I've reached an impasse trying to figure out what's going on with unRAID on my system.  It runs for a while, sometimes a couple days, sometimes a few minutes, then crashes.  The only thing I have is the dump on the screen:

 

http://my.jetscreenshot.com/12412/20150503-gle0-224kb

 

When the server crashes, the network crashes with it.  No devices on the network can access the internet.  When i run a network trace I see all the devices trying to figure out who's who.  I do know that the unRAID server was the network master.  Here's a snapshot of the LAN traffic using Wireshark.  As soon as I unplug the server from the network, it recovers and traffic moves without issues.

 

http://my.jetscreenshot.com/12412/20150503-fbih-451kb

 

Any wisdom?

Have you run memtest yet? (Ideally for 24 hours?)

 

What kind of power supply?

99,9% sure its not hardware. I've run memtest for 36 hours, no problems.  I've even installed Windows 8 on the system and run that with no issues whatsoever.  I've got a 700W Thermaltake power supply.  Here's whats in the system now:

 

PCIe 16x Slot - Syba PCIe x4 2 Port SSD Caddy (http://sybausa.com/productInfo.php?iid=1389)

PCIe 16x Slot - Supermicro AOC-SAS2LPMV8 8 Port HBA

Fresh build (all repurposed equipment)

Asus P8Z68-V Pro i5 2500K, 16GB Memory

 

Drives:

Parity: 5TB Seagate (On Mobo)

Cache: 750GB Samsung EVO (In Syba)

Apps: 128GB Samsung Pro (In Syba) (Pulling this drive shortly since I learned should just put apps on cache drive)

Data1: 2TB Hitachi (on Mobo)

Data2: 2TB Hitachi (on Mobo)

Data3: 2TB Hitachi (on AOC)

Data4: 2TB Seagate (AOC)

Data5: 2TB Hitachi (AOC)

Data6: 2TB Hitachi (AOC)

Data7: 2TB Hitachi (AOC)

Data8: 3TB Hitachi (AOC)

Data9: 3TB Hitachi (AOC)

Data10: 3TB Hitachi (on Mobo)

Data11: 4TB Seagate (Mobo)

 

I had this problem happen a couple weeks ago, http://lime-technology.com/forum/index.php?topic=39435.msg367919#msg367919

 

No responses, so this must be difficult territory...I'm about to give up and go with WHS (*sigh*)...

It would still be useful to know the exact power supply. Just knowing it is a reputable manufacturer and has a big number of watts is not the whole story when trying to run a lot of drives. Single 12v rail is recommended for this use. Some PS divide the watts into different uses, diverting part of the power to graphics cards, for example.

Do you have this problem too without any add-ons active (reboot your system in safe mode) ?

 

If you believe it isn't hardware related, it might be worthwhile to do a process of elimination, remove plugins, remove dockers, and re-add them one-by-one. This might be a lengthy investigation though.

 

My impression has long been that it's a defective network card, especially since it floods the network, creating a denial-of-service to the entire network.  But your testing seemed quite good, comprehensive, so most of us haven't been able to come up with any additional advice.

 

The cause of the issue HAS to include the fact that it floods the network, and that really limits the possibilities.

I would agree that it seems like a defective network card - except...

 

1. The same behavior (crashes) appear with either the motherboard (Intel 82579) adapter AND the Intel Pro/1000 card I bought.  Both behave the same way.

2. The fact that both work fine with Windows 8 - and most importantly UR5.0.6 works without errors, seems to point me to a driver issue? 

3. Both the NIC's use the e1000e drivers and both behave the same way.  And...given the likelihood that BOTH NIC's are invalid -

 

All these seem to me to conclude that it is a driver issue?  I'm always open to correction...

 

Is there someone on the development team that can review the differences in the drivers between 5.0.6 and 6b15?  Perhaps we could roll back the drivers to 5.0.6?  I would be willing to test that out if someone can walk me through the installation process...seems like that would be the next phase of debugging.  Eliminating the driver as the issue...  Would these instructions be appropriate but using the older drivers?  http://www.intel.com/support/network/sb/CS-032514.htm

 

Better yet, is there some sort of logging or tracing we can do internally to a file to try to trap the occurrence?

 

Is anyone else using the ASUS P8Z68-v Pro motherboard NIC or the Pro/1000 NIC and UR6b15?

Just had someone ask me about this post.  So updating it as solved.  The problem turned out to be the Asus P8Z68-V Pro motherboard.  It just does not work with unRaid 6.1.  My ultimate solution (after several weeks of trying) was to replace it with a different ASUS motherboard (See my signature for system config).  Works perfect.  Never did find out why.