Scheduler in Environments crashed after RancherServer Crash

1

Hi,
I had a strange thing with my RancherServer 1.4.2 environments. I know that I am not running the latest version, but I would like to understand what happend.

I have a RancherServer 1.4.2 serving multiple environments. One day the master crashed due to a problem with the VM it was running on.
No Problem, containers on the nodes kept running and all was fine.

I fixed the master and since then all Schedulers in the Network Stacks of the Environments wouldnt start. Which resulted in containers not starting after a change.
Only a restart of the Docker Daemons on the nodes solved the problem.

Any idea why the schedulers would crash, if the master vanishes?
I never saw this behavior before when I stopped the master on purpose.

Thanks, Chris

2

Did the IP address of the master change? That could have done it potentially. Did the logs of scheudler container indicate why it was crashing?

3

No, IP is the same.

But I see errors in RancherServer, which I didnt recognize before:

2017-11-16 16:11:59,667 ERROR [:] [] [] [] [torService-1457] [.e.s.i.ProcessInstanceDispatcherImpl] Unknown exception running process [instance.create:4116214] on [156662] io.cattle.platform.util.exception.ExecutionException: Failed to find network for networkMode appl_default
        at io.cattle.platform.docker.process.instance.DockerInstancePreCreate.handle(DockerInstancePreCreate.java:64) ~[cattle-docker-compute-0.5.0-SNAPSHOT.jar:na]

And I have no idea, what networkMode appl_default means.