when using preemtible vms https://cloud.google.com/compute/docs/instances/preemptible
these instances can be without noticed be restarted on any given time.
this will cause the worker to disappear for a a short time which then will result in the following errors
volume not foundworker concourse-worker-* disappeared while trying to reach it
for the stemcell and bats tests tis means that there are leftover instances that are not beining cleaned up.
so this needs to be handled.
when using preemtible vms https://cloud.google.com/compute/docs/instances/preemptible these instances can be without noticed be restarted on any given time. this will cause the worker to disappear for a a short time which then will result in the following errors
volume not found
worker concourse-worker-* disappeared while trying to reach it
for the stemcell and bats tests tis means that there are leftover instances that are not beining cleaned up. so this needs to be handled.
we can maby prevent many outages to just restart the vm on our terms within the 24h period see https://github.com/estafette/estafette-gke-preemptible-killer and https://gist.github.com/ahume/d56699f3eb2292dbbc1ba3825d44e4b5