A user is not usually allowed to connect to an arbitrary compute node through an normal `ssh` connection. This is a way to protect the node on which your jobs are running:
```
#[ctchiling@pollux running]# squeue -u ctchiling
JOBID PARTITION NAME USER ST TIME NODES NODELIST(REASON)
#[ctchiling@pollux running]# ssh p040
Warning: Permanently added 'p040,10.3.0.80' (ECDSA) to the list of known hosts.
Access denied: user ctchiling (uid=34213) has no active jobs on this node.
Authentication failed.
```
You can't enter in the node when you don't have a job in there.
But what if your own job is running on a compute node ? then you have a legitimate motive to connect to that job (to run a [top](https://en.wikipedia.org/wiki/Top_(software)), [gstack](https://linux.die.net/man/1/gstack) or [link](https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7/html/developer_guide/debugging-running-application) a debug session on you running process) and `slurm` has been configured to allow
