Motherboard dying? - Advice needed - Unraid 4.7

[Queball]

My server was unresponsive today. I tried to connect via unmenu, the stock web interface, telnet, and at the console and could get no response. When trying to connect via telnet I could enter user and password but no other commands would work after logging in. When I was at the console the screen was blank and no response from the keyboard commands (powerdown, restart, control alt del, etc.) it appeared that the usb keyboard was not working (no lights, even when toggling caps lock, num lock, scroll lock). Against my better judgement I decided to press the restart button as I could see that none of the drives where spun up. The server restarted normally and the keyboard at the console was working however a parity check did not automatically start which I though was going to happen because of the ungraceful restart. I started a parity check and decided to look on the flash share for a saved sys log. I could'nt access the flash share but I could access all other shares. I took a look at the current sys log and noticed some errors concerning high speed usb initialization failures and restart attempts (should have grabed a screen shot). I started to suspect there was a problem with the usb ports so I decided to cancel the parity check. The parity check successfully stopped but the server became unresponsive again. I checked the console and the keyboard would not work in any usb ports. I decide to shutdown and test different keyboards and usb ports by entering the bios and navigating around. Before doing this I pulled the unraid flash drive and disconnected the hard drives. I could'nt get any of the keyboard/usb port combos to work.

 

So I think there may be a problem with my motherboard. Does anybody have any other troubleshooting advice?

 

Here is my setup:

Unraid 4.7

MB - Biostar A760G M2+.

1 x SATA2 Serial ATA II PCI-Express RAID Controller Card (Silicon Image SIL3132)

1 x Supermicro AOC-SASLP-MV8 8-Port SAS/SATA Add-on Card

3 x Norco SS-500

Azza Helios 910

Corsair TX650

 

I run the following add-ons:

SabNZBD

Couchpotato

Sickbeard

Unmenu

Crashplan

 

Thanks in advance for any advice.

[JonathanM]

Hopefully it's just your USB key that's having issues, not the motherboard. Try to make a full backup of your USB key with another PC, and run a disk check on it.

[Queball]

Hopefully it's just your USB key that's having issues, not the motherboard. Try to make a full backup of your USB key with another PC, and run a disk check on it.

 

Thanks, I hope so too. I'll try this tonight.

[dgaschk]

Boot with the bare MB.

[Queball]

Hopefully it's just your USB key that's having issues, not the motherboard. Try to make a full backup of your USB key with another PC, and run a disk check on it.

 

I copied the flash then ran check disk - no errors reported.

 

I booted the server and again and it came up stable just long enough for me to grab a syslog (see attached). Then the monitor at the console went blank and I could not longer access the flash share. Telnet works, my other shares are accessible and my add-ons seem to be running fine except for unmenu which is sluggish and most of the pages are not accessible. For example, when I go to the syslog page in unmenu all I get is "e </BODY></HTML> 0".

 

While it is still on (for how long who knows) is there any troubleshooting anyone can think for me to do?

 

Thanks

syslog-2013-11-21.txt

[JonathanM]

Before you start doing major hardware replacement troubleshooting, try preparing another usb key with the free version of unraid, just don't assign any of your current array drives. You can install and run unmenu on the new key without any array drives, and that should be a good test of whether it's the usb on the board, or your key that is causing the issues. When my 1st key died, I saw similar behaviour, where it would boot ok, but after a period of time, the usb drive basically just disappeared from the system.

 

Unfortunately, the blank monitor symptom makes me think it may be the motherboard, CPU, or RAM, but who knows, you still could get lucky.

[Queball]

Before you start doing major hardware replacement troubleshooting, try preparing another usb key with the free version of unraid, just don't assign any of your current array drives. You can install and run unmenu on the new key without any array drives, and that should be a good test of whether it's the usb on the board, or your key that is causing the issues. When my 1st key died, I saw similar behaviour, where it would boot ok, but after a period of time, the usb drive basically just disappeared from the system.

 

Unfortunately, the blank monitor symptom makes me think it may be the motherboard, CPU, or RAM, but who knows, you still could get lucky.

 

I setup a new key and pulled all the array drives and it is exhibiting the same symptoms. Seems like something is faulty with the hardware. I am running memtest right now to check the memory.

 

Does anyone know of some other ways to test the hardware?

 

Thanks

[garycase]

Since you're getting the same symptoms on a new flash drive with UnRAID Basic, you've pretty well eliminated the flash drive as the problem.

