Lock ups. No local shares, no webgui, frozen screensaver.

[Beans]

This just started taking place today :( Go look at the machine, the BSD devil screensaver isnt moving, NIC lights flashing. No keyboard response.... fans all spinning.

Machine isn't nothing spectacular, but it was working fine before this, hard reset fixes it for 30 mins to couple hours.

This is my third freenas. I'm running version 8, 32bit

Intel BOXD915PDTL (using on board Sata controller)
Celeron 2.66
2gb of ram
HP NC380T
2 x 1tb in softraid. UFS
booted off a 2gb usb stick

Was using a generic pci video card, pulled it thinking maybe it was the issue. No change.

AWFULLY stumped.

 

[cyberjock]

Have you tried a RAM test? A SMART test of the hard drives?

Those are the two I'd probably try first because they are simple to run.

 

[William Grzybowski]

It probably paniced, disable screensaver and wait it happen again to get the panic message.

 

[Beans]

Well, it successfully has been up for 18 hours now without locking up.... Didn't really change anything. turned off the screensaver before you said anything William, maybe that did it.


I'll run some smart tests, and a ramtest if it comes to it.


Also where is the system log in this version? I can't find it. 8.1

 

[Beans]

Still going strong. Hopefully it resolved itself.

 

[cyberjock]

I'd definitely do the RAM and hard drive test soon. One person about 2 weeks ago lost all of his data when a bad stick of RAM corrupted his zpool beyond repair. He had no backup :(

 

[Beans]

I'd definitely do the RAM and hard drive test soon. One person about 2 weeks ago lost all of his data when a bad stick of RAM corrupted his zpool beyond repair. He had no backup :(

I have a back up on my old nas, 90% of it anyway. :D

Running memtest now, scheduled smart test... How do i know when it is done? Will i see it in the webgui? Or should i test with 3rd party software?

 

[cyberjock]

The Long SMART test is the one you want to run.

You can monitor it by doing smartctl -a /dev/ada0 and watch it under the SMART self-test logs. The remaining column will count down from 100% to 0%.

 

[Beans]

The Long SMART test is the one you want to run.

You can monitor it by doing smartctl -a /dev/ada0 and watch it under the SMART self-test logs. The remaining column will count down from 100% to 0%.

Thank you.

Ran memtest all day while at school, saw one clean pass before i left. Came back to a completely turned off! I turned it back on and...

Well there's your problem! Going to go pull the heatsink and cpu and see if there is a bad mate or if i just need a better heatsink/fan all together.

 

[Beans]

Pulled the cpu, it appears the dipshit i bought it from (new) off ebay (i bought it in a bundle) decided thermal paste on the pins is a good idea. I'm not even sure how he managed it, it's a stock fan for fucks sake, with the correct amount already applied.

 

[cyberjock]

That would definitely do it! I've seen that kind of workmanship before. Just get the highest % rubbing alcohol you can(so it evaporates the easiest) and put some in a bowl and stick the CPU in there for 10 mins. The stuff will rinse itself off. Let the CPU dry and you are back in business.

I would do a scrub after that is fixed just to make sure you have no errors and if there are they are fixed. :)