As Xpander said is either PSU or RAM but I will go for PSU .
It doesnt matter if you have a lot of stuff or not, if your PSU doesnt give enough power to your RAM,CPU and GPU when they need it then you gonna have these kind of crashes.
Sometimes your computer is working fine because maybe at that time your PSU handles everything ok and sometimes NOT because your PSU doesnt handle everything correctly , the same goes with RAM.

If you dont have spare parts you better take your computer to a technician,tell them your problem and that you suspect RAM or PSU to be the problem.
Hopefully if they are good technicians they will identify the problem and change the faulty part.

As a technician, if I had to deal with such a problem I would change the PSU and make you pay for it.You will have 14 days to test it and return it back in case is not that the problem.
In case PSU was not the problem I would do the same with the RAM.Finally if none of the above was the problem I would have to keep your computer for some days to test it.I would upgrade BIOS and do some other motherboard tests etc. Maybe I would check as well for Kernel Upgrade , maybe try to install a fresh OpenSuse distro or another Distro with a different hard disk and test it out and so on...
If they tell you is because you install Linux and have such problems then laugh at them and go to another technician :P

Hi guys,

I have some news. I was able to reproduce the freeze on Windows, after all. It wasn't easy, but I was playing Just Cause 2, and I got 2 freezes, 1 each day. It isn't as much, but it happens there too.

This is important because I can now scratch the premise that it was not an hardware problem.

What I also did was to test the RAM, using only one card at a time, and running that RAM test from the GRUB. The cards past the test, and I got freezes on Linux using either one of the ram cards. So I think it's safe to assume it's not a RAM problem.

So, the PSU it's still a strong possibility. But it could be a motherboard problem, too (I hope not), or even CPU? Itś not the hard drive, because windows and linux are on 2 different ones.

I'll contact a tech guy and take the computer to him to see if he can figure it out the problem. Thank you for the help.
Please feel free to keep your suggestions and thoughts coming, it's always useful to have opinions from experienced people.

Edit:
@Xpander

Quotecheck your dmesg, journalctl -b logs to see whats up, maybe there are hints somewhere.

Oh, I forgot about this. Will do, when I get home.

Is 2 year old that old, for a PSU, Feda?

@Xpander

This is so much information:
http://pastebin.com/PrJkfFtT

This got my attention:


[    6.608976] NVRM: Your system is not currently configured to drive a VGA console
[    6.608981] NVRM: on the primary VGA device. The NVIDIA Linux graphics driver
[    6.608983] NVRM: requires the use of a text-mode VGA console. Use of other console
[    6.608985] NVRM: drivers including, but not limited to, vesafb, may result in
[    6.608987] NVRM: corruption and stability problems, and is not supported.

But it's all chinese to me.

Quoting: PTLyonIs 2 year old that old, for a PSU, Feda?

@Xpander

This is so much information:
http://pastebin.com/PrJkfFtT

This got my attention:


[    6.608976] NVRM: Your system is not currently configured to drive a VGA console
[    6.608981] NVRM: on the primary VGA device. The NVIDIA Linux graphics driver
[    6.608983] NVRM: requires the use of a text-mode VGA console. Use of other console
[    6.608985] NVRM: drivers including, but not limited to, vesafb, may result in
[    6.608987] NVRM: corruption and stability problems, and is not supported.

But it's all chinese to me.

nothing to worry about the nvidia messages, its all fine.

you dmesg seems fine also, just seems your firewall throws lots of shit into it :)

if you get freeze it would be still good to see your

journalctl --since today or similar output, there might be either some system dump or some clues what went wrong

I can't find much information about it in english, most of it is in spanish/portuguese. Probably never got to the US market, but in Portugal/Spain seems to be a fairly common PSU.
http://www.game-debate.com/psu/index.php?ps_id=821&psu=Nox%20Urano%20SX%20500


The system was built by a tech guy that I know. He made some build suggestions, I chose some parts based on them, and he built it. It's the same guy that will try to diagnose the problem, early next year. The problem was that, while I was saving for a new desktop, my old laptop just died. So I had to anticipate the purchase, with a limited budget (and I had to buy a monitor too). I was the one that chose the AMD cpu because it had slightly better benchmarks than the Intel counterpart for the same price. I also went for a very cheap nvidia gpu because I had no budget for more and nvidia drivers worked better on linux (or so I was told), and I was planning to change it later. It took me 2 years, though, to change it for a GTX 960. But it was a cheap build, that's true.

I didn't try to disable the on-board graphics, no. It never occurred me that that could be a problem. I'll try that, thanks ;)

Quoting: Xpanderif you get freeze it would be still good to see your

journalctl --since today or similar output, there might be either some system dump or some clues what went wrong

Just got 2 or 3 in a row, but the log covers only a few minutes...

http://pastebin.com/uPbL5i5H


Edit: And this is the auth.log for today. I've been at the computer for a short period of time, and got 3 crashes. So, it's a short one. Does it help?

http://pastebin.com/fD7Ma7dA

How freaking frustrating.

I had a similar problem, with nearly identical symptoms. Although the system was incredibly different than yours. I'll share the story, just in case it might help.

It was my wife's Asus laptop. It has intel CPU and AMD GPU. Upon receiving the computer new, I immediately installed whatever the current version of Ubuntu was (a couple years ago). It was technically a dual-boot machine, but never booted to Windows 7 again, after the very first day. It ran perfectly for about 6 months. Then it started freezing/crashing randomly in linux.

I closely monitored temperatures, tested all hardware as best I could. I dug through the logs looking for clues (nothing found). I tried everything in my power (software wise) to make it run linux without crashing (I'm a fairly experienced linux user). I could spend hours telling you the various ways I tried to figure out the issue. But it would crash at least a few times per day, sometimes even more often than that. I could not figure it out.

