Debian kernel 4.9.0-16
KDE Plasma version 5.8.6
System is an HP Elitebook 8770w, 24gb ram

When I try to put the computer to sleep, there is at least a 50% chance that it won't go to sleep, but instead freeze up (and then will be quite hot when I return), requiring a hard reboot. That reboot then also has a 50+% chance of freezing, sometimes right after I enter my password, other times waiting until it's looking like the windows are going to come back up.

Other times if it goes to sleep properly, I try to come back from sleep mode, and again, it often lets me put my password in but freezes shortly afterwards.

The machine does not generally freeze up apart from those times.

Any hints?

Any hanging processes you see in starting "htop" ?
Any error message in the output of "dmesg" ?
Currently I see a notebook hanging often with firefox. It looks like some applications have leaks.

That laptop is rather old. RAM doesn't always last. Use memtest86 to give it several hours' workout, at least 4, and the more the better.

My system doesn't seem to have htop?
Dmesg outputs pages and pages. The two that it highlights in red are:

[ 1228.318811] Web Content: Corrupted page table at address 2be425680980

and

[ 1228.319057] RIP [<00007f105e6fe2c5>] 0x7f105e6fe2c5

Firefox is the main thing I use this computer for. I did wonder if that had something to do with it.

And mrmazda, that's a good idea. It is indeed an old system. I will try to run memtest tonight.

Just to follow up: I did go through one full test just now. I realize that's only 2 hours, so I should go back and do more later, but as a preliminary it seems to point away from RAM as the culprit. I do agree with checking it, I once had a system with a hardware fault that caused random crashes.

One more clue I had forgotten to mention. In the course of rebooting the machine a few times while getting memtest86 set up, the machine kinda reminded me that it generally freezes while attempting a reboot -- as in, before it starts the reboot. The screen will go blank, or in the most recent crash it went to the Debian text login prompt (at the top left of a black background, so outside KDE) but it wouldn't let me enter any text.

That's what is so frustrating here, it's like there is sort of a pattern to when it freezes, but only sort of.

Those indicate that potentially something is corrupted in the file system.
It would be a good thing to boot from a live install media and run fsck on all the different partitions/filesystems on your drive just to make certain the file system is good. It also would still be good to test the memory as well

24GB of ram is an unusual configuration for something coming direct from the manufacturer. 8GB of ram would be the usual for a laptop of that era. I suspect that at some point someone may have done a ram upgrade and the upgraded ram may not be compatible with that motherboard. AMD boards tend to be more finicky about their ram. Please post your ram configuration, i.e how many ram sticks and their size.

Ram issues tend to result in the type of erratic problems you're describing, i.e random freezes. You probably have 3 sticks of 8GB ram. You can try pulling two of the sticks and run with one ram stick to see if your problems go away. Check your manual to see where a single ram stick should be installed.

1 members found this post helpful.

In my case this was the solution:
I use nouveau driver.

https://bugzilla.kernel.org/show_bug.cgi?id=56461

I found it odd for Linux to call a RAM error page table, this is typically a MS Windows name for SWAP. Little googling found this link. While Chrome is not the root cause of this issue, it is something that can expose the error.

This does look, like as mentioned above, to be a hardware issue.

Run both the check on the RAM and the HDD. One of them is potentially on the fritz.

Yeah. I had one chromebook that would lock up after a hour or so. Running ChromeOS even that is built for that netbook.

Replacing the ssd fixed it. Aint saying that is your problem. But there are tools to test ssd drives.

https://www.simplified.guide/linux/disk-health-check

Well, that's an interesting point.

I'm the guilty party there, I'm afraid. It came with two "Ramaxel" 4gb sticks for a total of 8gb, and I added two 8gb Crucial sticks. On the other hand I just looked up when I had bought them, and it was in early 2019. So they've been there for almost 3 years, and the problem didn't crop up until the past month or two.

Perhaps I should pull the two Ramaxel 4gb sticks and try just running the Crucials?

I had tried an ssd awhile back and didn't like it (I forget what the odd behavior was, but there was one particular area where it was not like a normal hdd) so both drives installed in this machine are normal hdd's. I did run the smartmontools check and both passed.

Skimming through the above, if not already done, check & re seat disk drive cables & ram modules; you may have a damaged drive cable, try a replacement.

Nomodeset is virtually always a workaround, not a solution. Some NVidia (non-FOSS) drivers may need it. All competent FOSS display drivers for AMD, Intel and NVidia GPUs, which includes nouveau, require KMS, which nomodeset disables. When using nomodeset, resolution and performance is seriously limited.

I recommend #6: run fcsk
htop not installed? in a terminal give "su -" then return, give your root pw, then "apt-get install htop"