[Solved] Crash at boot after adding additional nvidia display adapter

keepitsimpleengineer · 2020-08-02 20:03:00

seth wrote:

Try to boot "nomodeset" and whether that allows you to
a) boot w/ the nvidia driver
b) inspect dmesg/the journal
So I made a hardware change, added a GTX560Ti to my GTX1070Ti system, and it borked my archlinux installation, crashing during boot.
After returning to original hardware configuration, arch still crashes during boot.
Would actually suggest a HW issue (ie. missing dedicated power supply, wrong PCI slot, …)
Otherwise, since you mentioned windows, see the 3rd link in my signature (if windows is hibernating, this can cause all sorts of weird issues)

So the Windows 10 hibernate turns on a (dim) light-bulb. So I booted both Win10 (SSD&HDD-backup) and I killed all sleep, hibernate and hybrid sleep options and turned off fast startup, plus moved all virtual memory to the respective system partitions (SSD&HDD), Best practice I think to isolate as much as possible.

Next I opened the system using nouveau and reinstalled nvidia, then reboot with nouveau blacklisted, The system again crashed during boot.

The journal for this attempt is @ https://pastebin.com/PbQtVYMM

Going throughthe journal I picked up these:

-- Logs begin at Wed 2020-03-18 17:52:50 PDT, end at Sun 2020-08-02 11:41:22 PDT. --

Aug 02 11:34:56 KISE-005 kernel: microcode: microcode updated early to revision 0x2f, date = 2019-02-17


Aug 02 11:34:57 KISE-005 audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 msg='unit=dracut-shutdown comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
Aug 02 11:34:57 KISE-005 mount[447]: Windows is hibernated, refused to mount.
Aug 02 11:34:57 KISE-005 mount[447]: Falling back to read-only mount because the NTFS partition is in an
Aug 02 11:34:57 KISE-005 mount[447]: unsafe state. Please resume and shutdown Windows fully (no hibernation
Aug 02 11:34:57 KISE-005 mount[447]: or fast restarting.)


Aug 02 11:34:58 KISE-005 kernel: nvidia: loading out-of-tree module taints kernel.
Aug 02 11:34:58 KISE-005 kernel: nvidia: module license 'NVIDIA' taints kernel.
Aug 02 11:34:58 KISE-005 kernel: Disabling lock debugging due to kernel taint

Aug 02 11:34:58 KISE-005 kernel: NVRM: loading NVIDIA UNIX x86_64 Kernel Module  450.57  Sun Jul  5 14:42:25 UTC 2020

Aug 02 11:34:58 KISE-005 kernel: [drm] [nvidia-drm] [GPU ID 0x00000100] Loading driver
Aug 02 11:34:58 KISE-005 kernel: [drm] Initialized nvidia-drm 0.0.0 20160202 for 0000:01:00.0 on minor 1

Aug 02 11:34:58 KISE-005 systemd[1]: Started Getty on tty1.

Aug 02 11:35:03 KISE-005 login[532]: LOGIN ON tty1 BY ljohnson

Aug 02 11:35:03 KISE-005 systemd[1]: Reached target Multi-User System.
Aug 02 11:35:03 KISE-005 systemd[1]: Reached target Graphical Interface.

Aug 02 11:35:49 KISE-005 systemd[1]: Received SIGINT.
Aug 02 11:35:49 KISE-005 systemd[1]: Stopping Session 1 of user ljohnson.
Aug 02 11:35:49 KISE-005 audit[532]: CRED_DISP pid=532 uid=0 auid=1000 ses=1 msg='op=PAM:setcred grantors=pam_securetty,pam_tally2,pam_shells,pam_unix,pam_permit,pam_env acct="ljohnson" exe="/usr/bin/login" hostname=KISE-005 addr=? terminal=tty1 res=success'
Aug 02 11:35:49 KISE-005 login[532]: pam_unix(login:session): session closed for user ljohnson

⇝Aug 02 11:34:57 "Windows is hibernated, refused to mount." ~ so I suspect there is a hanging hibernation out there

Then ⇝

Aug 02 11:34:58 KISE-005 kernel: nvidia: loading out-of-tree module taints kernel.
Aug 02 11:34:58 KISE-005 kernel: nvidia: module license 'NVIDIA' taints kernel.
Aug 02 11:34:58 KISE-005 kernel: Disabling lock debugging due to kernel taint

and then ⇝

Aug 02 11:34:58 KISE-005 kernel: NVRM: loading NVIDIA UNIX x86_64 Kernel Module  450.57  Sun Jul  5 14:42:25 UTC 2020

But if you check a successful nvidia boot from 7/15 { https://pastebin.com/fRJLpEWK } all these are there

Next step is to try nomodeset boot and see what can be done about the hanging hibernation.

keepitsimpleengineer · 2020-08-02 20:33:05

seth wrote:

Try to boot "nomodeset" and whether that allows you to
a) boot w/ the nvidia driver
b) inspect dmesg/the journal

Boot successful with nvidia

Journal @ https://pastebin.com/LhXe2p17

Dmesg @ https://pastebin.com/drrYnx9f

Terrific!

