My Desktop PC occasionally freezes yet the panel clock seconds keep ticking.
 

Hi.

My 2013 machine makes a sound like it's about to shut off. It did power off once about 3 weeks ago for no apparent reason.

It freezes for up to 30 seconds and then comes back. Some days this never happens. Other days like today, it can occur multiple times.

In GNOME Disks Utility, the HDD looks ok. I tested the RAM with memtest86+ and it's was ok. Could it be a PSU issue? Is there a way in Debian to find out? Is '/var/log/syslog' the best place to look?

I'm using the Nouveau drivers and kernel 5.4.0-0.bpo.4-amd64. I'm not sure if this is a hardware or software issue.

Thanks.

I would also look at sensors (lm_sensors package): https://packages.debian.org/sid/lm-sensors
And, some BIOS have area to show hardware events (even shows when system powers off/on).

My machines make no sound at all when they're about to shut off.
Do you mean the sound it makes when it actually shuts off?
Then that's probably what's happening.

I like the lm_sensors idea - maybe it is a heat issue? How long since you've cleaned out dust bunnies? Or is this a laptop and you can't...

If the clock keeps time then the system is really working. The Desktop if having an issue. I'd start with top htop or other system metric to see more.

Beep might mean some keyboard buffer full?

When it freezes for a short while, the sound that the PC makes sounds like it comes from the DVD/CD drive.

Hi dc.901.

Here's the output for:

Hi ondoho. No, it's not the same sound.

Hi sevendogsbsd. It's a Desktop PC. I cleaned it out in March/April.

Hi jefro. There's no beep.

Temps look OK to me so probably not heat related.

Take a look at dmidecode output; it can be a lot, so this guide will help:
https://www.cyberciti.biz/tips/query...nd-prompt.html

For powersupply; compare the output below with actual specs of PSU, do they match:
dmidecode -t 39

And see if there is IPMI device:
dmidecode -t 38

There is system event log, but it may not be easy to read:
dmidecode -t 15

So, if there is IPMI device, it will be much easier to read hardware layer system event log, but I will be surprised if it is there on a desktop.

I disconnected the DVD/CD drive cables to see if that would stop the freezes but they still happen. I think the sound could be from the PSU or HDD - sounds a bit mechanical.

I've noticed lines with kernel in syslog around the time of the freezes. I wonder if I should revert to the original non-backport kernel? Could the backport kernel (5.4.0-0.bpo.4-amd64) be the issue?

In my experience, this can be a symptom of a failing HDD.

Hi allend. In GNOME Disks, it seems ok. See attached image.

I'm currently running this:

Anyone know roughly how long should it take on a 2TB SATA 7200 HDD?

I'd say it's the HDD alright but cannot confirm.

Some days my PC freezes a few times a day - once even during bootup recently. Yesterday, there were no freezes. As of typing this, 1h45m uptime - no freezes either today.

I reckon you're right there, allend.

The freezes usually happen when the system is under a heavy load.

No freeze for a couple of days now.

If the seconds are still going and you have another PC/Laptop. Try and ssh to the 'hung' machine and see if it is active. I think it is still active and once in you can poke around.

My guess, the video card is having an issues.

Next time this happens, ssh into it when it's frozen. Check the usual logs (/var/log/messages, /var/log/Xorg.0.log or ~/.local/share/xorg/Xorg.0.log, dmesg -t, any journalctl stuff that might be relevant, etc).

In my experiences, these freezes usually mean that the desktop is stuck on a READ operation that's taking an unexpectedly long time. So it can be a failing hard drive. I've also seen it in cases where the home directories are network mounts (I assume that's not your setup?), and I've personally seen Flash lock up X like this (when I forced the "experimental" 3D acceleration on).

It's been ages since the PC froze. 😕

Yestereve, I saw 16 bad sectors in GNOME Disks.

Edit: Today, it says '8' [see attached image]. 🤔

Another edit: It now mentions no bad sectors [see 2nd attached image]. 😕

Today, I heard my PC beep. Then I noticed that it was a read-only file system. I did a reboot and was met with a console. I ran 'fsck path_to_my_luks_partition' and exited to reboot. All is ok again. However, I've the feeling this is a striking warning.

GNOME Disks now says 8 bad sectors.

linustalman,

Install gsmartcontrol:

Right click on drive > Perform test > Extended self-test (30 minutes).

Read the output and if items #5 or # 197 are not zero, then you have a problem with that drive.

Hi beachboy2.

#197 had 8 - just as GNOME Disks said. Please find attached 2 screenshots re GSmartControl.

Thanks.

linustalman,

As I expected, your drive is about to fail.

Get your data off it pronto and then replace it.

Thank you, beachboy2. I will do so. 👍🏻

Not exactly but it can take quite a while. If you think you need to do this, kick it off and find something else to do for a few hours.

Checking a drive of that size takes long enough that I wound up creating a cron job that runs in the wee hours on Sundays to handle fscking the disks in an external USB drive cabinet containing three 2TB drives (it hits one drive per week). This has kept me from being blind-sided by a forced fsck should I be forced to reboot in the middle of the day.

Cheers...

^ yes but isn't badblocks much more time-consuming than a (standard, default, quick) fsck?
Anyhow, my first reaction when seeing 2TB on spinning rust: start the process, come back tomorrrow...

also, I think the process can be further slowed down by bottlenecks in RAM and CPU speed

Hello again, ondoho.

---

One day no bad sectors are detected in GNOME Disks, then 16, then none, then 112, now none again. That's very odd.

^ Oh, I think it has been very clear for a while that your storage is dying.
Posting screenshots & asking questions won't change that.

Can you answer my query from post #36?

I finally moved to a new HDD - another 2TB Seagate. No freezes so far, so it was likely the old HDD was the only cause.

Bad sectors must have been the issue with the old install.