I'm running a few dockers and a Windows 8 VM with GPU passthru.  Whenever the VM is running, and either SABnzbd is downloading, or unRAID is doing a parity check, unRAID will just reboot itself.  Most recent syslog attached (syslog3.zip).  I have a couple of previous ones also.

 

I left unRAID to do a parity check from about 10pm to about 11am today, and it was about 75% finished.  My wife started playing some TV show from the media player in the windows VM, and after about an hour, the video just froze, and I knew; unRAID rebooted itself.  Sure enough, after a couple minutes, unRAID had restarted and restarted a parity check.  This happens pretty much daily, if not more often.

 

I've been having a lot of problems with unRAID freezing, but realized it wasn't freezing, I was just losing network connectivity.  I made some changes, including turning off one of my onboard NIC's, and removing network bonding.  I also installed a new router, with updated firmware.  The network loss is much less frequent, but not totally gone.

 

I suspected the issue might have to do with CPU pinning, so I removed all pinning from my XML, and it seemed to help (too longer to reboot), but it's still happening.

 

I have another post on the forum about this, but since it's changed to simply rebooting, I feel like it's another problem, so I'm starting a new thread.

 

I'm really, really tired of fighting the server, so I'm very motivated to get this issue resolved, so I'll pretty much do anything to help find the source of the problem, I just don't know what else to do.

 

PLEASE help me fix this.

syslog3.zip

syslog2.zip

syslog.zip

I know this is probably not the answer you're looking for, but to keep the wife and family happy I would go back to basics. I would remove the windows VM completely and stick to what unraid does best.. Storage.  If the windows VM is for video playback then replace it with a small media player.  I would then revisit the windows VM when your tinkering is not going to impact the use of the server or when vm's are closer to being a supported feature.  I just feel you are causing yourself too much frustration with something that is still in beta.

I know this is probably not the answer you're looking for, but to keep the wife and family happy I would go back to basics. I would remove the windows VM completely and stick to what unraid does best.. Storage.  If the windows VM is for video playback then replace it with a small media player.  I would then revisit the windows VM when your tinkering is not going to impact the use of the server or when vm's are closer to being a supported feature.  I just feel you are causing yourself too much frustration with something that is still in beta.

 

Not a bad idea, but not realistic either.  I'm not going to buy more hardware to rebuild the HTPC, which I neither need, nor want, and will eventually end up doing away with again, once unRAID stops fighting me.  Especially when my issues could very well be an issue with unRAID, and/or the hardware I have in the server, or the plugins, or the state of flux of the KVM/XEN system, and/or my network.

 

I think a better solution is to get a better system in place for error/failure tracking, which persists thru a hard-boot.  It doesn't need to be the default error tracking system, but considering we're pretty much all dealing my multi-terrabyte systems, I think putting a few MB onto the array to have a better error reporting/diagnosing/resolution system is a small price to pay.

 

If I could provide a syslog from right before a random reboot, and if that syslog was capturing all the pertinent information, it seems like it should be pretty easy to figure out the problem.  from there, a solution could be crafted/determined.  As it is now, I'm blind as to what's causing this, so any attempt to resolve is a guess.

 

I don't think anyone would really suggest a mechanic should just start replacing parts on a malfunctioning car, or to 'just buy a new car' when my current car acts up.  Proper, systematic diagnosis is the solution.

 

Unfortunately, I can't diagnose a syslog any better than an electrical issue in an 80's Jaguar

just out of curiosity, what sort of stability testing have you performed on your server?

I've run memtest for about a dozen hours.  I'm not sure what else I could do actually.  Suggestions?

I now suspect this has to do with touching the same file from inside and outside the VM.

 

Playing videos for 2 days has been fine, but I just tried to move a file using JRiver inside the VM, and it didn't look like it was moving, which often means a permissions issue.  So, I went to my laptop, and did the newperms command on the folder containing the file, and the command finished just fine.  Then I lost access/control from within the VM, I could not access unRAID from my laptop, nor Putty.  All within a couple/few seconds of issuing the command.

 

Then the VM went black, and Putty gave me a network error, and unRAID has rebooted itself.

 

Sadly, somehow my logging wasn't being stored to he external location for some reason.  Once unRAID comes back up, I'll try to get the logging moved, and see if I can re-create this issue.

 

I am curious though; if I can provide a log, does anyone actually care to review and see if we can fix this; or am I just out of luck, since it only seems to be happening to me

I now suspect this has to do with touching the same file from inside and outside the VM.

 

Playing videos for 2 days has been fine, but I just tried to move a file using JRiver inside the VM, and it didn't look like it was moving, which often means a permissions issue.  So, I went to my laptop, and did the newperms command on the folder containing the file, and the command finished just fine.  Then I lost access/control from within the VM, I could not access unRAID from my laptop, nor Putty.  All within a couple/few seconds of issuing the command.

 

Then the VM went black, and Putty gave me a network error, and unRAID has rebooted itself.

 

Sadly, somehow my logging wasn't being stored to he external location for some reason.  Once unRAID comes back up, I'll try to get the logging moved, and see if I can re-create this issue.

 

I am curious though; if I can provide a log, does anyone actually care to review and see if we can fix this; or am I just out of luck, since it only seems to be happening to me

 

Justin, you have some odd issues, but we want to help you.  We are laser-focused right now on finalizing the dynamix integration and preparing content for guides and videos to help better explain things.  Tom is about to get back into bug-squashing mode again, and at the minimum, I'd like to get his input on this.  I have your same motherboard, and I do not have these issues.  I've shared with you my settings, you matched them, and you have problems still.  You have a different CPU and your GPU is known to be a little flakey from our testing with the same model device.

 

Let me ask this though:  if you stop using a VM / GPU Pass through and just use docker + unRAID OS, everything is fine, right?  Or do you have issue independent of your VM usage?

Thanks for the feedback jon.  I suspect my motherboard is crapping out on me...

 

http://lime-technology.com/forum/index.php?topic=36372.0

 

I ordered a new one

 

It will probably be after thanksgiving before I can install it.  I'm currently clearing everything off of a drive, so I can remove it from the array, so I have SATA port to use for my SDD.

 

As for whether or not I have issues without the VM, it's hard to say.  I think it's crashed without it running, but it crashes so often, it's hard to keep track