Hello, I was using jupyterhub on casper today and I got kicked off and now it won't let me start a new session. It's stuck on the "Your server is starting up" page, which eventually times out. I've got a new jupyterhub session going on Cheyenne, but I will need to be back on casper to access certain files. @Jared Baker , would you be able to help me sort this out?
I'm also having issues getting onto casper with jupyterhub. Stuck on "Your server is starting up" with a red bar.
Others say their jobs that were running crashed
Might not be jupyter specific, but PBS on casper because no jobs show up for me with qstat @casper
CISL resource status page (https://www2.cisl.ucar.edu/user-support/cisl-resource-status) shows Casper job scheduling system is down.
Thanks @Stephen Yeager !
Yes, something more related to PBS as noted. We're working the issue.
I've had multiple occurrences of "Server unavailable or unreachable" errors this morning with jupyterhub on casper batch nodes.
Are others experiencing this as well?
An aspect of these errors that is confusing to me is that the underlying casper batch job is still running.
Me too. I've gotten kicked off 3 times now... yep, like @Keith Lindsay , I just looked and my jobs are still running on casper too.
me three
When attempting to restart, I sometimes get a popup with the following message
Spawn failed: The Jupyter batch job has disappeared while pending in the queue or died immediately after starting.
@Jared Baker any idea of what is going on there?
Same here, started on and off yesterday pm for me....
Do you have approximate times for these this morning?
Maybe around 11:30?
There were some PBS woes earlier this morning, but I can go digging more.
Mine have all happened between 9 and 11am
Around 11:20 for me.
It was working well till ~10:30am-present
One of the occurrences for me was right before I posted, so about 11:15.
I'm not sure about the times before that.
@Jared Baker it is working for me now
Okay, I'm still skeptical on a number of things right now. But I'll post more once I feel a bit more confident.
I filed a CISL ticket yesterday at ~4pm reporting Jupyterhub instability on casper batch nodes. This seems to be a very recurring issue.
Okay, I think I've got things back to a reasonable state over the course of the afternoon and tried to address some issues when the PBS server fails to communicate back to the JupyterHub server. Hopefully this helps with stability.
Thanks @Jared Baker Quite stable this pm on my end!
I'm having the same error message as @Keith Lindsay reported above this morning... :/
Last updated: May 16 2025 at 17:14 UTC