Hi All,

I've made a pretty awful mistake, and was wondering if you could help me. I had a PSU let loose the magic smoke and fry all of my disks, but managed to get the two parity drives fixed so that I could rebuild from them.

Unfortunately, I selected "New Config" and now get a message stating that the parity drives will be erased on starting the array. Is there absolutely anything I can do to revert that change, so that I can use the new drives to get my data back? I have a copy of the diags from before the incident - and one from today (both attached). The only thing I can think of at this point is to move one of the old parity disks into a data disk slot, tick 'parity is valid' and 'maintenance mode' and see if it can rebuild. Would I be right in thinking so? My other thought would be to copy over the super.dat file from the original diags.

Please could you let me know? I read the docs in a bit of a rush and made what looks to be a massive mistake, so anything I can do to fix it would be much appreciated.

Thanks,

JB

n1-lon-diagnostics-20251226-1322.zipn1-lon-diagnostics-20260121-2057.zip

25 minutes ago, jpants94 said:

I had a PSU let loose the magic smoke and fry all of my disks

Did you try to use a modular PSU cable that wasn't supplied with that particular PSU? There is no standard pinout for modular PSU cables.

27 minutes ago, jpants94 said:

but managed to get the two parity drives fixed so that I could rebuild from them.

Parity by itself can rebuild nothing. The parity algorithm reads parity plus all other disks to calculate the data to rebuild a missing disk. Dual parity can rebuild 2 disks by reading all other disks. You are missing more than 2 disks so parity can't help at all. Parity contains none of your data.

Do you have backups? Parity is not a substitute.

Hi,

Thanks for this. I could get the rest of the data disks repaired, if that would work - then I might be able to rebuild if I'm understanding what you're saying?

Re: backups - I've backed up the really critical stuff, but getting back the other stuff would be very helpful. If having the data disks back would work, then I could do that.

Thanks!

Re: PSU - no, it was a non-modular one - it gave up the ghost mid-day pretty spectacularly, complete with a bang.

You need at least 2 of the data disks with your 2 parity disks to rebuild the other 2 data disks.

OK, so if I got those repaired - using the files in the diags I attached, would I be able to restore everything? I still have the drives and could have them fixed up.

Thanks,

Jordan

Often the data is still on the platters and you just need to replace the circuit board using one from an identical model.

1 minute ago, trurl said:

You need at least 2 of the data disks with your 2 parity disks to rebuild the other 2 data disks.

Assuming there are no other problems we don't know about. And sometimes other problems can be worked with.

There are some specific steps you would need to follow to have any chance. We can get into those details if you get 2 data disks repaired. Whatever you do, don't start the array without further advice.

Thanks for the advice @trurl - based on that, I've gone and got the disks repaired and in situ (minus the cache/non-pool disks). Please could you let me know what I need to try to have any chance of bringing things back up?

As a result of all of this I've invested in a separate backup, which hopefully will minimise the likelihood of going through this ever again!

Do you have 2 disks to rebuild to?

I have all of the original disks together, with the data present. I just need to restore the config somehow - it should rebuild at that point, right? The only missing disks are cache/non-array disks. Forgive me if I'm not correct in assuming so, but I thought that with all of the original drives and data it should just be a case of restoring config and starting the array?

I thought you were paying to get 2 disks repaired so you could rebuild the other 2.

You just need to New Config. Parity might be valid, but you can let New Config rebuild it or check the box before starting the array. You should probably at least run a correcting check.

https://docs.unraid.net/unraid-os/using-unraid-to/manage-storage/array/array-health-and-maintenance/#reset-the-array-configuration

If his pool disks are functional, shouldn't he be able to drop them into a non pool or mount them with Unassigned Devices, and pull all the data off them into a new pool? Seems like les risk that trying to rewrite the cache.

Following the link @trurl posted, I was able to rename the super.old to super.dat, which has brought the disks into an array (as they were) in the correct order. For safety's sake, would you recommend using Maintenance Mode to run any checks that might need to be done here? They appear to be in the right order, but I don't want to run any more risks than I have to.

Also - just to say - thank you very much for your input on this. It's made things a lot less scary than they otherwise might be!

It would be better to start the array in Normal mode so you can see if the filesystems mount and you can inspect their data. In Maintenance mode, you can't access anything since nothing is mounted.

Thanks @trurl - I'll do that. At the moment I've followed @Arbadacarba's suggestion and begun pulling off anything I'd be scared to lose, so I'll kick it over into normal mode once done and update with progress.

Thanks again both!