Help: How can I swap a disabled drive with an unutilized drive in array?

[akennealy]

Hi,

I ran into a drive failure (no idea what the failure is -- brand new WD drive!) but I need to replace the disabled drive.

In the past, I've simply replaced with a new drive. However, this time I'd like to use a drive that's already part of my array that has no content on it, 0% utilized.

 

What's the best way to perform this swap and array shrink? I couldn't find a good answer here. My array is currently unprotected due to the disabled drive.

 

 

My thought was to stop the array and set disk 11 to none, then set disk 10 to what disk 11 was set to. But I have a feeling this isn't what I want to do!

[Squid]

You can't do that because the system won't be able to rebuild the bad drive due to the missing drive (single parity), or you'll be running the array in a degraded state (dual parity)

 

Actual drive failures are exceedingly rare.  The usual problem is poor connections to the drive / sata ports.

 

Post your diagnostics.

[akennealy]

Thanks for the quick reply! 🙇‍♂️

 

As for the connection issue, I have seen a number of drives have with udma crc error. I'm trying to track that done, trying new cables etc.

however, this disabled drive is making a fair bit of unusual noises while running...

I've attached the diagnostics for the disabled drive. 

WDC_WD80EFAX-68KNBN0-diagnostics-20210419 disk10 (sdj) - DISK_DSBL.txt

[John_M]

One option would be to copy the files from the emulated Disk 10 to Disk 11, followed by shrinking the array to eliminate Disk 10 and re-sync parity. Another would be to buy another disk to rebuild Disk 10 onto, which would allow you to investigate the "failed" disk at your leisure, possibly (probably) resulting in it being usable as a spare. The other option is, as Squid suggests, to rebuild Disk 10 onto the original disk if it turns out to be still good, guided by your diagnostics.

[trurl]

SMART for disk looks OK, but it hasn't had a complete extended SMART test.

 

You should post the complete Diagnostics ZIP.

[akennealy]

OK, I'll take a look at what's included in the full Diagnostics ZIP. I don't want to post anything here that I shouldn't...

[akennealy]

Thanks all for the help! I can't figure out what's wrong with the drive.

The extended SMART test doesn't seem to complete. Looks like maybe timing out? Says aborted by host.

 

Here's the full full Diagnostics ZIP. 

 

Edited April 20, 2021 by akennealy
remove zip

[John_M]

21 minutes ago, akennealy said:

The extended SMART test doesn't seem to complete. Looks like maybe timing out? Says aborted by host.

 

It's a bug in Unraid 6.9. Reported and acknowledged. Set the spindown delay for that disk to "Never" until it has completed.

 

 

Edited April 19, 2021 by John_M
Added link

[John_M]

Diagnostics don't show why it was disabled because you've rebooted since, but as indicated it's most likely a cable problem. If it passes the extended SMART self-test you should replace the SATA cable, then you can re-build onto the same disk.

[akennealy]

12 minutes ago, John_M said:

Diagnostics don't show why it was disabled because you've rebooted since, but as indicated it's most likely a cable problem. If it passes the extended SMART self-test you should replace the SATA cable, then you can re-build onto the same disk.

 

Thanks! I'm running the extended again with spin down disabled. I can post results when complete. 

[6of6]

Never spin down disks, ever!... Disks DO fail even if constantly running. (My two cents and 30 years on disks with 6 failures [3xWD/3xSeagate].)

 

That said... From the top of the post, I don't understand why Disk 11 can't replace Disk 10?... I understand it in hardware->software (disk 10 and 11 must be removed and replaced by disk 11, but... well... I know the OS can't know a good drive is waiting there... or, can a good OS know it?

 

6.

[itimpi]

1 hour ago, 6of6 said:

rom the top of the post, I don't understand why Disk 11 can't replace Disk 10

The problem is that even an disk with no files but has been formatted to create an empty file system  (what a user normally means by an ‘empty’ disk) is included in the parity calculations and cannot be removed without invalidating its role in the parity calculations.    This would mean that there is no longer enough redundancy information to allow disk10 to be rebuilt.

 

this is one reason why it is frequently recommended to not add a disk to the array before you require (or will soon require) the space.   Once added the drive is no longer available to be used as a replacement for another drive that then fails.

 

[akennealy]

The extended SMART test passed. I swapped out the SATA cable and port and triggered a rebuild on the same disk 🤞

 

Once my array is healthy again, I'll plan to downsize and remove disk 11. I'll have to look up how to do this properly.

[trurl]

1 hour ago, akennealy said:

I'll plan to downsize and remove disk 11. I'll have to look up how to do this properly.

New Config without the disk and rebuild parity.

 

https://wiki.unraid.net/Manual/Storage_Management#Removing_data_disk.28s.29