Swapping a parity drive and adding a data drive, may have messed up

[Varean]

So, I am new to UnRaid and I was doing my research before I started this process to make sure I had all the information and I thought swapping out my smaller parity drive for a new one, and then adding my old parity drive to the array should be easy. My Array had 5 devices, 4 data and 1 parity. I ran a parity check before doing anything and parity showed as valid so I thought I would be good to go, I didn't want to risk a loss of data during a parity rebuild so I placed the new Parity drive (after preclear) as a second parity drive and set it to sync, during the sync process (which is still going) I have one data drive that is showing 43 errors. that up until now passed all my smart tests (I try to run them once a month after parity). I am concerned that I may have gotten myself into a bind as I am not sure if I can stop the array, remove the smaller parity drive and set the bigger one as Parity, instead of Parity 2. I also am not that good at reading SMART tests and I can't really tell what the errors are coming from in Unraid for my Drive 2, (link to smart test output for data drive 2)

 

Any help is appreciated as I am concerned I am going down a road that might lead to some bad drive failures and/or data loss.

skynet-diagnostics-20200212-0853.zip

[trurl]

Quick skim over SMART for all disks, didn't notice any issues. SMART for all disks are included in diagnostics so no need to give any separately. Do any of them have SMART warnings on the Dashboard other than CRC? CRC, typically connection issues, can be acknowledged by clicking on the warning.

 

Those are read errors on disk2, likely you disturbed a connection when installing drives. You must always double check all connections, power and SATA, including power splitters, whenever you are mucking about in the case.

 

Also noticed your appdata share is cache-only, but it has files on the array. Probably that isn't what you intend. We can discuss how to fix that later.

 

Do you intend to have dual parity? It doesn't sound like that is what you want. Parity and parity2 are not interchangeable, so even if you succeed in your approach, you will not have a parity disk, you will have a parity2 disk. That is not necessarily wrong, but maybe not what you had in mind. It would have been much simpler if you had just replaced and rebuilt parity to the larger disk. And it is still possible to do it that simpler (and more standard) way if you want to do that instead.

 

Stop the array, go to Settings - Disk Settings and disable autostart. Then shutdown, double check all connections, reboot.

 

Then tell us what you want to do.

[Varean]

8 minutes ago, trurl said:

Quick skim over SMART for all disks, didn't notice any issues. SMART for all disks are included in diagnostics so no need to give any separately. Do any of them have SMART warnings on the Dashboard other than CRC? CRC, typically connection issues, can be acknowledged by clicking on the warning.

 

Those are read errors on disk2, likely you disturbed a connection when installing drives. You must always double check all connections, power and SATA, including power splitters, whenever you are mucking about in the case.

 

Also noticed your appdata share is cache-only, but it has files on the array. Probably that isn't what you intend. We can discuss how to fix that later.

 

Do you intend to have dual parity? It doesn't sound like that is what you want. Parity and parity2 are not interchangeable, so even if you succeed in your approach, you will not have a parity disk, you will have a parity2 disk. That is not necessarily wrong, but maybe not what you had in mind. It would have been much simpler if you had just replaced and rebuilt parity to the larger disk. And it is still possible to do it that simpler (and more standard) way if you want to do that instead.

 

Stop the array, go to Settings - Disk Settings and disable autostart. Then shutdown, double check all connections, reboot.

 

Then tell us what you want to do.

For errors on the dashboard, I don't see anything other than the CRC error, I did some searching and I found some articles say its a bad SATA cable, and others say it could be a bug with the drive itself potentially/being misreported. The Dashboard is actually showing Drive 2 as 'Healthy'.

 

As for the AppData share, I am not sure if it's because of this, but I have a Plugin to backup my docker containers and I store the backup on the array. I didn't otherwise notice that there were files on the array that should be on the cache drive.

 

 

My original intention was to, instead of rebuilding parity on a brand new drive, have Unraid copy the parity information from Parity to Parity 2, so I wouldn't be without parity drive protection (especially if a data drive died in the rebuild). Then simply remove the old parity drive and re-add it as a Data drive, if I can't set the Parity 2 drive as Parity drive and have to leave it as Parity 2 that's fine as long as it doesn't cause any issues down the line.

[JorgeB]

40 minutes ago, Varean said:

that up until now passed all my smart tests

That disk has been failing the extended test for some time:

SMART Extended Self-test Log Version: 1 (1 sectors)
Num  Test_Description    Status                  Remaining  LifeTime(hours)  LBA_of_first_error
# 1  Short offline       Completed: read failure       10%      3775         4296582856
# 2  Extended offline    Completed: read failure       20%      3743         4296582856
# 3  Extended offline    Aborted by host               80%      3738         -
# 4  Extended offline    Completed: read failure       10%      3731         4296582856
# 5  Short offline       Completed: read failure       10%      3716         4296582856
# 6  Extended offline    Completed: read failure       20%      1600         4296582848
# 7  Short offline       Completed without error       00%      1595         -
# 8  Short offline       Completed: read failure       10%      1509         4296582848
# 9  Extended offline    Completed: read failure       20%      1481         4296582848
#10  Short offline       Completed without error       00%      1449         -

 

You should replace it, or try preclearing, it might turn good, though it's more likely to fail again in the near future.

