-
Notifications
You must be signed in to change notification settings - Fork 937
Description
Thank you for taking the time to submit an issue!
Background information
What version of Open MPI are you using? (e.g., v4.1.6, v5.0.1, git branch name and hash, etc.)
4.1.6
Describe how Open MPI was installed (e.g., from a source/distribution tarball, from a git clone, from an operating system distribution package, etc.)
It's the .deb packages that are installed on Ubuntu 24.04.3
If you are building/installing from a git clone, please copy-n-paste the output from git submodule status.
Please describe the system on which you are running
- Operating system/version: Ubuntu 24.04.3
- Computer hardware: One self-built PC with an AMD CPU, and another is the NVIDIA DGX Spark
- Network type: Ethernet
Details of the problem
Please describe, in detail, the problem that you are having, including the behavior you expect to see, the actual behavior that you are seeing, steps to reproduce the problem, etc. It is most helpful if you can attach a small program that a developer can use to reproduce your problem.
Two systems, both run Ubuntu 24.04.3, on the same GigE network.
One is a homebuilt AMD CPU with an RTX 3090 GPU, called media, booted in graphical mode but no one is logged in locally.
The other is the NVIDIA DGX Spark (ARM CPU, Blackwell GPU), called spark, booted in text mode.
I'm ssh-ing into both machines from a MacBook Pro.
I cannot get rid of this error when trying to do mpirun from spark (text mode machine) to media (GUI machine):
florin@spark:~$ mpirun -H media uptime
Authorization required, but no authorization protocol specified
Authorization required, but no authorization protocol specified
18:01:02 up 17:51, 3 users, load average: 0.00, 0.00, 0.00
I've tried various workarounds I've seen online:
florin@spark:~$ HWLOC_COMPONENTS=-gl mpirun -H media uptime
Authorization required, but no authorization protocol specified
Authorization required, but no authorization protocol specified
18:04:52 up 17:54, 3 users, load average: 0.06, 0.02, 0.00
florin@spark:~$ mpirun --mca plm_rsh_args -x -H media uptime
Authorization required, but no authorization protocol specified
Authorization required, but no authorization protocol specified
18:05:02 up 17:55, 3 users, load average: 0.05, 0.02, 0.00
florin@spark:~$ mpirun -H media --mca plm_rsh_agent "ssh -x" uptime
Authorization required, but no authorization protocol specified
Authorization required, but no authorization protocol specified
18:05:12 up 17:55, 3 users, load average: 0.04, 0.02, 0.00
It's very frustrating. I need spark to be the "base of operations", to launch most commands from here. It's spark that has the most RAM, so I don't always need media.
The other way around it works fine:
florin@media:~$ mpirun -H spark uptime
18:01:29 up 19:01, 2 users, load average: 0.32, 0.11, 0.05
florin@media:~$
What else can I try?