Parity Device Disabled in Slot 1, works fine in Slot 2?

[Paul Rockliffe]

I've been struggling with my Parity drive for a while now, it shows absolutely fine in the logs, in SMART tests and if I remove it and put it in another machine it tests and performs faultlessly.  But setting it as parity and turning on the array would cause it to be disabled.  I've just removed it from Parity Device 1 and put it in Parity Device 2 and now it is happily rebuilding.  It's run for 5 hours now, so I'm fairly confident that it'll be OK.  

 

Is this a bug, or something that has happened to others before?

 

Secondly, while I've had no parity one of my drives has started showing errors.  It's been working fine, but showing maybe 500 errors, then a few days ago it started reporting that it is "Unmountable: Unsuported or not file system."  As Parity rebuilds it is reading the drive though.  Will the rebuild be good enough that I can swap that drive out for a new one and have it rebuild most of the data, or is that just lulling me into a false sense of security?

 

Thanks!

[dboonthego]

1 hour ago, Paul Rockliffe said:

Is this a bug

If you change the disk configuration it is normal to undergo a rebuild.

 

1 hour ago, Paul Rockliffe said:

As Parity rebuilds it is reading the drive though.  Will the rebuild be good enough that I can swap that drive out for a new one and have it rebuild most of the data, or is that just lulling me into a false sense of security?

False sense.  Parity is calculated against all your data disks and requires 0 read errors to be valid.  

 

Sounds like you have single parity with multiple disk failures.  If the errors are due to bad sata/power connections, then you can probably correct them by reseating or replacing the cables (eliminate splitters if possible).  If read errors are due to cables, you can do a new config and retain all disk assignments to rebuild parity. Don't rebuild the data disk on-top of itself because your parity as of now is no good.  You can also try  mounting the disk with Unassigned Devices plug-in and copy the contents elsewhere.

Edited October 21, 2023 by dboonthego

[JorgeB]

Please post the diagnostics.

[Paul Rockliffe]

Thanks, I meant a bug in that I have never managed to get the Parity drive to initiate a rebuild in slot 1, it would always either be disabled immediately or very soon after.  In slot 2 it immediately started to rebuild and is still rebuilding now without being disabled.

 

I'm not sure what's what with the drive with errors, it doesn't make sense to me that it says it unmountable, but is otherwise appears active in the array and is green, the parity rebuild has read the whole thing now and shows 2,114 read errors.

 

As you can probably tell, I don't really know what I'm doing with this, but if I know the best course of action from where I am then I think I understand enough to follow it.  I've attached diagnostics.

 

Thanks!

unraid-diagnostics-20231022-1004.zip

[itimpi]

9 minutes ago, Paul Rockliffe said:

Thanks, I meant a bug in that I have never managed to get the Parity drive to initiate a rebuild in slot 1, it would always either be disabled immediately or very soon after.  In slot 2 it immediately started to rebuild and is still rebuilding now without being disabled.

 

That strongly suggests there is some sort of hardware issue with slot1 (likely power or SATA connections) as there is no software related reason it should work in one slot and not another.

[Paul Rockliffe]

Sorry, I meant that in my array setup under devices I have the option for both Parity 1 or Parity 2.  If I put the drive in Parity 1 it is disabled, if it's in Parity 2 it is rebuilding.  The drive is physically connected the same way and in the same slot, it's purely an UnRaid setup thing.  I think,

[itimpi]

Just now, Paul Rockliffe said:

Sorry, I meant that in my array setup under devices I have the option for both Parity 1 or Parity 2.  If I put the drive in Parity 1 it is disabled, if it's in Parity 2 it is rebuilding.  The drive is physically connected the same way and in the same slot, it's purely an UnRaid setup thing.  I think,

 

Never heard of such an issue before, and cannot think what might cause it.

[dboonthego]

1 hour ago, Paul Rockliffe said:

In slot 2 it immediately started to rebuild and is still rebuilding now without being disabled.

Party2 is showing invaild by the way. 

 

There are read errors with Disk4 and it has has a bunch of re-allocated sectors and failed the S.M.A.R.T. test.  You won't get a successful parity sync until that disk is corrected.  

[JorgeB]

Diags only sow the disk assigned to parity2, but first you need to look at disk4, looks like it's failing, you can run an extended SMART test to confirm

[Paul Rockliffe]

OK, thanks.  it's not a problem to remove Disk 4 entirely, I have space on the array.  What's the process for getting as much of the data off that drive and onto other drives on the array as possible?

[JorgeB]

You can try copying the data manually or using ddrescue to clone it.

[Paul Rockliffe]

Just coming back to this now, I managed to work through ddrescue, which identified a number of errors on disk 4.  I emptied another disk and cloned to that disk, however that disk now will not mount, with the log stating, "Metadata corruption detected at xfs_btree_sbloc........"

 

I've run through the clone process twice now, with the same result both time.  I've run xfs repair too, that finds no Super Blocks.

 

What should my next steps be?

[JorgeB]

17 minutes ago, Paul Rockliffe said:

"Metadata corruption detected at xfs_btree_sbloc........"

Check filesystem.