Re: JNLP java.nio.channels.ClosedChannelException with kubernetes-plugin

Hi Carlos, I am running into this same issue when I try to run pods in parallel within the same job. As the number of pods launched by the job increases, the number of timeouts/disconnects increases. See attached images for the failure details. I believe it is the same issue. Regards, Art

Re: JNLP java.nio.channels.ClosedChannelException with kubernetes-plugin

what flavor of kubernetes ? cloud? on prem? how was it built? what networking is it using? there are a lot of variations so the best chance to find out is to relate with other people's environment and try to isolate On Thu, Apr 25, 2019 at 6:36 PM Gopi wrote: > Hi > Using latest Kubernetes 1.13

Re: JNLP java.nio.channels.ClosedChannelException with kubernetes-plugin

Hi Using latest Kubernetes 1.13 with 3 master nodes. On Thu, Apr 25, 2019 at 2:15 PM Carlos Sanchez wrote: > what kubernetes cluster are you using and with what networking? > my guess is that the underlying network layer is killing inactive > connections > > On Thu, Apr 25, 2019 at 4:24 AM Gopi

Re: JNLP java.nio.channels.ClosedChannelException with kubernetes-plugin

what kubernetes cluster are you using and with what networking? my guess is that the underlying network layer is killing inactive connections On Thu, Apr 25, 2019 at 4:24 AM Gopi Naidu wrote: > I have got into the same issue. Could you please post the solution if > anything is found? > > On

Re: JNLP java.nio.channels.ClosedChannelException with kubernetes-plugin

I have got into the same issue. Could you please post the solution if anything is found? On Sunday, 26 August 2018 04:28:47 UTC+5:30, fabio...@singlepoint.ie wrote: > > I'm using Kubernetes plugin 1.12.3 > > The plugin works well, but whenever the agent does streams the log to the > master, it

Re: JNLP java.nio.channels.ClosedChannelException with kubernetes-plugin

Fabio, Did you find the solution for this issue or the root cause? Thanks Roey On Sunday, August 26, 2018 at 12:12:27 PM UTC+3, fabio...@singlepoint.ie wrote: > > Interesting that if I have a loop with some activity as follows, the job > completes successfully, so it seems to be a constraint

Re: JNLP java.nio.channels.ClosedChannelException with kubernetes-plugin

Interesting that if I have a loop with some activity as follows, the job completes successfully, so it seems to be a constraint of 60 seconds of inactivity perhaps between the agent and the jnlp? for i in 1 2 3 4 5 do echo "Looping ... number $i" sleep 50 done On Sunday, August 26, 2018

Re: JNLP java.nio.channels.ClosedChannelException with kubernetes-plugin

Hi Mark, thanks for the response. Yes, the pod meets the requirements: - I haven't provided a jnlp container - cat is the command - tty is enabled - Additional default container: maven:3.5.4-jdk-8 (tried different ones) What I've observed: - A build step with a sleep of 60 seconds always works -

Re: JNLP java.nio.channels.ClosedChannelException with kubernetes-plugin

Does the pod template meet the Kubernetes plugin constraints mentioned in readme? Specifically, is ttyEnabled and is there a long-running process that will allow the pod to continue running until the job completes? On Sat, Aug 25, 2018

JNLP java.nio.channels.ClosedChannelException with kubernetes-plugin

I'm using Kubernetes plugin 1.12.3 The plugin works well, but whenever the agent does streams the log to the master, it causes a connection exception as follows. To reproduce, simply create a free style job, restrict to run on the kubernetes pod, and add a step to sleep for 300 seconds.