After performing several reboots in a row, SNO cluster API does not respond anymore:

The connection to the server api.cloudransno-site1.slcm1.bos2.lab:6443 was refused - did you specify the right host or port?

4.12.16

We run a test that performs several reboot in a row. We see this issue with a high rate every time we run that test. We say in 4.12.16 100%of times, and now also in 4.12.21 happened the first time we run the test.

1. Reboot SNO cluster 5 times
2. Check API

Node does not respond anymore. I left it several hours but it did not come back.

Node recovers properly

System Impact: Very severe. Node cannot be longer used

ACM reports: 

The kube-apiserver is not ok, status code: 0, Get "https://172.31.0.1:443/livez": dial tcp 172.31.0.1:443: connect: connection refused

oc adm must gather cannot be performed. Only SOS report. Logs attached