Random Rebooting

[pro_trouble]

I have had my freenas box for the better part of year now and over that time the longest uptime I've had is about 43 days. It just seems to randomly reboot... and I intentionally haven't been using it... so it just sits idle all the time and randomly reboots and leaves nothing in any logs.

My version of Freenas is FreeNAS-9.2.1.2-RELEASE-x64 (002022c). I had the same problem with the previous release, except that it didn't reboot... it just froze and I needed to manually reboot it.

Yesterday I replaced the PSU, and then it ran for the day then rebooted... so it wasn't the PSU all on its own. I have previously run memtest and it produced no errors. All 5 fans (CPU and case fans) are all running and the CPU temperature is ok.

Next I will be replacing the memory then the mobo/cpu, but I would like some input on other things to check first.

I've included a couple of the reporting graphs to show the behaviour I am seeing. If anyone has any tips/opinions of things to do other things to look at I would appreciate the input.

Hardware specs are included below.

R.

---------------------
Motherboard: ASUS H61M-A/USB3 LGA 1155 Intel H61 HDMI USB 3.0 Micro ATX Intel Motherboard with UEFI

CPU: Intel Core i3-3220 Ivy Bridge Dual-Core 3.3GHz LGA 1155 55W Desktop Processor Intel HD Graphics 2500 BX80637i33220

Hard Drives: 3 x Western Digital Red NAS Hard Drive WD20EFRX 2TB IntelliPower 64MB Cache SATA 6.0Gb/s 3.5" Internal Hard Drive

RAM: Crucial Ballistix Sport 8GB 240-Pin DDR3 SDRAM DDR3 1600 (PC3 12800) Low Profile Desktop Memory Model BLS8G3D1609ES2LX0

 

[anodos]

My best guess is that it is the RAM. You can run memtest86 on it. If you're replacing the hardware, make sure you buy an actual server motherboard. Friends don't let friends use desktop components for FreeNAS.

 

[pro_trouble]

I was going for relatively inexpensive, but you have a point. I'll give some new memory a try.

 

[danb35]

You could buy a Socket 1155 server motherboard (one of the SuperMicro X9 series, for example) and keep your CPU, and still have ECC support (indeed, requirement).

 

[no_connection]

Run OCCT. If it errors you know something it wrong.
http://www.ocbase.com/index.php/software/20-occt

 

[joeschmuck]

If you have run Memtest86 for 3 days solid without a single error, your RAM is likely good. If your RAM fails, and if it's running at 1600MHz, drop the clock rate to 1333MHz and try again. Of course you could just drop your clock rate and see if that stops the rebooting issues. There are a lot of compatibility issues out there when running 1600MHz rates, even if they say they will work.

The problem with random reboots is the difficulty in isolating the problem, as you are aware. If you could figure out if they are occuring during a backup (I/O intensive), or just sitting idle, or whatever, that might help isolate it.

If you have powerd enabled, disable it to see if that fixes the issue.

 

[pro_trouble]

I'm coming back to this to provide some follow up. After a bunch of testing I believe that the problem I was having was related to the NIC (long ram test was very successful, so probably not a problem with the ram, other than it wasn't ECC, and got a new power supply). Hit it with a bunch of data "quickly" and I would get a crash (i.e. over a 100mbps connection). Try the same transfer using a computer connected via wifi (Wireless G only) and the thing would sit there and transfer all day long, albeit very slowly.

I wasn't willing to put any money into this setup anymore, so I just built a new machine with Xeon processor, 16GB ECC ram, and dual Intel NICs. It works great so far.