 

It's interesting that your system will run MemTest => this means both the memory and the CPU are functioning.  But clearly there's still an issue with the board -- I'd suspect the SATA controller or possibly the onboard NIC.  But neither is worth spending $$ on to replace via add-on boards on this old board => I'd simply upgrade to a nice new motherboard/CPU combo.

 

(and while you're at it, upgrade to v5)

 

[Queball]

Since you're getting the same symptoms on a new flash drive with UnRAID Basic, you've pretty well eliminated the flash drive as the problem.

 

It's interesting that your system will run MemTest => this means both the memory and the CPU are functioning.  But clearly there's still an issue with the board -- I'd suspect the SATA controller or possibly the onboard NIC.  But neither is worth spending $$ on to replace via add-on boards on this old board => I'd simply upgrade to a nice new motherboard/CPU combo.

 

(and while you're at it, upgrade to v5)

 

Yeah that's what I'm thinking. I put off upgrading to 5.0 because my server was doing everything I wanted it to on 4.7. So I guess its an early Christmas this year!

 

I'll start browsing around the forum for 15 drive 5.0 hardware.

 

 

[dgaschk]

Before you start doing major hardware replacement troubleshooting, try preparing another usb key with the free version of unraid, just don't assign any of your current array drives. You can install and run unmenu on the new key without any array drives, and that should be a good test of whether it's the usb on the board, or your key that is causing the issues. When my 1st key died, I saw similar behaviour, where it would boot ok, but after a period of time, the usb drive basically just disappeared from the system.

 

Unfortunately, the blank monitor symptom makes me think it may be the motherboard, CPU, or RAM, but who knows, you still could get lucky.

 

I setup a new key and pulled all the array drives and it is exhibiting the same symptoms. Seems like something is faulty with the hardware. I am running memtest right now to check the memory.

 

Does anyone know of some other ways to test the hardware?

 

Thanks

 

Boot with a bare MB.

[garycase]

I haven't used this board, but it looks like a good choice for what you want:

http://www.newegg.com/Product/Product.aspx?Item=N82E16813157369

 

It's an ATX board (larger than your current uATX) ... but your case holds ATX boards, so that shouldn't be a problem.  With 8 SATA ports, you'd only need to add one 8-port card to have 16 drive capability => and the board has 3 PCIe x16 slots, so you could easily expand beyond that if you ever wanted to.

 

In addition, it's a Haswell board -- the Haswell chipsets and CPUs are VERY power-efficient.  The system would undoubtedly draw far less power than you're using now.

 

[garycase]

... if you want more choices, here's a list of all the Socket 1150 (Haswell) boards Newegg stocks with 8 SATA ports:

 

http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&N=100007627%20600438202%20600054097&IsNodeId=1&name=8%20x%20SATA%206Gb%2fs

 

Edit:  There's one more excellent choice that's not listed above (It's under their Server motherboards, so didn't show up in the search I did):

 

http://www.newegg.com/Product/Product.aspx?Item=N82E16813182836

 

[Queball]

What about this CPU:

http://www.newegg.com/Product/Product.aspx?Item=N82E16819116947

 

and this MB+Mem:

http://www.newegg.com/Product/Product.aspx?Item=N82E16813128592

 

I'll think about the Supermicro. Do you know if it is picky about which memory sticks you can use?

[garycase]

What about this CPU:

http://www.newegg.com/Product/Product.aspx?Item=N82E16819116947

 

and this MB+Mem:

http://www.newegg.com/Product/Product.aspx?Item=N82E16813128592

 

I'll think about the Supermicro. Do you know if it is picky about which memory sticks you can use?

 

That motherboard/CPU combo is a good choice [and free RAM is a nice bonus ].  You can save $10 and get more performance by using the 4130 instead of the 4130T => the lower TDP throttled CPU isn't really necessary ... the vast majority of the time you'll be drawing very little power from either.  The T version simply throttles ("governs") the max power the CPU can draw -- but all of the Haswell CPU's are very power-efficient.

 

The Supermicro boards I've used in recent builds have had no issues with any quality DDR3 modules I've used (Crucial, Kingston, Corsair, etc.).

 

[Queball]

What about this Pentium: http://www.newegg.com/Product/Product.aspx?Item=N82E16819116950 with multiple addons (see original post). Is the I3 overkill or worth the extra cost.

 

Also I am checking that a ATX MB will fit in my Azza Helios 910 with 3 Norco 500s (they protrude pretty far into the case). Looks like it will though.

 

 

[garycase]

The Pentium will work fine, although personally I'd go with the i3 for the extra CPU headroom it provides.  I've never seen anyone complain about having too much CPU "horsepower"