Random reboots

[dv310p3r]

I've jumped through the forum enough now that I see in most similar cases the answer is that it's a hardware issue. So I've prepared as much logging as I can. I'll add that prior to turning this into my unRAID server it was running Windows 10, with no issues, pretty much ever.

 

I'll go through my setup first and outline as much detail as I can, then I'll give the best description of the problems I'm having and when they have happened, and what I've tried to do to troubleshoot the issues.

 

My Setup:

 

• HARDWARE:

  • MSI ATX DDR3 2400 LGA 1150 Z97 PC Mate

  • Core i7 4770 (not K)

  • Cooler master hyper 212 EVO - CPU Cooler with 120mm PWM Fan

  • 32gb Crucial Ballistix Sport (8gb x 4) DDR3 1600 MT/s PC3-12800 memory

  • EVGA SuperNOVA 750 B1 80+ Bronze, 750W Power Supply

  • MSI GTX 1070 GPU

  • 10 Drives

    • Cache Pool (BRTFS)

      • 2 500GB Samsung Evo 850 SSD

    • Parity

      • 1 3 TB WDC_WD30EZRX

    • Array

      • 2 3TB Toshiba_DT01ACA300

      • 1 3TB ST3000DM008

      • 2 1TB WDC_WD10EADS

    • Unassgined

      • 1 250gb Sandisk SSD

      • 1 1TB WD Green

• SOFTWARE:

  • Dockers:

    • Plex

    • Sabn

    • Couchpotato

    • Sonarr

    • Splunk

    • Deluge

  • VM

    • Windows 10

      • I have both the unassigned drives being used by the windows 10 VM where I pass the whole disk to the VM

        • the SSD is for the OS and the 1TB is for games / data

      • Used for gaming, I have my tower on my desk

      • I pass the 1070 through to the windows 10 VM for gaming

 

My Issues / Occurences:

 

• Sometimes if / when I'm gaming on my VM the system will lock up, and ten 15 seconds later, restart
• Sometimes it's 1 in the AM and I'm watching a stream from Plex on my Firestick and the system will restart (no gaming going on)
• Sometimes its 3 in the Afternoon, no gaming, no streaming, and the system will restart (no gaming, no streaming)
• I do / have been getting warnings about Disk's getting warm but they always go back to normal

 

Stuff I've Done To Troubleshoot:

• I have pressed the system while watching the temps and haven't noticed any temps getting out of control
  • not able to reproduce
• I've replayed video streams that were running when a reboot happened to see if there was a problem with the file
  • note able to reproduce
• I've done plenty of logging and searching through logs, but nothing evident really pops out
  • Installed Splunk to try and catch the last few log events before a shutdown (I think I have it, but not sure)
  • the SMART logs show a couple of errors, but I'll note that I have just recently, (like two weeks ago), moved the drives and the flash out of a server (R720) into this case and setup. I never had the reboot problems in that server, so that leads me to the conclusion that the drive errors reported in the SMART logs can't be what's causing these problems (or very unlikely).

 

Lastly I've attached my logs:

 

Hope you can help...

 

ceres-station-diagnostics-20170809-2322.zip

Edited August 10, 2017 by dv310p3r
Added some additional troubleshooting steps I took

[Squid]

Have you checked for BIOS updates?  Will it restart even if the VM isn't running?

[Frank1940]

I am wondering how VM's handle the auto-update feature of Win10?  Most of the time, windows requires a reboot of the system install the updates.  Could this be part of the problem? Does your VM run all of the time?  Was it always running when the reboot occurred?

 

You mention " prior to turning this into my unRAID server it was running Windows 10 ".  What exactly did you do?  Was it strictly installing unRAID as the OS and then setting up a Win10 VM or was there some hardware changes also in the conversion?

 

You mention temperatures being on the high side.  Temperatures of what devices and how high? 

 

That power supply (apparently) has (4) +12V busses of 20 amperes each.  How are you spreading the load for the hard drives and the video card?

 

EDIT:    Do you have a UPS on this system?

Edited August 10, 2017 by Frank1940

[dv310p3r]

On the latest BIOS, and yes, it'll restart even if the VM isn't running.

 

I'm was tailing the syslog at just now trying and make it crash again and I did, but no data in the syslog at all. Rather no data that is from after the VM started up. It ran for like 5 minutes before restarting.

 

Id' love to know if it was temps in some way, but I can't get the temp data over time.

