UnRAID unresponsive + all network down

[snidera]

I'm having an issue where UnRAID becomes unresponsive, as well as my entire network loses connectivity.  Ping unresponsive on multiple (probably all) devices connected to the router.  Remove network cable from UnRAID box & everything goes back to normal instantly (or as fast as I can run over to a PC & check).  Plug it back in, everything fails.   yellow LED on UnRAID ethernet adapter is flashing steadily/slowly - green is activity, it is solid.

Even after removing the ethernet cable, UnRAID seems frozen.  Power button doesn't respond, have to cut power (hold button or unplug AC)

This happens ~4+ hours after boot, has happened 3 times in about 12-15hrs.

It happened 1x before the 3 week win10 run, but I chalked it up to lightning/power blip + it was on a different power supply (same power supply is currently in use & stable on another system)

Same HW/SW config was steady for 8+ days on a different router/location (Asus RTn-10)

Same MB/proc/RAM was steady for 3+ weeks running Win10 off a HDD (thrashing CPU with handbrake)

killed my win10 VM before it happened this last time - don't think that is affecting it.

 

This last time I was copying files from my other server (win7), the HDD activity LED is solid on.   The times before, server was idle & HDD LED was off.

 

Archer C7 3150 

UnRAID 6.3.5

Asrock H170M Pro4

i5-6500 (intel cooler)

16GB RAM

2x 6TB Toshiba's

1TB WD blue (Cache)

HAF912 case - fans running in 'performance' mode.  HDDs never get above 35*

Corsair TX550m

 

No IP conflicts on network - I have set static DHCP for all devices & place devices in ranges for easy QOS control 

 

I thought about power (AC) - but the 3+ weeks was on the same surge protector, along with another PC that never missed a beat (using 2 win PCs to compress DVD/BRs via handbrake).  I now have it plugged into UPS to try to rule out dirty AC.

 

Suggestions to diagnose?

MB was used/ebay, it's my prime suspect - suggestions on diagnosis/fixes?  All other components are new.  But again, stable with windows & for 8 days on another router

Router?  How do you get ethernet wrong?  This router has been great for 8 months when my other one got struck by lightning.

UnRaid - this USB has been stable & in use for over a year?  I wiped it, downloaded 6.3.5 & set this server up fresh to take over main duties.  Thought 8+ days of burn in was enough.

RAM?  Ran memtest for a while, not my first suspect

CPU?  Same as RAM - it's been pushed via handbrake & was stable

Heat?  doubt it, but never know.

[FreeMan]

Post the complete diagnostics.zip file (Tools -> Diagnostics).

 

There may be hints in there that someone will be able to find for you.

[snidera]

I checked logs when it happened last time & couldn't see any trace.   I thought they were cleared on reboot (I'm a bit dumb in this area). 

 

Attached are from after a hard reset.

tower-diagnostics-20171004-1930.zip

[snidera]

Well, 22hrs & 4TB of data transferred later, it looks like a crappy powerstrip/surge protector might have been the issue.  Either that or my server knows I made this thread & is screwing with me.