Hello,

 

I have the severe problem, that after some hours of copying data, the sever shuts down.

 

This means, no webinterface, no samba shares and also no response when trying telnet.

 

Nothing happens when trying to telnet in- no response, Putty does not display anything.

 

I already tried to add the suggested lines in the first post of the other "Unraid webconsole unreachable ? Transport endpoint not connected ? Read this."

thread to my GO script, but nothing changed.

 

If I restart the server, everything works again.

Also, I don't have the parity up yet, as i just started to copy files over. I have to start the array manually after booting up everytime, if this helps to identify the problem.

Config is 3x3TB HDD 5700RPM and unRAID runs off a CF/IDE Adapter Plus reads config from Flashdrive. (cant boot from USB) - its the basic version, no addons. RC12A.

 

Is this a known issue? Any suggestions?

I am quite new to linux.

Post a syslog and while your waiting if you have any add ons disable them and reboot to see if the problem goes away.

Sorry misses the no add ons installed.

I can post a syslog before the failure- but when it goes offline I cant even contact it by Telnet, so how would I go about posting a syslog containing the important bit?

Hello Fatpig,

 

1. Restart your server, or telnet to your server while it is running properly.

2. Type the command:  tail -f /var/log/syslog

3. This will keep a running log of the syslog, when your unraid server freezes the telnet session will retain part of the syslog, then it's as simple as copying and pasting here in the forums.

 

Note: If you are using putty or some other telnet client, make sure it is set to not automatically close the window on session disconnect.

 

Hope that helps.

I have a feeling that you are running your unraid as a headless system like we all do.

 

But while you are having this problem if you could connect a monitor and keyboard and use that to access your system that will be extremely useful.  That way if telnet goes off line you still have control from the console of the unraid system.

 

We don't have any information about your system.  What system board, network cards, processor, memory PSU number of disks etc so we can begin to get the whole picture.  This info would be very helpful.

 

Remember that rebooting your unraid, you could loose your syslog, that is why its important to have a monitor and keyboard connected to the unraid server so that you can copy it to the flash drive before reboot.

 

If the console is locked up and not responding, then you will be looking at a hardware failure with the motherboard, memory or processor.

 

In which case you need to run the memtest 86+ for at least 2 days (should give you 10 complete loops of memory testing).  If you get no errors after 10 complete loops of memory testing then we can move on to other problems.  You could also have a cooling issue with your processor.  The motherboard is just protecting the processor from overheating by locking up the processor.

 

Can you ping the server when this happens?

 

Good luck!

 

-- Harold

Hey guys, thanks a bunch!

 

I did not even think about a hardware error or connecting via keyboard and monitor-

i will run it now until the error happens and then try to catch the syslog via all the options.

Okay guys, connecting a monitor when frozen was all that was necessary.

the computer remained frozen but displayed:

 

[Hardware Error]: CPU 0: Machine Check Exception: 4 Bank 4: b200000000070f0f

[Hardware Error]: TSC 2fdaeb7ef3bf

[Hardware Error]: PROCESSOR 2:20fc2 TIME 1365388551 SOCKET 0 APIC 0 microcode 0

[Hardware Error]: Run the above through 'mcelog --ascii'

[Hardware Error]: Machine check: Processor context corrupt

Kernel panic - not syncing: Fatal machine check on current CPU

 

 

I am using an old Sempron 2800+ that ran fine the last time i checked on windows. It is run with a cooler that may be too weak. Temps are around 50 deg celsius though when running for about an hour or so.

Do you think this could be overheating related? I have 2 case fans installed.

Mainboard is an old K8NF4G-SATA2 from ebay. The CPU was mine since it was new.

Maybe I can get a "sensors" output automatically every few seconds to check? //EDIT: Now running a sensors output every 5 secs.

Thanks you guys! I really appreciate it.

First ensure you CPU fan rpm is where it should be. If it is then I would start by removing the CPU and cooler, clean both and reapply thermal paste. While it's apart you can inspect and clean the fan blades. Reassemble and see if that fixes you issue. Here is a wiki page that may help. http://en.wikipedia.org/wiki/Machine_Check_Exception

I would agree with burt's procedure, and agree that the diagnosis is likely a hot CPU if it's happening consistently after a certain period of operation. 

 

The problem may be that you've already done the damage, and that even cooling the CPU may not 100% work, as it's already been 'cooked' so to speak.  CPU's that have been frequently run too warmly are often already subtly damaged in a way that can't be fully addressed.

I would add that although 50C is a little warm it is well within its operating range of 61 high to 95 critical. What is your motherboard temp? If you do get it running again monitor your temps with a load and see where they top out.

thanks guys, I have a spare processor so it would not be such a problem.

 

I have put the NAS in the middle of the room now and it has not crashed since then- I will keep it running now.

(much cooler than in the badly ventilated area it was before - under the cupboard.)

 

If it doesn't crash for a while we can count this case as solved.

 

I am also not sure if "sensors" really does put out the cpu temp-

as the only temp reading it puts out is "pci adapter core0"

 

- in bios i can read the cpu temp though.

I'd take a can of compressed air and blast your CPU cooler before taking things apart, especially if you can't see in between the blades to see how dusty it may or may not be.  Simple things first, then onto more involved things.

 

Always good to blast dust with the PC outside!

Actually I just cleaned the whole NAS (just put it together) - especially the cpu cooler i even cleaned the non-moving parts with an ultrasonic cleaner. amazing thing

 

Well, I think it can be called solved now.

It might all be the fault of my weak cpu cooler, that was not made for this socket - and the poor placement under a cupboard. :-P

When I put the pc out in the open it runs fine for almost a day now.

 

Thanks to all of you, I think it would have taken a great deal longer to figure this one out, as ridiculous as it may sound.

Have a great day

It might all be the fault of my weak cpu cooler, that was not made for this socket

 

That sounds very possible!