[Squid]

The diagnostics you posted (you had FCP running in trouble shooting mode) shows the temps every 10 minutes.  And the last entry (Aug 9@23:12) is a reasonable 44C

 

Primary causes of random reboots are

• Power Mains Problem (No UPS and power dips)  (Air Conditioner cycling on causing a power dip?)
• Bad / Overloaded / Poor Quality Power Supply
• Memory Bad (Run Memtest)

[Frank1940]

3 hours ago, Squid said:

Memory Bad (Run Memtest)

 

Yes and let it run for 24 hours.  Having 4 memory chips in some MB's can sometimes be a problem.  Double check and see if the MB manufacturer has the chips you  are using on their list of 'preferred/tested/certified' memory. 

[dv310p3r]

14 hours ago, Squid said:

Primary causes of random reboots are

• Power Mains Problem (No UPS and power dips)  (Air Conditioner cycling on causing a power dip?)
• Bad / Overloaded / Poor Quality Power Supply
• Memory Bad (Run Memtest)

 

I'm on a UPS so I'm going to run Memtest next. Let me elaborate a bit on where I'm at.

 

So, in my possession I have an i5 4690K which is the same socket type as the i7. I have a somewhat convoluted amount of hardware sitting around my house. Anyway, I decided last night to pop the i5 back into the system and give my son the i7. So far, no reboots. There is some premise for this, which is I had tried the unRAID setup last year on the i7 and had similar issues. The problem is that when running the i5 the VM is noticeably, laggy, and not just in gaming, it's laggy even moving the mouse around on the desktop. On the i7 it's not laggy at all, it's like a regular desktop. So it sucks to have to go back to the i5.

 

Anyone have an idea why the i7 would cause that kind of a problem?

 

My last resort honestly is going to be to by a Socket LGA2011 board (not sure if I should go Dual or Single) and scavenge one or both E5-2670 v2 chips I have in a Dell R720 server. I didn't want to go this route because I wanted to sell the R720 for as much as possible. However I'm guessing I can still sell the R720 barebones. I had been using the R720 just as the unRAID server, but I had to consolidate a bit because of the wife, also, the R720 chassis that I have only fits a max of 8 drives. The tower that I have (the Corsair 750D) can hold at least 12, and possibly more if I try. Also, i really like how cool it looks on my desk.

 

My one and only concern about the E5-2670 v2 route is gaming performance. I'm really concerned about the hit I'll take when gaming. For the record, I'm not a 120hz gamer, rather I'd like to be, but am ok not being. I have a 4K Samsung panel that does max 60hz, so anything more than that is pointless. I guess I'm asking for opinions / experience on this topic. Will I suffer majorly in my gaming, or will that processor along with my 1070 be more than fine?

 

Thanks again,

[Frank1940]

23 minutes ago, dv310p3r said:

So, in my possession I have an i5 4690K which is the same socket type as the i7. I have a somewhat convoluted amount of hardware sitting around my house. Anyway, I decided last night to pop the i5 back into the system and give my son the i7. So far, no reboots. There is some premise for this, which is I had tried the unRAID setup last year on the i7 and had similar issues. The problem is that when running the i5 the VM is noticeably, laggy, and not just in gaming, it's laggy even moving the mouse around on the desktop. On the i7 it's not laggy at all, it's like a regular desktop. So it sucks to have to go back to the i5.

 

 

Assuming that the i7 is not flaky, this could be a PS issue (since nothing else is changing) .  What is the difference in TDP of the two processors? 

[dv310p3r]

i5 4690 88w TDP

i7 4770 85w TDP

[dv310p3r]

Figured I'd follow up. Got home today and got to work trying to solve the problem of the Windows 10 VM being very laggy. So, I tried spinning up a new VM same settings, same Disk, etc... but still, super laggy. I tried that because I felt like maybe there was some setting that was stuck in thinking that it had all the cores of the i7. 

 

Anyways, that thought train got me thinking that this had to be a processor issue. So I did some more digging and low and behold, somewhere in the dark recesses of this very forum, was another post about a guy with an i5 having lag issues in his Windows 10 VM. Apparently, there's a setting that I can use to isolate cores, or something like that. So i went into my syslinux.cfg file and added "isolcpus=2,3,4" then restarted.

 

When i got back up, I went into the unRAID GUI, started the VM, turned on my Gaming PC monitor and BAAM! It works like a peach. Honestly, I've been streaming a 1080p video while playing DOOM 2016 and alls well, no crashes at all. 

 

Not sure what the deal was with the i7, it works just fine when not in the unRAID box. Anyhow, thanks for all the help.