[Varean]

10 minutes ago, johnnie.black said:

That disk has been failing the extended test for some time:

SMART Extended Self-test Log Version: 1 (1 sectors)
Num  Test_Description    Status                  Remaining  LifeTime(hours)  LBA_of_first_error
# 1  Short offline       Completed: read failure       10%      3775         4296582856
# 2  Extended offline    Completed: read failure       20%      3743         4296582856
# 3  Extended offline    Aborted by host               80%      3738         -
# 4  Extended offline    Completed: read failure       10%      3731         4296582856
# 5  Short offline       Completed: read failure       10%      3716         4296582856
# 6  Extended offline    Completed: read failure       20%      1600         4296582848
# 7  Short offline       Completed without error       00%      1595         -
# 8  Short offline       Completed: read failure       10%      1509         4296582848
# 9  Extended offline    Completed: read failure       20%      1481         4296582848
#10  Short offline       Completed without error       00%      1449         -

 

You should replace it, or try preclearing, it might turn good, though it's more likely to fail again in the near future.

I don't know if it's me or the drive or what, but there have been multiple occasions where I try to run the long SMART test, and it just sits on 90% complete for hours, and when I check it later either it shows as completed, or it will show as a lower progress percentage and I just abort it and try running it again. It might just be me being new to UnRaid, but the dashboard has been reporting that disk as Healthy, I just run the SMART tests and assume its going to update the dashboard if something comes up wrong. But I did have plans to upgrade a second drive, I can swap them and run a preclear on it to see if it clears up, if not I can just chuck it.

Edited February 12, 2020 by Varean

[trurl]

5 minutes ago, Varean said:

As for the AppData share, I am not sure if it's because of this, but I have a Plugin to backup my docker containers and I store the backup on the array. I didn't otherwise notice that there were files on the array that should be on the cache drive.

No the backup will not be the reason for the appdata share on the array, since the backup is not (or at least should not be) in the appdata share. And Mover ignores cache-only shares, so it won't move them to cache. And Mover can't move open files either. So you would have to disable docker service, set appdata to cache-prefer, run mover, enable docker service.

 

9 minutes ago, Varean said:

instead of rebuilding parity on a brand new drive, have Unraid copy the parity information from Parity to Parity 2, so I wouldn't be without parity drive protection

Unraid does not copy parity except in the case of the "parity swap" procedure, which is not what you were doing. Parity2 is calculated according to a totally different algorithm, so it is not interchangeable with parity. If parity and parity2 were the same there wouldn't be any extra redundancy for any of your data drives and the extra parity would only be useful if the other parity failed.

 

And if you notice in that first screenshot you linked, it was reading all the other disks in order to calculate parity2 and write it.

 

Speaking of "parity swap", that might be the way forward if you want to replace parity with that larger disk and disk2 with the old parity.

 

[JorgeB]

10 minutes ago, Varean said:

but the dashboard has been reporting that disk as Healthy,

Dashboard only reports the more common SMART attributes, it doesn't check SMART tests.

[Varean]

1 minute ago, trurl said:

No the backup will not be the reason for the appdata share on the array, since the backup is not (or at least should not be) in the appdata share. And Mover ignores cache-only shares, so it won't move them to cache. And Mover can't move open files either. So you would have to disable docker service, set appdata to cache-prefer, run mover, enable docker service.

 

Unraid does not copy parity except in the case of the "parity swap" procedure, which is not what you were doing. Parity2 is calculated according to a totally different algorithm, so it is not interchangeable with parity. If parity and parity2 were the same there wouldn't be any extra redundancy for any of your data drives and the extra parity would only be useful if the other parity failed.

 

And if you notice in that first screenshot you linked, it was reading all the other disks in order to calculate parity2 and write it.

 

Speaking of "parity swap", that might be the way forward if you want to replace parity with that larger disk and disk2 with the old parity.

 

So as for resolving the parity issue, are there any issues with my current data drives that you see that cause any issues if I were to simply stop the array, remove the current 6TB Parity drive, and place the 8TB drive as Parity and rebuild, then stop the array and set the 6TB drive as a Data disk drive to the rest of the array? My current hardware configuration only supports 6 disks total, so I only want to have a single Parity drive to maximize my total capacity.

I guess my weakness here is understanding/interpreting SMART test results.

[Varean]

6 minutes ago, johnnie.black said:

Dashboard only reports the more common SMART attributes, it doesn't check SMART tests.

Then that is a failure on my part, thank you for letting me know. I'll be ordering a drive today since I had planned on upgrading one of my 3TB disks anyway, I'll just upgrade that one first.

[trurl]

As mentioned, it is possible to "parity swap". That would get the new larger disk into parity slot and the original parity disk replacing disk2.

 

[Varean]

Just now, trurl said:

As mentioned, it is possible to "parity swap". That would get the new larger disk into parity slot and the original parity disk replacing disk2.

 

I apologize I misread your post and thought you were referring to making the bigger disk parity, and the old parity disk parity2. Yes I think that would be the procedure I would like to do, make the larger disk Parity and the old parity disk a replacement of Disk2.

[trurl]

Study this:

 

https://wiki.unraid.net/The_parity_swap_procedure

 

I think simply unassigning parity2 and starting the array will get you back to where you can begin this procedure.

 

Is that correct @johnnie.black or will New Config / Trust Parity be needed?

 

[JorgeB]

Just now, trurl said:

I think simply unassigning parity2 and starting the array will get you back to where you can begin this procedure.

Yep, that will do it.

[Varean]

7 minutes ago, trurl said:

Study this:

 

https://wiki.unraid.net/The_parity_swap_procedure

 

I think simply unassigning parity2 and starting the array will get you back to where you can begin this procedure.

 

Is that correct @johnnie.black or will New Config / Trust Parity be needed?

 

I think you both are the best and have been extremely helpful to someone new like me. I'll go through and read the documentation on this procedure, and report back once I've completed it. I will also do my due diligence and read up further on the SMART reporting done in UnRaid so I don't run into another issue like I did with Disk2.