Help, my Disk 2 (sdd) is showing red ball, what do I do?

[bobekdj]

I am afraid of another catastrophic failure due to issues with my unraid, so i figure i will ask the community what i should do, attached is the sys log file.

 

Any way I can figure out what to look for?  I have ran the smart test and it passed.

 

Thanks

syslog.txt

[trurl]

What issues are you having?

[bobekdj]

Sorry thought it was clear in the title.  Log is attached to first post.

 

 

Disk 2 (sdd) is showing red ball

 

Main dashboard photo attached.

 

unMenu page attached.  Appears the disk was disabled for some reason?

 

[Squid]

Attach a smart report for (sdd) WDC_WD30EFRX-68EUZN0_WD-WCC4NJPSF9CK

[trurl]

unRAID disables a disk when a write fails. Do you have a replacement disk? The usual advice is to rebuild on another disk, and then determine if the problems with that disk can be resolved.

 

Post a smart report of the drive.

[Squid]

You've rebooted since the redball occurred, so the syslog is clean and nothing in it indicates any issues with any drive.  A smart report will hopefully tell us if there is a problem with the drive, or if it was a cabling issue or what-not

[bobekdj]

Squid,

 

I ran the quick one and it passed.  I am running the full test now. 

 

 

How can I easily upload the results?

[Squid]

Been awhile since I've used unmenu, but I seem to recall that there's an option in there somewhere to run smart on a particular drive.

 

If you can find it, then you can always log into the system via telnet / putty and run

 

smartctl -a /dev/sdd

 

Then copy and paste the result

[bobekdj]

Ok.  I am actually running the test via 6b12 menu under the disk properties itself.  Not sure where it outputs when done other than on my screen?

 

 

Been awhile since I've used unmenu, but I seem to recall that there's an option in there somewhere to run smart on a particular drive.

 

If you can find it, then you can always log into the system via telnet / putty and run

 

smartctl -a /dev/sdd

 

Then copy and paste the result

[Squid]

just copy and paste the results then after you hit "Disk Attributes"

[Squid]

And the Disk Error Log

[bobekdj]

ok will do.

 

Its a 3TB drive... any idea how long the extended test takes??  Also does the extended tell me any extra info more so than the short self test??

[Squid]

It should tell you when you hit disk capabilities.  One of mine says 341 minutes.  The extended test checks all of the sectors.  You can see its progress under disk self-test log

 

But, it wouldn't to hurt to throw up the results from disk attributes now while you've got people's attention

 

[bobekdj]

412 minutes... lol.

 

should I just cancel at this point?

 

attributes to from previous quick scan to come:

 

It should tell you when you hit disk capabilities.  One of mine says 341 minutes.  The extended test checks all of the sectors.  You can see its progress under disk self-test log

 

But, it wouldn't to hurt to throw up the results from disk attributes now while you've got people's attention

[Squid]

The drive redballed...  Let it run...

 

Post the attributes now, and after its done.

 

Before and after will also gives us more info

 

[bobekdj]

{old}

6/6

 

Any recommendations on how to get around the 192kb limit of the message boards and quickly post multiple screenshots? lol

{/old}

 

[new]

 

Here are the new uploaded images to save posts :-)

 

 

 

 

 

 

 

 

 

 

 

 

[Squid]

6/6

 

Any recommendations on how to get around the 192kb limit of the message boards and quickly post multiple screenshots? lol

Upload them to photobucket then include the url.

 

[Squid]

The smart status looks ok.  There's no reallocated sectors or pending sectors.  There's only one error listed "READ DMA"

 

Best guess right now:  If you had the syslog when it redballed (before a reset), there would have been a whack of ATA errors in it along the lines of ata11.00: failed command: READ FPDMA QUEUED

 

This usually happens when there's a bad / loose cable and/or the power supply is weak / too many splitters, etc.

 

I'd wait for the extended smart test to finish and lets see what the results are.

 

But, right now I'd say that your course of action is going to be to reseat all of the cables to the drive and mobo, and then rebuild the same drive onto itself, while watching the syslog for any errors similar to the above.

 

[Squid]

6/6

 

Any recommendations on how to get around the 192kb limit of the message boards and quickly post multiple screenshots? lol

Upload them to photobucket then include the url.

Or just highlight the relevant text and paste them here into a "code" box

[bobekdj]

Ok... that makes sense.

 

considering its at 30% i am not sure if it will complete tonight...

 

So if the smart report in unchanged then I need to shut down, unplug and plug back in each connector.

 

Then how do I rebuild on itself?  I dont think I see that option.  I have learned in the past to not reset the config because that can cause more errors than good,

  - so how do i rebuild the drive onto itself?

 

 

The smart status looks ok.  There's no reallocated sectors or pending sectors.  There's only one error listed "READ DMA"

 

Best guess right now:  If you had the syslog when it redballed (before a reset), there would have been a whack of ATA errors in it along the lines of ata11.00: failed command: READ FPDMA QUEUED

 

This usually happens when there's a bad / loose cable and/or the power supply is weak / too many splitters, etc.

 

I'd wait for the extended smart test to finish and lets see what the results are.

 

But, right now I'd say that your course of action is going to be to reseat all of the cables to the drive and mobo, and then rebuild the same drive onto itself, while watching the syslog for any errors similar to the above.

[Squid]

Ok... that makes sense.

 

considering its at 30% i am not sure if it will complete tonight...

 

So if the smart report in unchanged then I need to shut down, unplug and plug back in each connector.

 

Then how do I rebuild on itself?  I dont think I see that option.  I have learned in the past to not reset the config because that can cause more errors than good,

  - so how do i rebuild the drive onto itself?

 

 

The smart status looks ok.  There's no reallocated sectors or pending sectors.  There's only one error listed "READ DMA"

 

Best guess right now:  If you had the syslog when it redballed (before a reset), there would have been a whack of ATA errors in it along the lines of ata11.00: failed command: READ FPDMA QUEUED

 

This usually happens when there's a bad / loose cable and/or the power supply is weak / too many splitters, etc.

 

I'd wait for the extended smart test to finish and lets see what the results are.

 

But, right now I'd say that your course of action is going to be to reseat all of the cables to the drive and mobo, and then rebuild the same drive onto itself, while watching the syslog for any errors similar to the above.

When plugging / unplugging be very careful not to disturb the other connections  (I'm a huge fan of hotswap bays because of this)

 

To rebuild onto itself, you're going to stop the array.  Change the disk to be not installed and then start the array.  Then you will stop the array, change the drive from being not installed to be the WD30... and start the array.  It will start rebuilding.

[bobekdj]

Here is the Extended SMART results...

 

Looking like from the error log it was some sort of Read DMA error?  Is this possibly a bad cable, or is it more likely a bad sector on the drive itself?

 

 

 

 

 

 

 

 

[Squid]

Cable.  Doesn't look like anything else changed

[bobekdj]

So should I replace the cable all together and not just reseat?  Also got a new drive, but should I try to rebuild on the old one first because the new one has not been precleared?

 

Cable.  Doesn't look like anything else changed

[Squid]

The odds of a cable actually going bad is miniscule.  More likely a reseating issue

 

If it was me, I would rebuild the existing drive and preclear the new one at the same time.  If the existing fails again, then you're already half way through the preclear.