Hardware: HP Z840 PC
Graphics cards: NVIDIA P2000, NVIDIA P6000.
OS: Red Hat Enterprise Linux 6.9, Kernel version 3.10.0-514.rt56.228.el6rt.x86_64 –
NVIDIA driver version: 367.55
The problem we are having is that most of the time we boot, x-Windows does not come up.
After the red hat version banner appears, the system appears to be hung. We can ssh into the system from a remote terminal.
We see this stack trace from what appears to be an NVIDIA device driver crash in /var/log/messages,
Feb 14 10:45:19 msg0-0 kernel: [] os_acquire_mutex+0x3d/0x50 [nvidia]
Feb 14 10:45:19 msg0-0 kernel: [] _nv016701rm+0x18/0x30 [nvidia]
Feb 14 10:45:19 msg0-0 kernel: [] ? _nv018815rm+0x3d/0x120 [nvidia]
Feb 14 10:45:19 msg0-0 kernel: [] ? _nv017365rm+0xfb/0x200 [nvidia]
Feb 14 10:45:19 msg0-0 kernel: [] ? _nv000815rm+0x225/0xbb0 [nvidia]
Feb 14 10:45:19 msg0-0 kernel: [] ? rm_execute_work_item+0x49/0xc0 [nvidia]
Feb 14 10:45:19 msg0-0 kernel: [] ? os_execute_work_item+0x50/0x80 [nvidia]
Our goal is to first, boot successfully into Red Hat with x-Windows, and then configure the P2000 to be used for the display, and the P6000 to be used for CUDA processing.
We removed the P2000 and tried rebooting(after updating the xorg.conf file), but failed to start x-windows 3 out of 4 times. The problem has always been intermittent, whether we have two cards installed, or one card.