Open Cryptophobia opened 6 years ago
From @timfpark on April 5, 2017 15:41
Deleting the router pod manually with kubectl seems to resolve this.
From @kris-nova on April 5, 2017 22:25
Are you running with EXPERIMENTAL_NATIVE_INGRESS
?
From @bacongobbler on April 5, 2017 22:33
@kris-nova I don't think experimental native ingress was released in v2.12.0. We have still yet to release v2.13.0. :)
ping @krancour, any clues? This is one bug I have not seen before.
@timfpark would you mind dumping the entirety of kubectl logs --previous
for the router in a gist? I'd also check and see what's happening on the kubelet's side as well just in case; that no such container
error usually indicates something wrong on the Docker level.
From @krancour on April 6, 2017 15:2
I have never seen anything like that before and cannot think of any reason the router itself would throw such an error. As @bacongobbler suggested, I'd love to see a more complete log if that's available.
From @timfpark on April 5, 2017 15:3
In my deployment of Deis Workflow v2.12.0, I am seeing a CrashLoopBackOff on the Router component:
deis deis-router-2126433040-drr9t 0/1 CrashLoopBackOff 85 5d
looking at the logs with kubectl, I see the following:
no such container: "169d0df2740125110584a290b4b7da6b1fa8c3d9ebd82120ec93ee621fd2070e"
Is this a known issue? Is there anything I can collect to help diagnose what's going on here?
Copied from original issue: deis/workflow#784