Unable to access SSH or ping instance after updating GCP OS

By | July 11, 2018

I have a cluster of instances in Google Cloud Platform that was accessible. I ran updates on a VM, then rebooted it. I have not been able to access the VM since. SSH fails, saying it cant be reached. I also cannot ping the instance. When I connect via the console port, it is flooded with martian source logs.

martian source x.x.x.x <affected machine> from x.x.x.x <some other ip>, on dev eth0
ll header: 42:01:0a:8e:00:05:42:01:0a:8e:00:01:08:00
Jul 11 20:07:40 <servername> kernel: __ratelimit: 3 callbacks suppressed

It is only affecting the machines that have been updated and rebooted. The ones that have been updated but NOT rebooted are fine (so far). I also checked the firewall-rules, and there is the default SSH rule.

I did find that if I clone an affected VM and give it a new OS disk, I can access it afterwards.

Any ideas? I am at a loss as to what to do with the ones that have been (and have not been) updated.