Question Prebuilt System freezes after restarting (two OSes installed) ?

Aug 11, 2023
2
0
10
Hi all! New to the forum with an intriguingly weird problem:

I have a new, prebuilt system that has two processors and one graphics card. I installed Debian 12 along with the existing Windows 11 (Linux installed on another M.2 drive). The system would freeze within around 10 minutes after startup after I perform an OS restart. A complete shutdown (either using OS menu or power button) doesn't seem to cause this. Specs below:

Desktop PC model: Lenovo Thinkstation PX

CPU: 2x Intel Xeon Gold 5416S
RAM: 8x 16GB DDR5-4400MHz
GPU: Nvidia RTX A4500
SSD: Windows 11 on Lenovo 4 TB SSD M.2 2280 PCIe Gen4; Linux on WD_BLACK 4TB SN850X

(sorry but didn't bother to open it up one more time to check the specs)

The problem is worse when I boot into one OS, select restart, and choose the other OS in GRUB, which guarantees a freeze within 10 minutes. It can also freeze after restart in BIOS settings, or even on the GRUB selection page.

Things I have tried:
  • Checking Linux Syslog (doesn't capture anything).
  • Updating all drivers and BIOS (both Windows Update and Linux apt).
  • Running hardware diagnosis (both from BIOS and on Windows; everything passes).
  • Turning off fast boot (both in Windows and in BIOS) and secure boot.
  • Resetting all BIOS options (have to change the boot order to get into GRUB).
  • Reseating GPU (partly for fun).
Some observations:
  • Both CPU and all RAM can be detected by both OS's and the BIOS.
  • The freeze can happen regardless of CPU temp. I have seen it happen at 40°C.
  • If I boot into Debian first and then choose to restart and try to go into Windows, the system seems to turn itself off immediately and then back on.
  • If the freeze happens when I am in Debian, the system will turn itself back on. Doesn't seem to do so when it freezes in GRUB or BIOS.
  • If I shut down completely and then start into either OS, the system can run smoothly for a few days but might crash again with CLOCK_WATCHDOG_TIMEOUT......
  • I have been getting CPU-related error codes on the chassis LCD, this only happened once when I was stuck in GRUB.

As this is prebuilt I do have warranty support, the support person seems to think that this is a motherboard issue and has ordered a new motherboard to be sent. As the part is backordered you're seeing my attempt to solve this by myself. Any advice is appreciated! (Sorry for the long text lol).
 
Welcome to the forums, newcomer!

Curious as to how you installed both OSes to get a dual boot going...? Once you'd verified that the BIOS was flashed to the latest version, did you clear the CMOS?
With Windows 11 preinstalled, I added an additional SSD to the system (WD_BLACK mentioned above) and prepared a bootable Debian 12 USB. Debian was installed on the second SSD. After installation, the boot order was changed so that the second SSD was booted first to get into GRUB.

I have tried to clear CMOS to reset BIOS options once. Used the factory reset option provided by the BIOS.

(I'm pretty sure that the issue was ongoing before Debian got installed since I distinctively remember my first attempt at installing Debian also froze while I was in the USB installer. Cold restart and then going into the bootable USB was successful and Debian is on.)
 
One thing I discovered many years ago when booting up a machine from three separate drives, with Windows XP, 7 and 10, was the computer would sometimes run CHKDSK on all the hard disks and SSDs, after switching from Windows 10 to 7 or XP and back to 10.

I "fixed" this anomaly by disabling Windows 10 "Fast Startup" mode in Power Options. For good measure, I also got rid of hiberfil.sys, which I don't use on desktop PCs.
https://www.howtogeek.com/243901/the-pros-and-cons-of-windows-10s-fast-startup-mode/

Another tip. When installing any operating system, make sure all the other hard disks and SSDs are physically unplugged. Windows has a nasty habit of occasionally installing the boot sector on a different drive, which can be awkward.

If you had the Debian drive connected when installing Windows, or the Windows drive connected when installing Debian, you may have got your boot partitions scrambled.

I don't experience many problems changing the BIOS on multi-OS computers and booting from the required drive. I keep a backup Windows SSD in most machines, in case the main SSD goes bad and stops booting. I access the backup drive using a Function key for one-time boot, or dive into the BIOS boot order for a more permanent boot drive change.

If all else fails, you could try installing Debian as a VM under Windows in Hyper-V or VirtualBox, unless Debian needs direct access to the GPU.

On a final note, if you're running your RAM at XMP 4,400MT/s, I recommend booting from a MemTest86+ USB memory stick, to see if your memory is 100% stable at this speed.

You could temporarily disable XMP when installing any new OS. It would be one less thing to worry about.