[SOLVED] Replace Faulty Disk with New Drive Larger than Parity - v6.0-beta12

[tomorrowsman]

Hi there,

 

I'd like to make sure I get this right:  I have a four-disk array; Parity is 2TB, then a 2TB, 1TB, and 750GB.  The 2TB -- which is my newest drive -- is now reading as faulty.  I purchased a new drive that is 3TB.  In which order should I upgrade, given that the new drive is larger than my Parity drive?

 

Also, does anyone know a way to test for the source of the fault?  The drive's self-test was fine, and it is a fairly new Seagate Baraccuda 7200.14 (and SMART test came back as PASSED).  I am wondering if there is a painless way I can track down if it is possibly voltage, the SATA connector, etc. without having to acquire an engineering degree (though I'd love to).

 

Cheers,

Chris

[JonathanM]

The 2TB -- which is my newest drive -- is now reading as faulty.

Red ball is not automatically a bad drive. It means a write to the drive failed, which can be any number of things. Hopefully you haven't rebooted the server yet, because a syslog that contains the time period where the drive was red balled should have clues as to what caused the failed write. Zip up the syslog and a smart report for the drive, and post it here, and someone should be able to give you a better idea what happened, and what your next move should be.

 

The parity drive must be the largest drive in the protected array, so having a larger replacement drive complicates things. Hopefully your current drive is fine, and you can rebuild back to itself and then worry about swapping in the larger parity drive.

[tomorrowsman]

Aggravatingly, we had a power-outage midday a few days back while I was at work, so even my UPS died.  I reckon that means my syslog is kaput.  So rule of thumb is always to ensure 'green' drives first and parity is good, then swap up the parity drive for a bigger one, correct?  Will parity remain invalid until I can clear the red dot?

[itimpi]

Aggravatingly, we had a power-outage midday a few days back while I was at work, so even my UPS died.  I reckon that means my syslog is kaput.  So rule of thumb is always to ensure 'green' drives first and parity is good, then swap up the parity drive for a bigger one, correct?  Will parity remain invalid until I can clear the red dot?

There is a process called swap-disable that allows for the new larger drive to become the new parity drive, and the old parity drive to be used for rebuilding the failed drive.  You should be able to get the exact steps to do this by searching the forum.

 

Having said that if the problem drive was red-balled due to an external issue and the drive is OK then simply rebuilding onto that drive might be the easiest way to go initially.

[garycase]

Here's the Wiki entry for swap-disable, which is a procedure which will let you use your new 3TB drive as parity while using the old parity drive to replace your failed drive:

 

--------------------------------------------------------------------------------------------

 

... If the replacement disk is larger than your parity disk, then the system permits a special configuration change called swap-disable.

 

For swap-disable, you use your existing parity disk to replace the failed disk, and you install your new big disk as the parity disk:

1. Stop the array.

2. Power down the unit.

3. Replace the parity hard disk with a new bigger one.

4. Replace the failed hard disk with you old parity disk.

5. Power up the unit.

6. Start the array.

 

When you start the array, the system will first copy the parity information to the new parity disk, and then reconstruct the contents of the failed disk.

 

--------------------------------------------------------------------------------------------

 

 

[garycase]

... However, as already noted, you should first confirm that you didn't just have a loose cable; poorly seated drive (if in a hot-swap cage); or some other external issue.    If you can simply rebuild the drive onto itself, then you'll be able to just do a simple parity drive upgrade.

 

[tomorrowsman]

Hmm, so if the swap-disable worked, would I recover only up to the day the 2nd drive went "red ball" on me, since Parity has been invalid since then?

[tomorrowsman]

Another thought -- if indeed the issue could possibly be related to the power problems we had, is there a way to ask the system to re-read that disk, i.e., re-integrate it into the array?  And what would happen if the disk is truly fried somehow (though that seems so unlikely)?

[JonathanM]

would I recover only up to the day the 2nd drive went "red ball" on me, since Parity has been invalid since then?

If you have 2 bad drives, nothing can be recovered from them using parity. I'm rather confused as to what state your array is in right now, as this is the first you mentioned a second failure. Can you post a screen shot of the main page that shows which drives are red balled?

[garycase]

Hmm, so if the swap-disable worked, would I recover only up to the day the 2nd drive went "red ball" on me, since Parity has been invalid since then?

 

Why do you think parity is invalid?    IF that's the case, you can't rebuild your bad disk.    But this is the first you've mentioned that.

 

IF parity is valid, then your "bad" drive has been emulated since it was red-balled, so the emulated data is actually more current than what's physical on the drive ... i.e. when you "write" to that drive, you're really writing to every drive EXCEPT that one.    So if you do a rebuild (or a swap-disable), your data on the rebuild drive will all be current.

 

But IF you've had a 2nd failure, then you simply need to start from scratch with a new configuration and rebuild parity (WITHOUT including the failed drive in the array); and then use some data recovery tools to try and read what you can from the failed disk.

 

 

[tomorrowsman]

Sorry for the confusion!  Only Disk2 is red-balled; but on my Dashboard, I have a "Data is invalid" alert.  I must have misinterpreted this as meaning Parity is invalid, but I'm now guessing it just means Disk2.  Here's a screenshot...

[storagehound]

Hello,

I've purchased a DESKTAR NAS 6TB to replace my 4TB parity drive (supplied by unRAID).  Is the procedure below the way I should handle the upgrade? I want the 6TB drive to now be the parity and the old 4TB one to become a regular drive.  I'd like to do this tonight.  Also, does a pre-clear process still need to be done in this scenario?

