[SOLVED] System Freeze Becoming More Frequent

coleman2247 · 2020-12-09 13:04:29

New SSD

I recently purchased a new nvme ssd to upgrade the storage I have. So I used dd to clone my old 128 GB sata ssd to my new 1 TB nvme ssd, I then resized the partition (with fdisk) and file system(resize2fs) so it would make use of the extra space. I did all of this from a live usb. After doing this I rebooted into my new ssd and disconnected the old one and everything seemed to work just fine, but about an hour in and my entire system just froze. I than rebooted and and everything was good for another 7 hours and it froze again. This freeze is nothing new and I have been trying to figure out the cause of it for quite some time, but it hasn't been a big issue for me and would happen maybe once a month before this hardware change, but with the new ssd it seems to happen much more frequently.

Error description

system completely freezes at random time(from 20 minutes to 12+ hours), not responsive to changing tty, ssh or any keyboard input
only way out of the error is to physically power off the system
display stays the same as the moment when the system froze
keyboard and mouse lights stay on

What I've tried so far

I suspect that it is hardrive related because the issues I'm seeing are very similar to ones on other posts
Checked for a /var/log/syslog and there is none
I've looked over
```
journalctl -b-1
```
and didin't seem to find anything that indicated a problem
I have ran
```
badblocks -v /dev/nvmen0
```
and it showed no bad blocks on the new drive
I have ran
```
smartctl -H
```
on all my drives and they all passed

Logs

journalctl

fstab

# /dev/sda2
UUID=ID0	/         	ext4      	rw,relatime	0 1

UUID=ID1       /swap           swap            rw,noatime      0 1


# /dev/sde2
UUID=ID2 /media/New\040Volume1 ntfs-3g remove_hiberfile,rw,nofail,uid=1000,gid=1000 0 0

# /dev/sdd2
UUID=ID3 /media/New\040Volume auto rw,nofail 0 0

Software Specs

x86_64 Linux 5.9.12-arch1-1
nvidia 455.45.01-5
i3-gaps 4.19-1
amd-ucode 20201120.bc9cd0b-1

Hardware

Ryzen 1700
Zotac Gtx 970
Corsair Vengeance LPX 32GB (4x8GB)
Asrock B450 AMD B450 AM4 ATX
Crucial P1 1TB 3D NAND NVMe
Seagate BarraCuda 6TB
Seagate BarraCuda 4TB
Samsung 850 128 GB (Windows drive and is usually never mounted)
Seagate 1TB SSHD (Usually never mounted)

Last edited by coleman2247 (2020-12-18 10:17:21)

xerxes_ · 2020-12-09 14:38:49

Did that freezes happen/happening when you use Discord or urxvt? Or with other apps too?

coleman2247 · 2020-12-10 00:41:36

xerxes_ wrote:

Did that freezes happen/happening when you use Discord or urxvt? Or with other apps too?

I have them open pretty much all the time, so it could be related to those two programs. I'll try using a different terminal emulator and use a discord alternative and see if the freezes persist.

Last edited by coleman2247 (2020-12-10 03:52:14)

Painless · 2020-12-17 20:16:09

Hi, I was wondering how your testing was going on, and if you use a dual monitor setup? I have a vaguely similar setup to you:

Latest kernel, nvidia 455, amd-ucode, fvwm

Ryzen 3 1300X
GT 1030 (Gigabyte)
Samsung SSD 960 EVO

I had been experiencing similar freezes since about the middle of August, not too often to begin with, but getting more frequent as time went on. Nothing unusual in journalctl output or in /var/log/Xorg.0.log.old, the log just ends without recording any errors.

I tried various options such as enabling nvidia-drm.modeset=1, also processor.max_cstate=1.

For me it seemed as though the freeze mostly occurred when moving the mouse pointer over a browser window (both Firefox and Ungoogled Chromium, but mostly Firefox) so I disabled hardware acceleration and vaapi in the browsers.

Sometimes it would lock when Xscreensaver was running, so I only enabled the blank screen option.

None of these strategies worked, I still got lockups. I stumbled across this quite old link - https://forums.developer.nvidia.com/t/r … 9/63708/15

I had purchased a second monitor in May, no issues at all to begin with (until August).

Just over week ago I tried using “UseNvKmsCompositionPipeline” set to “false” and ForceFullCompositionPipeline turned on for both monitors as per the link above and so far so good.

HTH

coleman2247 · 2020-12-18 10:16:23

Painless wrote:

Hi, I was wondering how your testing was going on, and if you use a dual monitor setup? I have a vaguely similar setup to you:
Latest kernel, nvidia 455, amd-ucode, fvwm
Ryzen 3 1300X
GT 1030 (Gigabyte)
Samsung SSD 960 EVO
I had been experiencing similar freezes since about the middle of August, not too often to begin with, but getting more frequent as time went on. Nothing unusual in journalctl output or in /var/log/Xorg.0.log.old, the log just ends without recording any errors.
I tried various options such as enabling nvidia-drm.modeset=1, also processor.max_cstate=1.
For me it seemed as though the freeze mostly occurred when moving the mouse pointer over a browser window (both Firefox and Ungoogled Chromium, but mostly Firefox) so I disabled hardware acceleration and vaapi in the browsers.
Sometimes it would lock when Xscreensaver was running, so I only enabled the blank screen option.
None of these strategies worked, I still got lockups. I stumbled across this quite old link - https://forums.developer.nvidia.com/t/r … 9/63708/15
I had purchased a second monitor in May, no issues at all to begin with (until August).
Just over week ago I tried using “UseNvKmsCompositionPipeline” set to “false” and ForceFullCompositionPipeline turned on for both monitors as per the link above and so far so good.
HTH

I've seemed to fix my issue (or at least it hasn't crashed in a long time). When I installed my new ssd I actually reset the bios and the reset changed the ram profile to "Auto" which set my ram to 2133 MHz (None of my sticks are 2133). I changed the XMP back to what it was and it hasn't crashed since. Granted I have only had the system on for at max 14 continuous hours with a total hour count of about 50. This still doesn't explain the less frequent crashes I was getting in the past. I do have 3 monitors so I'll try what you suggested. Since the crashes are no longer happening frequently I'm going to mark this thread as solved.

Arch Linux

#1 2020-12-09 13:04:29