GPU occupied by unkow process at boot
Hi everyone, I have just initialized a pod with four 3090 GPUs, but one of them keeps being occupied by an unknown process, as shown in the attached screenshot.
I have tried to terminate and re-initialize another pod with the same config, but it doesn't help. Could anyone help resolve this issue? Thanks.
7 Replies
Did you face this problem on Secure cloud or Community cloud? In my case I faced this on Secure Cloud as well.
@Aaron49
Escalated To Zendesk
The thread has been escalated to Zendesk!
Ticket ID: #20592
I am also on the Secure Cloud and this issue has also been posted before. I think we just need to wait till the tech support help us fix this problem 👀
Hope they really do something about it this time. Rather than simply saying - "try a different GPU".
I frequently get this when the pod before me was still shutting down but in most cases it clears up within a minute or two. If not switch pods if you can (and report the faulty one)
Mine has still been occupied for about 24 hrs. Switching pods is not feasible for me since I need exactly 4x3090, but there are only 7 of them available in my region, so I could only wait for help from the tech support...
@Aaron49 replied to you in the ticket