Hey guys.

 

So I just replaced a 2TB drive with a  3TB drive a few days ago.  I connected another new drive to the system today to pre-clear it and when I ran "preclear_disk.sh -l" it showed the new drive and the drive I just added.  That's wierd, so I stopped the array so I could reboot.  Once the array had stopped my receipt replaced drive is showing a red ball.  I was able to start the array, it's still showing red ball but I can read and write to it just fine.  I did a SMART test and it looks fine.

 

Here's my syslog from today - https://mega.co.nz/#!XRcEUCQa!Wbbv7U08usW09ZHG7sKP4VLKyFM9vizwg6MEJPVvMmw

 

The drive in question is disk9 (sdn)

 

Thanks.

 

If it shows a red ball it's been disabled.  You can still read/write to it because UnRAID is emulating it with the other disks plus parity => that's what fault tolerance is all about 

 

You need to determine what's wrong BEFORE you do anything else with your array -- you're currently "running at risk" ... so any other failures will result in data loss.

 

Shut down;  reseat the disk (if in a caddy);  unplug/replug the cables (in fact use a different SATA cable if you have one);  then reboot; Stop the array;  unassign the failed drive;  Start the array;  Stop it again and re-assign the failed drive; then Start the array again and it should rebuild the drive.    If that works okay, you're fine.  If not, replace the drive.

 

Does the fact that it was triggered by an array stop tell you anything?  It seems rather odd that it would just decide to fail when i initiate the stop.

When you stop the array a sync is sent, writing to the drives, if that write failed, the drive will red ball

 

Sent from a phone, sorry for any typos

 

 

If it shows a red ball it's been disabled.  You can still read/write to it because UnRAID is emulating it with the other disks plus parity => that's what fault tolerance is all about 

 

You need to determine what's wrong BEFORE you do anything else with your array -- you're currently "running at risk" ... so any other failures will result in data loss.

 

Shut down;  reseat the disk (if in a caddy);  unplug/replug the cables (in fact use a different SATA cable if you have one);  then reboot; Stop the array;  unassign the failed drive;  Start the array;  Stop it again and re-assign the failed drive; then Start the array again and it should rebuild the drive.    If that works okay, you're fine.  If not, replace the drive.

 

I'm doing the rebuild now.  When I swapped the power/data cables the first time the drive didn't detect then re-seated again and it did.  The drive might be bad, it's a brand new WD Red and we all know they've had a troubled past. If I get another red ball on it, I have a 3TB green ready to go.

rebuild went fine overnight, doing a parity check now.  I just noticed that my cache drive is no longer present, maybe it was failing and caused a bad write onto disk9.  I'll have to check it's connections when I get home from work.  My SAS cables are a bit short in my new D8000 case so maybe they're coming lose, hopefully my monoprice order comes sooner than later.

Definitely sounds like your cables may not be well-seated.  Installing longer (and I assume locking) cables should make this much more stable.

 

My buddy saved the day with a pair of longer cables while I wait on monoprice.  I did some fiddling with mine and it's a specific one of the 4 breakouts that's not 100% reliable.

 

Cache is back online and i'm all green.  Thanks so much for the help!

 

Now I can move forward with rolling rebuilds of my two non MBR-aligned disks.

Now I can move forward with rolling rebuilds of my two non MBR-aligned disks.

Why are you risking your data on a cosmetic non-issue? You are gaining nothing by changing from sector 63 to sector 64 unless you are running WD EARS drives with no jumper. You are risking a double drive failure by intentionally failing a drive simply to rebuild it. If you happen to have bad luck and a different drive fails while you are in process, you stand a good chance of losing the content of both drives. You already seem to be having stability issues, I wouldn't push my luck.

I was rock solid before this cable issue and I believe the two unaligned EARS are negatively affecting my speeds.  Worst case scenario, I lose a TB or two of data that can be replaced.  One of the two rebuilds is going to be an upgrade anyway.

 

I just want everything to be as perfect as possible and my OCD agrees

(a)  Completely unnecessary to do that

 

(b)  I completely understand  :)

 

[Hint:  I'd probably do the same thing -- I have an equally bad case of OCD !! ]