bear with me because i am typing this on a phone and am about to go to bed...

 

came home yesterday to disk6 red balled.  stopped array, ran smart, passed.  unassigned drive, started array, stopped array, reassigned drive, started array.  disk rebuilt, green ball.  ran parity check, came home this morning to same disk (6) red balled.  stopped array, and five additional disks went red after the array was stopped.  disk6 red square, other disks (5, 7, 8, 9, 10) red balls.

 

latest v5 running on this system.

 

could i be looking at controller failure?  drives that are red balled -- some on onboard sata, some on supermicro pcie controller.  lightning storm last night while at work... maybe something fried?

 

will troubleshoot when i wake up but next days off are thurs/fri so limited time to really dig into it,

 

any ideas welcome.

also, all red balled disks say "no device"

Sounds like a controller issue

Or power issue, or motherboard issue.  Do you happen to have the syslog from that session (before any reboot)?  (ALWAYS grab the syslog when there is trouble!)

No syslog, unfortunately.

 

Upon further inspection, a power cable was loose from one of my iStarUSA 5-in-1 cages (each cage takes two power cables).  However, the loose power cable was on the cage WITHOUT the red-balled drives.  The cage with (all) the red-balled drives had both power connectors on it.

 

I reset the system and the Supermicro 8-port controller found all drives.  Unraid then showed everything good-to-go except Disk 6, still red-balled.  I powered down and reseated the SATA controller and all SATA cables, unraid came up and I removed/added Disk 6, rebuilt the drive, checked parity, and now am at 100% and all green balls.  Unfortunately, last time I rebuilt the drive, I went to bed with all greens, which is what I am about to do again.  So I will wake up and check to see if I am still all green.  If not, I am going to swap out Disk 6 for my spare and then run some long SMART and preclear sequences on it.  Its a WD Green 3TB, and recent feedback on Newegg shows a LOT of them dying premature deaths.  So we'll see.

 

Right now everything seems OK.

Happened again today, this time disk 8.

 

Powered down the server and rebooted, four disks came up as missing.

 

All connectors (power and SATA) were seated properly.  Moved the Supermicro 8-port to a different pcie slot and rebooted, all green except disk 8, removed, re-added, and am now doing a re-build.

 

I am trying to slowly eliminate variables here.  If the array pukes again, I'll throw in a couple of older PCI sata controllers and see what happens.  Hard to believe the Supermicro card just suddenly went bad, although I guess its possible.  If problems arise with other controllers, I suppose I'll replace the power supply, although during parity checks and other (presumably) power-intensive tasks, the system appears fine.

 

Starting to get frustrated.

 

As an aside, whenever the array goes into a state where a disk fault has occurred, its entirely unresponsive for telnet, connection via windows shares, etc.  The only thing that works is the web server, and its s-l-o-w.  Not sure what all that means.  When everything is green/green the system runs speedy and very responsive.