Dear Stephan,

Thanks for your quick help and suggestions!

The session listing commands you mentioned didn’t reveal the users’ sessions. But while I was investigating those ideas, I was advising my users to try using an ssh -X session in an xquartz terminal as a workaround. that produced similar errors as were being reported in x2go (X11 connection rejected because of wrong authentication). Even when attempting to log onto the server in person using their user accounts, they were booted back to the login screen. So the problem was clearly outside of x2go.

I googled this new but related X11 message, and noticed it was sometimes associated with users having exceeded their disk quotas: https://www.cyberciti.biz/faq/x11-connection-rejected-because-of-wrong-authe...

This was the problem! I have home folder quotas, and the users had neglected our server rules and were saving large image files to their documents folder rather than our data volume. With quota enforced on their home folders, x11 and x2go were unable to write any of the files needed for authentication. This manifests as an “authentication error” even though this is a red herring. I will now know what to check if I see this error again for specific users.

I’m guessing this arises because a server can’t reveal errors during authentication for security reasons, so there's no way for a client to know whether there’s free disk space on the server until it has successfully authenticated. This likely means there's nothing an x2go client can do to know this was the problem, but perhaps it would be worth considering a helpful suggestion in this error that the problem *could* be attributable to insufficient disk space in the user’s home folder on the host machine?

As for brains: it's thoughtful of you to make these suggestions, although I’m afraid I'm not a radiologist or clinician, just a brain scientist who likes to use the interface for persistent matlab desktop sessions and running numbers. We do have quite a few users, so you'd be forgiven for thinking my lab was an outfit with actual resources, but all are students and we operate on a shoestring :-) This just makes us all the more grateful for your development efforts.

Cheers, Jordan

Jordan Poppenk, Ph.D. Canada Research Chair in Cognitive Neuroimaging Department of Psychology and Centre for Neuroscience Studies Queen's University http://popmem.com 613-533-6009