In a nutshell, what does nomodeset do or not do... .. ?

seth · 2020-08-02 20:51:59

It prevents https://wiki.archlinux.org/index.php/Ke … de_setting - you won't be able to run a graphical display server (except on the VESA driver)

Aug 02 13:11:01 KISE-005 kernel: You have booted with nomodeset. This means your GPU drivers are DISABLED
Aug 02 13:11:01 KISE-005 kernel: Any video related functionality will be severely degraded, and you may not even be able to suspend the system properly
Aug 02 13:11:01 KISE-005 kernel: Unless you actually understand what nomodeset does, you should reboot without enabling it

The kernel module loads nevertheless and there're no errors.

However windows still seems to hibernate:

Aug 02 13:11:02 KISE-005 mount[436]: Windows is hibernated, refused to mount.
Aug 02 13:11:02 KISE-005 mount[436]: Falling back to read-only mount because the NTFS partition is in an
Aug 02 13:11:02 KISE-005 mount[436]: unsafe state. Please resume and shutdown Windows fully (no hibernation
Aug 02 13:11:02 KISE-005 mount[436]: or fast restarting.)

And my money remains on this being the cause.

keepitsimpleengineer · 2020-08-03 15:13:37

seth wrote:

It prevents https://wiki.archlinux.org/index.php/Ke … de_setting - you won't be able to run a graphical display server (except on the VESA driver)

Aug 02 13:11:01 KISE-005 kernel: You have booted with nomodeset. This means your GPU drivers are DISABLED
Aug 02 13:11:01 KISE-005 kernel: Any video related functionality will be severely degraded, and you may not even be able to suspend the system properly
Aug 02 13:11:01 KISE-005 kernel: Unless you actually understand what nomodeset does, you should reboot without enabling it

The kernel module loads nevertheless and there're no errors.

However windows still seems to hibernate:

Aug 02 13:11:02 KISE-005 mount[436]: Windows is hibernated, refused to mount.
Aug 02 13:11:02 KISE-005 mount[436]: Falling back to read-only mount because the NTFS partition is in an
Aug 02 13:11:02 KISE-005 mount[436]: unsafe state. Please resume and shutdown Windows fully (no hibernation
Aug 02 13:11:02 KISE-005 mount[436]: or fast restarting.)

And my money remains on this being the cause.

Well in any case your suggestion about 'nomodeset' solved the problem, adding it to the kernel resolved the issue regards properly booting with nvidia. Thanks a bunch. Turns out this was occasionally happening in other distributions particularly Ubuntu. 'nomodeset' instructs the kernel to defer modesetting for the X to do it, the legacy method, the only difference, so far, is to deny fancy screen stuff until X starts running.

With regards to hibernation in Windows, I had turned off hibernation in the Windows GUI but this did not delete the 'hiberfil.sys' in Windows system root directory. To do this requires using CMD or powershell ~ which I have now done.

journal for success @ https://pastebin.com/BJtre06N

Last edited by keepitsimpleengineer (2020-08-03 18:24:15)

loqs · 2020-08-03 21:26:42

keepitsimpleengineer wrote:

'nomodeset' instructs the kernel to defer modesetting for the X to do it, the legacy method, the only difference, so far, is to deny fancy screen stuff until X starts running.

nomodeset disables kernel mode setting for that boot. The vesa X11 display driver works without KMS not sure on the nvidia one.

seth · 2020-08-04 06:16:43

The xorg log will tell, but I'm pretty sure it runs on the VESA driver.
You could try to avoid a framebuffer console instead of nomodeset, https://wiki.archlinux.org/index.php/GR … ramebuffer

keepitsimpleengineer · 2020-08-04 17:47:57

seth wrote:

The xorg log will tell, but I'm pretty sure it runs on the VESA driver.
You could try to avoid a framebuffer console instead of nomodeset, https://wiki.archlinux.org/index.php/GR … ramebuffer

Xorg.0.log @ https://pastebin.com/euqbx6i4

seth · 2020-08-04 18:54:53

You're using the nvidia driver despite nomodeset - you're good.

Arch Linux

#26 2020-08-02 20:03:00

Re: [Solved] Crash at boot after adding additional nvidia display adapter

#27 2020-08-02 20:33:05

Re: [Solved] Crash at boot after adding additional nvidia display adapter

#28 2020-08-02 20:51:59

Re: [Solved] Crash at boot after adding additional nvidia display adapter

#29 2020-08-03 15:13:37

Re: [Solved] Crash at boot after adding additional nvidia display adapter

#30 2020-08-03 21:26:42

Re: [Solved] Crash at boot after adding additional nvidia display adapter

#31 2020-08-04 06:16:43

Re: [Solved] Crash at boot after adding additional nvidia display adapter

#32 2020-08-04 17:47:57

Re: [Solved] Crash at boot after adding additional nvidia display adapter

#33 2020-08-04 18:54:53

Re: [Solved] Crash at boot after adding additional nvidia display adapter

Board footer