Dear Riccardo,
We have a cluster at Science Cloud (S3IT, project conti.systbot.uzh) with a frontend node and five compute nodes. We have problem to connect to compute nodes either via ssh from frontend or direct ssh to the compute nodes' IP addresses. This we noticed because all of submitted jobs fail after a few seconds without any error.
A possible cause of this issue could be that one of the users has done an upgrade on the frontend (sudo apt-get upgrade) and this has obviously changed some of the configuration files. Can you please help me solving this connection problem? Thanks a lot in advance!
All the best,
Narcis