Kernel panic - not syncing: Fatal exception in interrupt

 

I keep getting this error, earlier I had a CPU overheat error after changing CPU and after that I didn't get the same transfer speeds, changing from ASUS Optimal to ASUS Normal

in BIOS settings, which should not change the network transfer rates.

So now the computer keep crashing, and I don't have time to run a parity check before the server crashes.

 

I tried searching Kernel panic - not syncing: Fatal exception in interrupt

and found this thread

http://lime-technology.com/forum/index.php?topic=42142.msg400504#msg400504

 

After reading it, I am running a parity check, and put it into troubleshooting mode, and it crashed.

 

Log and diagnostic added.

syslog-20160615-024545.txt

skynas-diagnostics-20160615-0249.zip

Number one step should be to make sure your RAM is fine, so run Memtest from the boot menu, several passes.

It's 2 passses and 82% on third, seems fine so far. Any other suggestions?

How many passes should I go for? Atleast the server isn't shutting down during this process

That's probably plenty for the memory.  If nothing else turns up, there's a better Memtest86 out there, not included here, sometimes finds issues this one doesn't.

 

Neither syslog shows ANY issues at all!  Nothing wrong in the diagnostics either that I could see.  The only evidence of something amiss is that when you started Troubleshooting mode the second time, there's no output afterward, the syslog just stops.  To see what happened, we would need a screen pic, if it happens again.

 

You are running quite fast (4.121GHz).  Any chance it's overclocked too fast for the hardware?  Or overheating CPU or motherboard chipsets?

How do I reduce the CPU speed? It's not overlcocked to my knowledge, so it shouldn't be, but its quad core 4.1ghz processor

 

AMD A10-7890K Black Edition, Socket-FM2+

Prosessor,4.1GHz,Quad Core,4MB,95W, 28nm, Radeon R7, inkl. Wraith cooler

 

I don't know how to overclock, so I doubt it is, but if you can tell me how to reduce speed, perhaps I could reduce it.

 

 

I will run a new syslog without parity check.

 

 

I have no /etc folder

 

I tried installing sensor-detect to monitor my CPU temprature, but only got this far

 

Do you want to generate /etc/sysconfig/lm_sensors? (yes/NO): yes

Copy prog/init/lm_sensors.init to /etc/init.d/lm_sensors

for initialization at boot time.

You should now start the lm_sensors service to load the required

kernel modules.

 

root@SkyNAS:~# cp prog/init/lm_sensors.init /etc/init.d/lm_sensors

cp: cannot stat ‘prog/init/lm_sensors.init’: No such file or directory

root@SkyNAS:~# cp /init/lm_sensors.init /etc/init.d/lm_sensors

cp: cannot stat ‘/init/lm_sensors.init’: Not a directory

root@SkyNAS:~# /etc/init.d/lm_sensors

-bash: /etc/init.d/lm_sensors: No such file or directory

 

What am I doing wrong?

Could you please tell me how to copy that file and to where?

 

The problem happened after this, as I haven't copied the file to the right place.

But I'm guessing this wouldn't affect my system

 

Not sure where you found those instructions, but if you can remember where, I'll try to remove them!

 

You don't need lm_sensors.  You just need the System Temp plugin installed (it is), and Perl installed (it is), and then start the System Temp plugin and click the Help button and follow the instructions.  Once temps are showing, uncheck Perl for installation, you don't need it any more.

It won't Detect the drivers, because of that I tried installing this script

 

https://web.archive.org/web/20150513093034/http://dl.lm-sensors.org/lm-sensors/files/sensors-detect

 

http://lime-technology.com/forum/index.php?topic=36543.msg430371#msg430371

 

It actually is recomended in help for System Temprature

Use 'detect' to search for available drivers using the 'sensors-detect' script. After detection proceed with saving the list.

Use 'save' to store the drivers list to disk, the drivers field may be edited before saving (e.g. add a non-specified driver). Saving is mandatory.

Use 'load drivers / unload drivers' to manually load or unload the available driver modules.

 

Sorry but I forgot to mention it was the sensor-detect script, that was just the last lines of installing the script.

The sensors-detect script is already installed with unRAID.  As far as I know, this should be the correct instructions -

 

1. Install Nerd Tools (NerdPack) plugin (if not already installed)

2. Click the On switch for Perl installation

3. Click the Apply button for it to download and install Perl

4. Install System Temp plugin (if not already installed)

5. Click the Detect button (should result in one or more modules being listed)

6. Click the Save button

7. Click the dropdown menu for Processor temp and select an appropriate one

8. Click the dropdown menu for Motherboard temp and select an appropriate one

9. Click the Apply button at the bottom

10. Done.  Check for temps in the bottom status bar

11. Go back to Nerd Tools and click the Off switch for Perl and click Apply (Perl will not be installed on next boot)

 

I did notice that nowhere in the Help does it mention the need for Perl loaded when Detect is clicked.  Hopefully bonienl will rectify that?!? 