 

Here's the Wiki entry for swap-disable, which is a procedure which will let you use your new 3TB drive as parity while using the old parity drive to replace your failed drive:

 

--------------------------------------------------------------------------------------------

 

... If the replacement disk is larger than your parity disk, then the system permits a special configuration change called swap-disable.

 

For swap-disable, you use your existing parity disk to replace the failed disk, and you install your new big disk as the parity disk:

1. Stop the array.

2. Power down the unit.

3. Replace the parity hard disk with a new bigger one.

4. Replace the failed hard disk with you old parity disk.

5. Power up the unit.

6. Start the array.

 

When you start the array, the system will first copy the parity information to the new parity disk, and then reconstruct the contents of the failed disk.

 

--------------------------------------------------------------------------------------------

[trurl]

Hello,

I've purchased a DESKTAR NAS 6TB to replace my 4TB parity drive (supplied by unRAID).  Is the procedure below the way I should handle the upgrade? I want the 6TB drive to now be the parity and the old 4TB one to become a regular drive.  I'd like to do this tonight.  Also, does a pre-clear process still need to be done in this scenario?

Do you actually have a faulty drive? If you are not trying to rebuild a bad drive, then there is no reason to use swap-disable. Just replace the parity drive and let it rebuild.

 

Do you want the old parity drive to replace one of your data drives, or do you just want to add it to the array? If replacing then just replace and let it rebuild. If adding, then you would need to preclear or unRAID will have to clear it for you and you will be offline while it does.

 

Also, while unRAID does not require parity to be clear, it is a good idea to preclear any new drive so it gets a good test.

[storagehound]

Thank you. I wanted to add a new drive.  Since my new drive is larger than my parity drive I understood I should use it to replace the old Parity.  Then I would take the old parity and add it to my array as an additional drive.  I'll figure out if I have the time to just let my server clear the drive or do a preclare.  I wish I ustility like Spinrite could do it.

Thanks again!

 

Hello,

I've purchased a DESKTAR NAS 6TB to replace my 4TB parity drive (supplied by unRAID).  Is the procedure below the way I should handle the upgrade? I want the 6TB drive to now be the parity and the old 4TB one to become a regular drive.  I'd like to do this tonight.  Also, does a pre-clear process still need to be done in this scenario?

Do you actually have a faulty drive? If you are not trying to rebuild a bad drive, then there is no reason to use swap-disable. Just replace the parity drive and let it rebuild.

 

Do you want the old parity drive to replace one of your data drives, or do you just want to add it to the array? If replacing then just replace and let it rebuild. If adding, then you would need to preclear or unRAID will have to clear it for you and you will be offline while it does.

 

Also, while unRAID does not require parity to be clear, it is a good idea to preclear any new drive so it gets a good test.

[BobPhoenix]

Thank you. I wanted to add a new drive.  Since my new drive is larger than my parity drive I understood I should use it to replace the old Parity.  Then I would take the old parity and add it to my array as an additional drive.  I'll figure out if I have the time to just let my server clear the drive or do a preclare.  I wish I ustility like Spinrite could do it.

Thanks again!

I've done that several times:  Run several preclears on the new parity drive before I add it as parity.  Backed up my flash drive.  Pulled the old parity drive and set it to the side and add the new drive as parity and rebuilt parity on it.  Once the parity build has completed do a parity check to make sure everything was built correctly.  Once that was done then put the old parity drive back into the computer and do a write only preclear before adding it to the array as a data drive.  Depending on the drives sizes and speeds that could take 1-3 days to complete.  If you like living dangerously and don't want to have a backup to fall back on you could do the write only preclear on a different box while the parity build/check is going on to cut down on the time.  But it is best to use the safe method and spend the extra time.

[storagehound]

This may be embarrassingly obvious, but... why does one need to backup the Flash Drive?  Is it just a good habit like changing smoke detector batteries with daylight savings time or something?

 

Thank you. I wanted to add a new drive.  Since my new drive is larger than my parity drive I understood I should use it to replace the old Parity.  Then I would take the old parity and add it to my array as an additional drive.  I'll figure out if I have the time to just let my server clear the drive or do a preclare.  I wish I ustility like Spinrite could do it.

Thanks again!

I've done that several times:  Run several preclears on the new parity drive before I add it as parity.  Backed up my flash drive.  Pulled the old parity drive and set it to the side and add the new drive as parity and rebuilt parity on it.  Once the parity build has completed do a parity check to make sure everything was built correctly.  Once that was done then put the old parity drive back into the computer and do a write only preclear before adding it to the array as a data drive.  Depending on the drives sizes and speeds that could take 1-3 days to complete.  If you like living dangerously and don't want to have a backup to fall back on you could do the write only preclear on a different box while the parity build/check is going on to cut down on the time.  But it is best to use the safe method and spend the extra time.

[JonathanM]

This may be embarrassingly obvious, but... why does one need to backup the Flash Drive?

Unraid will refuse to assign a smaller drive to a slot. So, if you don't backup the flash drive, and something goes wrong with a drive upgrade, then you are stuck trying to finish the upgrade because you can't easily put back the old smaller drive. If you have a backup of the flash, you can just put all the drives back the way they were before the upgrade, restore the flash, and unraid will happily start.

 

BTW, the array MUST BE STOPPED to get a good backup of the flash. Otherwise the backup will contain the running status of the array, and if you start the array in that state, unraid will start a parity check because it assumes it was shut down incorrectly.

[interwebtech]

BTW, the array MUST BE STOPPED to get a good backup of the flash. Otherwise the backup will contain the running status of the array, and if you start the array in that state, unraid will start a parity check because it assumes it was shut down incorrectly.

 

Did not know that. Filed in noggin array