So we resorted to booting Windows on it, and the freezing apparently stopped. Our solution was to have her computer make a remote desktop connection to a desktop that was running linux upstairs, to run specific apps remotely. With a fast local connection, this remote desktop set-up worked for a while. But then we realized the crashing had not stopped. It only was just much less frequent in Windows than linux. So, it was a hardware problem after all. It would run fine for days. Then it would crash 5 times in an hour. Then it would run fine for a week. Then it would crash once, and run for two weeks again without problem. It was just random.

But when it would crash while running windows, sometimes it borked the operating system so badly we needed to reinstall windows, from scratch, from the recovery DVDs (blue screen of death immediately upon boot, recovery from the recovery partition failed). Of course it took hours and hours (pretty much all day) to reinstall windows from scratch and get it updated each time. So although the crash happened much less often in windows, it was even more difficult to deal with than the more frequent crashes in linux. And on really frustrating days, it could crash during the re-installation, making it take forever.

After wasting enormous amounts of time reinstalling Windows multiple times, I contacted Asus customer support, and went through their trouble shooting routines on the phone. Of course this did not help. They agreed it was hardware and I sent it in as directed (no battery, no PSU). It was authorized for repair on at least 5 separate occasions (not kidding). They reinstalled windows once, which of course did not fix the problem. They replaced the motherboard twice, which did not fix the problem. They tried replacing other parts too. Every time they worked on it, the tech would note that the problem was duplicated in their repair facility, and then tested fine after whatever service they did.

I was pretty angry at this point. I was frustrated with the limited knowledge of customer support. I was frustrated with the repair techs sending it back with the same issue present. I was frustrated Asus just wouldn't give me a refund. This had been going on for about 6 months now, with our computer being under repair (or shipping back and forth across the country) for weeks at at time. Finally, I got a phone call directly from the tech who was working on it. He wanted to just give us a replacement B-stock machine. I was so angry with Asus, I just told him I didn't want it. I was just going to buy a new computer. He was trying to convince me to at least try the replacement computer, but I didn't really care if they ever got windows to work on it. What I (and my wife) really wanted was to run linux. I was just going to buy a more expensive linux laptop (system 76, zareason, etc).

But after being on the phone with the tech guy a few minutes, we actually started talking about the problems, and exactly how and when they would happen. We talked about thermal issues. We talked about load issues. He realized that I was a competent computer user, and trouble shooter, and I knew what I was doing. And then he finally told me that they HAD NOT EVER duplicated the problem in their repair facility. They ran it all the time when it was there, and used it extensively to even handle their own work duties while it was in the facility. They were stumped to find the problem. I was surprised by this because the service ticket had said that they had duplicated the problem EACH TIME it had been sent in. But that was apparently a lie. He told me they just marked that they had duplicated the problem, because he needed to do that in order to open it up and do any work on it.

And at that moment, we realized what must be happening. It was the PSU (which had never been sent in for service, as Customer Service had directed me not to send it in).

With the new PSU the computer has ran linux, absolutely stable, for several years now. And the windows partition has been collecting cobwebs, for years, too.

There's no way to be sure what your problem is, but you really need to look into the power supply.

Hi Kirkgun,

thanks for sharing your story. That was quite an adventure. I hope mine doesn't take that long to figure out the problem!

So, I have some recent news:

I noticed my sound system (cheap one that uses audio jack and external power supply) was making an odd noise on one of the speakers. I moved the cable around a bit, trying to understand what was the problem, but it didn't seemed to be from the cable. But the problem disappeared after a minute a so, when I wasn't even moving the speaker. I didn't paid much attention, until it happened again the next day. Then I though: "That's it! The sound system is having some kind of power related issue that only happens now and then, for some reason, but that problem it's also causing the PC to crash!". I removed the sound system, and for around 1 and a half, 2 days, I had no crashes. I was happy, and about to call the tech guy saying I had solved the problem. Then, in the next morning, BANG, 5 crashes! "Nooo, sh*t, sh*t, sh*t!!!". It wasn't the sound system.

Then, for 2 days, no crashes at all. On those days I played a lot XCOM 2 on Linux, and run an benchmark test on Ultra quality, on Windows. No problems at all on those 2 days. Until I tried to play Just Cause 2 just to test it on Windows (it was the last game where I had a crash on Windows). I restarted the checkpoint were I had had a crash (a car chase with a lot of shooting around), and it did crash again. I reseted and reloaded. Played that part of the game again, twice. No crashes.

Then it was the day I had appointed with the tech guy to deliver the computer, and I told him all this.

It's been 24h+ hours, and wasn't able to duplicate the error at all.

Random. I hate random.

You can check the power needed by your PC using this link (if your hardware isn't too recent because the calculator isn't up to date, but you can always look for equivalences by reading technical docs): http://apc.canardpc.com/index.php

From what I've read, I also think it's your PSU that cause the crashs you are experiencing (some PSU aren't reliable if the power consumption is too close to their theoretical maximal power, and some "China Export" PSU will not give you more than half the power they theoretically should deliver). :)

Riusma, indeed.

Ok, the tech guy said the PSU wasn't delivering enough power AND the temperature of the CPU is somewhat higher than it should be. The crashes are related to the PSU, he says.

He will change the PSU, add some watercooling, change the computer case (it's too small for watercooling), change the thermal paste, clean the fans, etc.

I could save some money by changing the PSU but not adding the watercooling, but the temperature problems are always a pain, so better safe than sorry. I really hope this solves the issue.

Thank you for allowing me free rant in here :P . I'll let you guys know if it worked.

PS: I just want to point out, once again, that this forum was the only place I did find help and advice. Linux gamers indeed needed a space for them only, like this one ;)