row, for 2 times I get the timeout instantly and the third time it takes a while to timeout.
|
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 5 replies
-
|
Do you still have connectivity to the machine over a public IP? Is the problem only with the VLAN? |
Beta Was this translation helpful? Give feedback.
I feel stupid but cilium was using the wrong interface (vswitch ID changed). Nothing related to Talos