Occasionally unable to connect to Yggdrasil

Primary informations

Username: ceva3
Cluster: Yggdrasil


I connect to yggdrasil via x2go. I have been able to connect and use yggdrasil successfully with x2go for nearly the past year, so I fairly certain my configuration in x2go for connecting to yggdrasil is correct and not the source of this problem.

Starting yesterday afternoon, I was unable to connect to yggdrasil via x2go. I would login using the correct credentials (which I verified are still correct via Password Self Service), but then got the error “socket unknown” or, primarily, “Timeout connecting to login1.yggdrasil.hpc.unige.ch”. Note that yesterday morning I was able to connect without trouble, it was only at some point in the afternoon that I was suddenly no longer able to log in.

This is again happening today. This morning I was able to connect just fine, then starting ~30 minutes ago, I can no longer connect.

I double checked that this was not an x2go issue by using Termius to test the ability to connect, using the same credentials in Termius that I have set up in x2go. Termius never successfully connects, it just gets stuck in the “connecting” step.

Disclaimer: my apologies if I should not have posted this here, this is the first time I have submitted an issue regarding the UNIGE hpc clusters or used these forums.

Steps to Reproduce

Use x2go or Termius to connect to x2go. Host: login1.yggdrasil.hpc.unige.ch Login: ceva3 SSH port: 22 Session type: XFCE

Expected Result

To be able to connect to yggdrasil with x2go.

Actual Result

Got the error “Timeout connecting to login1.yggdrasil.hpc.unige.ch” when trying to connect to yggdrasil vai x2go yesterday afternoon (but not in the morning), as well as today, but only starting ~30 minutes ago.

I can suddenly, just now, connect again.

I also noticed that, via SSH.

Dear @William.Ceva and @Giuseppe.Chindemi is this possible that you had too many bad password attempt and that you ip was blacklisted for 15 minutes?

It is as well of course possible that login1 was under high load from a user who launched by “mistake” a job on it instead of using a compute node.

Please let us know the following detail if you have the issue again:

  • your ip address
  • the client you tried (x2go, putty, etc)
  • the hour and minutes when the issue occurred
  • the output of ssh -vv <user>@login1.yggdrasil.hpc.unige.ch or a screenshot