Unstable Desktop

Forums › Desktop computing › Unstable Desktop - Ideas?

Xeon

302 posts

Ultimate Geek

#196290 26-May-2016 00:20

Hey,

I built myself a desktop to use a NAS/Plex Server and its been a little unstable.

System is:

PSU: Corsair HX750i

RAM: 2x 8GB Crucial Ballistic Sport LT DDR4

Motherboard: Gigabyte Micro ATX DDR4 GA-H170M

CPU: i5-6400

Boot Drive: SanDisk SSD Plus 480GB

Storage Drives: 5x 3TB WD Blues currently installed + 6 more to put in

PCI-e Cards: Intel PRO/1000 Dual Port NIC & Dell Perc H200 SAS HBA Card

I initially installed Windows 10 and ran that for ~1 hour while I was testing some things and had no issues. Following that I installed OpenMediaVault (Debian Wheezy) and I never really had the system crash but I would randomly get segfaults - in particular anytime there is high disk I/O/load in general it seems to cause a segfault (eg. Plex Server transcoding, high number of concurrent torrents, moving files from SSD to HDD (when I use a program to move them), SnapRAID Sync) and would crash the program involved and potentially some of my services.

I put it down to it being an issue potentially with the old libraries in Debian and planned to install a new OS, in the meantime (today) I need to install my other 6 hard drives and my Dell Perc H200 so I can plug them in. Before anything I run a memory test with MemTest86 (all drives unplugged at this point) but the memory test simply will not complete, 1 stick/2 sticks; different slots. I never get a memory error but it will randomly just reboot under EFI memtest. Under a legacy booted MemTest it would often stopped saying 'memtest halting unexpected interrupt cpu 1', I did catch it under EFI booted once for half a second before a reboot say something about an invalid address. I decided to go onwards, and I need to flash my Dell Perc H200 to a different firmware by booting FreeDos and running a couple of programs but just sitting on the FreeDos shell the computer just randomly reboots (aka the same issue as for MemTest).

I'm thinking its the motherboard giving me grief but not entirely sure. Any ideas on things to try/whats up?

Cheers.

ubergeeknz

3344 posts

Uber Geek

Trusted

Vocus

#1559755 26-May-2016 00:27

Could be almost anything, probably not the RAM (unless it's both sticks OR it's the wrong speed/type for your setup)

Make sure your heatsink is properly installed and interfaced with the CPU

Then you could try swapping power supply, mainboard out

Very unlikely to be a faulty CPU.

gzt

17111 posts

Uber Geek

Lifetime subscriber

#1559808 26-May-2016 09:18

No idea. Maybe check the memory is running at the expected timings / spec / clock.

timmmay

20578 posts

Uber Geek

Trusted

Lifetime subscriber

#1559822 26-May-2016 09:48

Memory fault sounds possible, fairly common issue. Memtest x86 isn't that good, try HCI memtest, it found a memory fault that memtest missed for me. Note that HCI only tests 2GB or RAM per instance so you have to run [ (RAM amount - 1GB) / 2 ] instances, and they need to run for up to 24 hours to find faults. Mine only found the fault on the 3rd pass. Replacing RAM fixed all problems immediately.

ubergeeknz

3344 posts

Uber Geek

Trusted

Vocus

#1559949 26-May-2016 11:15

timmmay:

Memory fault sounds possible, fairly common issue. Memtest x86 isn't that good, try HCI memtest, it found a memory fault that memtest missed for me. Note that HCI only tests 2GB or RAM per instance so you have to run [ (RAM amount - 1GB) / 2 ] instances, and they need to run for up to 24 hours to find faults. Mine only found the fault on the 3rd pass. Replacing RAM fixed all problems immediately.

But a memory fault is unlikely to reboot the machine when it's just sitting in DOS.

I would first suspect a heatsink issue, if this were me. What temps are you getting in the BIOS?

timmmay

20578 posts

Uber Geek

Trusted

Lifetime subscriber

#1559953 26-May-2016 11:19

Heat is worth checking. Memory is pretty easy to test, and is a really common fault, so no harm testing it.

ubergeeknz

3344 posts

Uber Geek

Trusted

Vocus

#1559961 26-May-2016 11:34

timmmay:

Heat is worth checking. Memory is pretty easy to test, and is a really common fault, so no harm testing it.

According to the OP the box cannot get through a memory test ... and has tried both sticks in isolation.

Xeon

302 posts

Ultimate Geek

#1560264 26-May-2016 21:11

Timings on the memory are at the ones specified by Crucial.

Temp shouldn't be issue - idle is 18 degrees Celsius, load (from memory) was reasonable too (I ran Prime95 under Windows when first installed and no issues).

RAM I have tested under linux and it doesn't report any issues.

PSU I can try swap, but its doubtful (voltages look good on it, plus is stable once in linux).

Running linux I never get a random reboot but just browsing the BIOS I managed to get it reboot on me just before. I also have had it freeze a few times on the boot screen (screen with press f* for x,y,z).

I'm leaning towards some motherboard/ram issue or incompatibility

GoodSync

GoodSync. Easily back up and sync your files with GoodSync. Simple and secure file backup and synchronisation software will ensure that your files are never lost (affiliate link).

ubergeeknz

3344 posts

Uber Geek

Trusted

Vocus

#1560312 26-May-2016 22:04

Sounds like you are on the right track ... just keep eliminating things until you find the culprit. Let us know though cos it's sure an interesting one...

lNomNoml

1807 posts

Uber Geek

ID Verified

#1560323 26-May-2016 22:24

ubergeeknz:

timmmay:

Memory fault sounds possible, fairly common issue. Memtest x86 isn't that good, try HCI memtest, it found a memory fault that memtest missed for me. Note that HCI only tests 2GB or RAM per instance so you have to run [ (RAM amount - 1GB) / 2 ] instances, and they need to run for up to 24 hours to find faults. Mine only found the fault on the 3rd pass. Replacing RAM fixed all problems immediately.

But a memory fault is unlikely to reboot the machine when it's just sitting in DOS.

I would first suspect a heatsink issue, if this were me. What temps are you getting in the BIOS?

Faulty RAM can do anything.

gzt

17111 posts

Uber Geek

Lifetime subscriber

#1560326 26-May-2016 22:37

I suggest removing the nic if you have not already, check for firmware updates and then set to factory.

It is not unknown for northbirdge issues to give similar symptoms but in any case submit a support ticket to gb and see what they suggest as a next step:

http://www.gigabyte.com/support-downloads/technical-support.aspx

Xeon

302 posts

Ultimate Geek

#1648522 10-Oct-2016 13:45

In case anyone finds this thread, my CPU was broken.

Lias

5589 posts

Uber Geek

ID Verified

Trusted

Lifetime subscriber

#1648541 10-Oct-2016 14:10

That is an exceptionally rare fault. I've worked in IT for 20 years and I've only ever seen maybe a dozen non overclocked Intel CPU's fail.

I'm a geek, a gamer, a dad, a Quic user, and an IT Professional. I have a full rack home lab, size 15 feet, an epic beard and Asperger's. I'm a bit of a Cypherpunk, who believes information wants to be free and the Net interprets censorship as damage and routes around it. If you use my Quic signup you can also use the code R570394EKGIZ8 for free setup.