Cannot Restart Array - Disk in parity slot is not biggest.

[BaconDroid]

 

Hi,

 

I've stopped my array before shutting down my server to install a second nvme for cache.

Turning the server back On, and now I can't restart my array.

Unraid says Disk in parity slot is not biggest.

My parity drive (sdg) is the same size has two of my data drive (sdd, sde). (same model etc.)

 

Looking around on the forum, I saw something about HPA, but it doesn't seem to be the issue?

When I do hdparm -N /dev/[hs]d[a-z] every drive except my USB thumb give me HPA is disabled.

 

 

Anybody got an idea? I didn't purchase unraid yet, still on the trial. I don't want this to be a dealbreaker, since it worked fine before.

 

 

Thanks,

 

tower-syslog-20231005-1714.zip

[itimpi]

The message is correct as in the syslog the parity shows as

Oct  5 13:03:57 Tower kernel: mdcmd (1): import 0 sdg 2048 3907017540 0 WDC_WD40EFAX-68JH4N1_WD-WX42D81H76FR
Oct  5 13:03:57 Tower kernel: md: import disk0: (sdg) WDC_WD40EFAX-68JH4N1_WD-WX42D81H76FR size: 3907017540 

whereas disk4 for instance reports as

 

Oct  5 13:03:57 Tower kernel: mdcmd (5): import 4 sde 64 3907018532 0 WDC_WD40EFAX-68JH4N1_WD-WX42D914AEYX
Oct  5 13:03:57 Tower kernel: md: import disk4: (sde) WDC_WD40EFAX-68JH4N1_WD-WX42D914AEYX size: 3907018532 

an HPA on the parity disk would be the obvious suspect but you say you have checked for that so not sure what else to suggest.

[JorgeB]

Post the output of:

fdisk -l /dev/sdg

 

[BaconDroid]

26 minutes ago, JorgeB said:

Post the output of:

fdisk -l /dev/sdg

 

 

Here it is :

[JorgeB]

That's not the Unraid standard partition, it should start on sector 64, not 2048, no idea how you only saw the error now, but you need to delete that partition and re-sync parity so a new one will be created with the correct starting sector.

[BaconDroid]

5 minutes ago, JorgeB said:

That's not the Unraid standard partition, it should start on sector 64, not 2048, no idea how you only saw the error now, but you need to delete that partition and re-sync parity so a new one will be created with the correct starting sector.

 

Is there a way I can easily do that from the web gui? or is there a specific command i should use?

 

I'm still a beginner, i've use OMV for a year before they move to version 6 which was causing me issues. This is why I'm trying Unraid now.

 

[BaconDroid]

I was able to delete the partition using fdisk.

I'm rebooting to see if I can start back the array.

[BaconDroid]

I'm getting Invalid Configuration - Too many wrong and/or missing disks!

Did I break something?

 

[JorgeB]

Post new diags.

[BaconDroid]

Here.

tower-syslog-20231006-1358.zip

[JorgeB]

That's just the syslog, complete diagnostics please.

[BaconDroid]

Sorry about that. Here's the diagnostic.

tower-diagnostics-20231006-1049.zip

[JorgeB]

All disks are being detected with a little more capacity than before, besides parity, did you do anything to the other disks?

[BaconDroid]

 

No, I just added a new ssd.

 

I've also followed your previous comment by delete the partition of the parity drive.

After that, since it still didn't work, I tried to copy the partition from sdf to sdg, so I could get a partition that started at sector 64.

Nothing to the other disks.

 

But yeah, i don't think i did much.

 

Edited October 6, 2023 by BaconDroid

[JorgeB]

This is vary strange, you can try doing a new config to see if the data disks still mount, if they don't it won't cause any more damage, just don't format, also leave parity unassigned for now.

[BaconDroid]

I get this. I've remove the parity drive but now the data drive look unmountable.

[JorgeB]

Not a good sign, post new diags.

[BaconDroid]

Here.

tower-diagnostics-20231006-1235.zip

[BaconDroid]

If there's no solution, I will start anew.

Just would like to make sure it doesn't happen again

[JorgeB]

Assuming the disks were xfs formatted you can try running xfs_repair, but this can make things worse, only do it if you have backups or clone one of the disks and rune it there, e.g.:

 

xfs_repair -v /dev/md1p1

 

2 minutes ago, BaconDroid said:

Just would like to make sure it doesn't happen again

It may well happen since it looks like something changed the partition layout of all your disks.

[BaconDroid]

It's running, will let you when it stops.

[JorgeB]

That won't work, if not using the md device you need to specify the partition, e.g,:

 

xfs_repair -v /dev/sdf1

 

[BaconDroid]

Thanks for the reply, I've launched the right command now.