Welcome to SparkyLinux forums
Zapraszamy również na polsko-języczne Forum https://forum.linuxiarze.pl

Help finding out why my system randomly shuts down, OS or hardware?

Started by PNWDrew, January 20, 2015, 07:09:35 AM

Previous topic - Next topic

PNWDrew

I'm full of questions recently...

edit: Forgot to add this--->Homebuilt PC.  AMD A10-7850K, Gigabyte F2A88XN-WIFI mobo. And of course Sparky....

My PC sometimes randomly shuts down during usage.  It also happens when it is idle with screensaver running. The fans remain on and the front power LED is lit; but I cannot wake the PC.  Reset and power buttons on the case not are functional.  Monitor shows no HDMI signal.   This is a recent development, it has happened 8-10 times in the last week, twice in about 10 minutes yesterday.  Earlier this evening there was an unprompted reboot also, shortly after starting it.  It has been up with no problems for 3hrs now however.

When this happens the only solution I've found is to turn off the power supply and leave it off for several minutes.; sometimes as much as 15 minutes. It will not power up if I try just flipping psu switch quickly.  Possibly related is the fact that neither shutdown nor hibernate work properly, each leaving the system unable to power back up without the same procedure. I have to use suspend to shut PC down if I want to restart it with the case button.

I returned all mobo settings to default, but only the RAM speed was altered. For whatever reason the OS only sees 1600mhz of 2133mhz RAM so I had increased it through mobo settings to see if that helped.  I don't really need the faster speed but since I bought higher speed I wanted to use it.  Returning to default has not stopped the problem.

Is there a way to see if the OS (Sparky) records anything odd prior to shutdown??  I looked at a variety of logs but I have no idea what I'm looking at. 

Drew



MoroS

Sorry for the lare reply. I've been a bit busy lately.

It seems like a power management issue. Did it also happen while you were working or only when the PC was idle or asleep? You can check /var/log/dmesg (and other log files in that directory) for anything useful. I'll look into any possible issues regarding that hardware.

Other than that, it seems that sleeping support has been troublesome lately. My old Intel-based Mac Mini also has similar issues when sleeping (doesn't wake up).
There's no such thing as "impossible". :)

PNWDrew

Quote from: MoroS on January 23, 2015, 02:28:38 PM
Sorry for the lare reply. I've been a bit busy lately.

It seems like a power management issue. Did it also happen while you were working or only when the PC was idle or asleep? You can check /var/log/dmesg (and other log files in that directory) for anything useful. I'll look into any possible issues regarding that hardware.

Other than that, it seems that sleeping support has been troublesome lately. My old Intel-based Mac Mini also has similar issues when sleeping (doesn't wake up).

Thanks for the response,

I think I've solved it, it's been up for 5 days no with no crashes.  I realized that the only hardware change I made since the initial build was to add a PWM case fan.  I unplugged it from the board since I didn't really need it and it is unnecessary noise.  AMD makes finding actual CPU temps difficult so I had decided to be cautious.
Since then not a single crash. 

Some things I've considered:
--Possibly the variable case fan speed (PWM) caused a power supply glitch at times. I haven't researched whether that is a known issue with these mobos. I hope not.
--The wires for the front switches run against that fan frame and the shutdowns were similar to a reset button push so maybe the fan motor was inducing a false reset signal or something.  I sometimes see some very odd wiring/noise issues at work so it isn't too far fetched.
--Perhaps the physical button is flaky and was actually closing the circuit. I actually had never used it prior to having this issue.
--I do recall that I had some problems determining polarity (if any) of the switch wires, a few of the connectors that run from the case to mobo headers were unmarked but I assumed that those were simply momentary switches and it didn't matter. 
--Maybe the PSU is bad
--It might also just be coincidence but I won't argue with success.


Thanks,
Drew

For the record this is my part list---I bought components thinking this would be a Windows PC, neither the motherboard nor CPU have good linux support so a hardware issue is possible.

CPU: AMD A10-7850K 3.7GHz Quad-Core Processor
CPU Cooler: Scythe BIG Shuriken 2 Rev. B 45.5 CFM Sleeve Bearing CPU Cooler
Motherboard: Gigabyte GA-F2A88XN-WIFI Mini ITX FM2+ Motherboard
Memory: Patriot Viper 3 Low Profile Red 8GB (2 x 4GB) DDR3-1866 Memory
Storage: Samsung 840 EVO 120GB 2.5" Solid State Drive
Case: Corsair 250D Mini ITX Tower Case
Power Supply: Rosewill Hive 550W 80+ Bronze Certified Semi-Modular ATX Power Supply

MoroS

Quote from: PNWDrew on January 28, 2015, 08:03:14 AM
I think I've solved it, it's been up for 5 days no with no crashes.

Does it sleep/hibernate without problems?
There's no such thing as "impossible". :)

PNWDrew

Quote from: MoroS on January 28, 2015, 03:52:53 PM


Does it sleep/hibernate without problems?

Nope.  the various power options have been flaky at best.
Hibernate has never worked.  If I get anything it comes up with an error involving Radeon when awakened. Something about bo still active in vm.
Reboot does not always reboot. or at least did not in the past. I've had enough problems when restarting the PC that I avoid doing it when possible. I'll save the constant reboots for my Win 7 work laptop.

I usually use suspend as shutdown when needed.   I'm not sure what suspend is actually intended to do but it shuts down immediately for me. 

If you are using a laptop of course these functions are rather important.

When I have time to risk it I will run through everything and get back to you. 

PNWDrew

After updating suspend behavior has changed.  It now asked for a password (twice?) and when awakened it was not a reboot.  It returned to full operation instantly.  It only took a few seconds to reboot before but if this is the new suspend action I like it.

pavroo

Upgrading from Debian repos changed something for sure.
Have you restarted after upgrading?
Nothing is easy as it looks. Danielle Steel

PNWDrew

Quote from: pavroo on January 30, 2015, 06:32:52 PM
Upgrading from Debian repos changed something for sure.
Have you restarted after upgrading?

Yes I did restart.  Upgrading definitely corrected my suspend/reboot/hibernate behavior. 

1) Reboot is normal now, previously would not work and caused difficulty upon restart.
2)Suspend powers all off instantly, restarts to Xscreensaver and requires a password.
3)Hibernate is the same as Suspend except a longer period before powering down.

Suspend occasionally asks for password prior to action, and always wants a pw when waking it up. 

I'm happy with current actions, seems to work as expected now.
Thanks!

MoroS

Quote from: PNWDrew on February 01, 2015, 11:45:04 PM
Yes I did restart.  Upgrading definitely corrected my suspend/reboot/hibernate behavior. 

1) Reboot is normal now, previously would not work and caused difficulty upon restart.
2)Suspend powers all off instantly, restarts to Xscreensaver and requires a password.
3)Hibernate is the same as Suspend except a longer period before powering down.

It's finally working as expected. ;) There were some kernel updates recenlty, so maybe some bugs were eliminated. :)
There's no such thing as "impossible". :)

View the most recent posts on the forum