I have configured kubernetes cluster where it has 2 masters, 3 etcds and 20 nodes. Masters fronted by a load balancer. I followed up https://coreos.com/kubernetes/docs/latest/getting-started.html when creating the cluster. Everything work as expected. I could deploy pods without issue and pods running fine. But when I tried to tail the logs, kubectl suddenly return unexpected EOF and stop tailing. Again I have to execute kubectl logs command to continue. This is very annoying as it not even keep one minute.
The command which I execute is, kubectl logs -f --tail=100 <pod_name> or kubectl logs -f <pod_name>
After less than one minute it return, error: unexpected EOF
Appreciate your input to sort out the issue.
AWS ELB has a default "IdleTimeout" of 60 seconds, that can be increased up to 3600 seconds. I can confirm that increasing this value solve this issue.
Here you can find more information about it: http://docs.aws.amazon.com/elasticloadbalancing/latest/classic/config-idle-timeout.html