 

The only reason you might want a newer sensors-detect script is if you need a more modern one, and unfortunately they aren't making them at the moment, but I've attached the latest one I could find, zipped.  Extract it to /usr/sbin on your unRAID server.

sensors-detect.zip

The sensors-detect script is already installed with the System Temp plugin.  As far as I know, this should be the correct instructions -

 

1. Install Nerd Tools (NerdPack) plugin (if not already installed)

2. Click the On switch for Perl installation

3. Click the Apply button for it to download and install Perl

4. Install System Temp plugin (if not already installed)

5. Click the Detect button (should result in one or more modules being listed)

6. Click the Save button

7. Click the dropdown menu for Processor temp and select an appropriate one

8. Click the dropdown menu for Motherboard temp and select an appropriate one

9. Click the Apply button at the bottom

10. Done.  Check for temps in the bottom status bar

11. Go back to Nerd Tools and click the Off switch for Perl and click Apply (Perl will not be installed on next boot)

 

I did notice that nowhere in the Help does it mention the need for Perl loaded when Detect is clicked.  Hopefully bonienl will rectify that?!? 

 

The only reason you might want a sensors-detect script is if you need a more modern one, and unfortunately they aren't making them at the moment.

 

1. I have nerd package installed with following

perl-5.22.0-x86_64-1.txz up-to-date 14.0 MB yes yes

2. System Temp is already installed

3. When I press "Detect" in my browser(tried Edge, FireFox&Chrome) the window refreshes, but no drivers found - it  only prompts me back to the same page, with "Click Detect"

CPU: AMD A10-7890K Black Edition, Socket-FM2+

Mobo: ASUS A88X-Plus, Socket-FM2+

4. The drop down menus are inactive

 

I believe I have done more harm then good to the system with installing\uninstalling,

and would like to completly start over with unRAID, I still have my Plus.key file from Lime-Tech download . ( I would have done this either way )

Are there any more files I should make a secure copy of?

My disks won't get reformatted, will they? ( I have taken a ss of the "Main Page" so I know which disk goes where, for any impact on my network)

 

 

This is the picture when my CPU was overheating, so I dont think thats the problem now

 

On my screen now I keep getting this error

"grep: /etc/sensors.d/sensors.conf: No such file or directory"

You may have a non-standard motherboard then.

 

To do some cleanup, I'd mark Perl for uninstallation, then remove System Temp, perhaps the Nerdpack too, and reboot.  The grep error is just from System Temp not being setup correctly, so it can't find sensors.conf.

 

If you feel you need to clean up further, I've created a section called Starting Over.

I completely reinstalled, and instead of PhAzE's plex plugin I went for the docker one,

but the problem still persists. Same error.

Syslog and diangostics ready for download

 

Could it be that the CPU is still geting to warm?

As this only happens when I scan through libraries etc, I changed my cpu and the new one has always been problematic.

syslog.txt

skynets-diagnostics-20160616-1340.zip

New syslog

syslog-20160616-194147.txt

I finally got around to checking your syslogs and diagnostics, but only saw one problem - your drives are running very hot, 48 to 50 degrees.  That implies your whole system is running quite hot.  You weren't able to get CPU temps, but have you tried the BIOS on boot - they often have a section with temps and voltages.  At the moment, I'm leaning toward your system being too hot, and causing the issues, especially since starting a heavy I/O operation crashes it.  That would raise the temps even higher, perhaps too high.

 

How is your system cooled?

Hmm, I migrated my server to my main computer, before I buy Lian Li PC-D8000, that system is water cooled, and now my harddrives are 30-40c (open case tho) I thought degrees around 44-50 was normal, but I guess this is better.

After migrating my system I've had no problems, I am currently scanning my library and doing parity sync without problems.

 

Specs on my main computer/current server is

i7-2600k

16gb ram (Will be nice when I setup VM's)

ASUS motherboard (Hope this one can accomodate the System Temp plugin)

Corsair Hydro Series H60 CPU Cooler

120mm Radiator, 115x/2011/2011-3/1366, AM2(+)/AM3(+)/FM1/FM2, 2000 RPM, 54 CFM