Hi Everyone - recently purchased a GB10 DGX for our academic instituition. Trying to set it up fresh out of the box. Went to do a full local install on AI workbench and was met with “failed to install service dependencies: error code 0x10060004” this is a fresh boot any ideas? I did the latest update it requested. Attached are the log files. Did I miss a step? Saw something that mentioned there might be an issue with WSL ?
Thanks!
error log.txt (233.1 KB)
I think you omitted the steps you took to attempt installation?
have you updated the os and firmware?
did standard updating when first turning on - then was requested to do a sudo apt update when opening ai workbench. We just got this device today. This was a first run. Also did update through DGX dashboard .. still get the same result
Opened app > locations manager > full local install > installing … > fails to install service dependencies > says try again > repeat
Can you try updating on the command line (ssh?) and see if it shows any errors or applies cleanly?
Previously some forum members were reporting that they couldn’t complete updates without intervention.
sudo apt update
sudo apt dist-upgrade
sudo fwupdmgr refresh
sudo fwupdmgr upgrade
sudo reboot
I was able to reproduce this issue after getting my apt in a bad state. I would recommend rebooting your machine and trying again.
If that does not help can you also attach logs from ~/.nvwb/logs/workbench.log
what would put the APT in a bad state? this was a fresh boot right out of the box. I will give this a try and update you.
I also have a GX10 and had something similar happened to me. I don’t remember everything so I can’t say we have the same issue but i’ll share it just in case. When installing NVIDIA AI Workbench, I would get the dependencies install error as well. It appears the Nvidia AI Workbench was still installed though.However, when opening it up and clicking through the menus it would failed. I forgot the exact error but it seemed like a docker issue. So I went to see if I could manually start docker in the terminal. It would fail with: “Job for docker.service failed because the control process exited with error code. See “systemctl status docker.service” and “journalctl -xeu docker.service” for details.”
After looking at the logs I did the following:
# stop docker + socket activation so it doesn’t immediately restart
sudo systemctl stop docker docker.socket
# Delete the corrupted buildkit state
sudo rm -f -r /var/lib/docker/buildkit
# clear the “start request repeated too quickly” latch
sudo systemctl reset-failed docker
# start docker again
sudo systemctl start docker
# verify
sudo systemctl status docker --no-pager -l
sudo docker ps
After this, the NVIDIA AI Workbench worked for me.
thanks for this! Apparently 4 repo’s were getting blocked by our firewall. Allowed the IP’s through and did these updates. Problem is solved!
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.