I had my usb drive fail, I copied over a 2 month old config backup.  I had one drive in the array before during the backup that is no longer there I removed it and added 3 others in its place.  All of my drives are labeled in the das Disk 1 , disk 2 etc and this corresponds with my unraid config, so I know exactly which drive is assigned what.  I matched them up accordingly and click start the array, then proceed, then …. Nothing at all happens.   I really don’t want to mess things up any worse than they are now, so can someone please advise me on what my next steps should be? I have dual parity and both of those are physically labeled as well, so at least there is no guessing which drive was which.  I’m just not sure where I’m going wrong now.  
 

thanks in advance. 

Please post the diagnostics.

2 hours ago, JorgeB said:

Please post the diagnostics.

Sorry about that I just wanted to get the info typed out while it was fresh.  
 

attached here.  Thanks!

oswald-diagnostics-20231215-0638.zip

I don't see any attempt to start the array, or is the problem that you cannot start it?

 

Post a screenshot of main.

27 minutes ago, JorgeB said:

I don't see any attempt to start the array, or is the problem that you cannot start it?

 

Post a screenshot of main.

I just noticed both cache say "Wrong"  the backup I have must have had the old SSD cache and in those 2 months I switched to 2x NVME Cache. 

I'm thinking I need to do something with the New Config I read in the manual but I'm a little gun shy and would like some confirmation first.

 

 

Unassign the pool devices, if all the array disks are all correctly assigned check "parity is already valid" and start the array, then stop the array, re-assign the pool devices, start the array and post new diags.

9 minutes ago, JorgeB said:

Unassign the pool devices, if all the array disks are all correctly assigned check "parity is already valid" and start the array, then stop the array, re-assign the pool devices, start the array and post new diags.

Thanks I’ll try that in a few minutes.  I assume by pool devices that means the two cache drives aka cache pool correct?   Really appreciate the help.  

Correct

21 minutes ago, JorgeB said:

Unassign the pool devices, if all the array disks are all correctly assigned check "parity is already valid" and start the array, then stop the array, re-assign the pool devices, start the array and post new diags.

Does this still apply if my config from the backup 2 months ago has different drive assignments.  I believe there are 3 data drives that weren’t there on the 2 month old config, and likely one drive missing that was on the old

config.  

If you did a new config that's not a problem, though strange the pool is looking for the old devices, did you do a new config?

17 minutes ago, JorgeB said:

If you did a new config that's not a problem, though strange the pool is looking for the old devices, did you do a new config?

No I didn't, should I?  I was thinking new config but I figured I'm better off taking this slow and accepting a day or two downtime vs trying my gut instinct and ending up in a real mess.  Truly appreciate the responses!   I think I did enough to avoid major disaster by labeling well, but in the future I know I'll need to have up to date backups of my usb (I do) but actually export them off the array more often (I wasn't).

34 minutes ago, serversonfire said:

up to date backups of my usb (I do) but actually export them off the array more often (I wasn't).

You mean you save your flash backups on the array? You need to rethink that.

 

Unraid Connect plugin will store your flash backups on cloud.

 

The screenshot seems to imply that the array assignments in that old backup were correct. Are you sure that backup was before the array changes?

8 minutes ago, trurl said:

You mean you save your flash backups on the array? You need to rethink that.

 

Unraid Connect plugin will store your flash backups on cloud.

 

The screenshot seems to imply that the array assignments in that old backup were correct. Are you sure that backup was before the array changes?

Correct just as part of the automated app data backups it takes flash backups too.  My big issue with that was I had an old one where there have been multiple changes to the drives since.  So that is why I'm thinking I need to do new config since I know which drives are which at least.

 

 

"The screenshot seems to imply that the array assignments in that old backup were correct. Are you sure that backup was before the array changes?"  As for this no to the data drives.  The backup could have happened after the data drives were already changed.  However the backup definitely predates the upgrade from SSD cache to NVME cache.  Perhaps all I need to do is unassign both cache drives and then start the array, then re-add the cache drives somehow?  My concern with that is that my appdata lives on the cache drive.  I'm just being very cautious because I'd rather not have to restore triple digits of TBs of data or have data loss etc. Slow and steady recovery since I'm not 100% confident in much of it.

 

 

FOR SURE keeping the backups only on the array is stupid, I just didn't realize what I was doing by not copying over a new one after making those changes.  I also run offsite backups of the array so I could have recovered the files from there, but I do that manually and haven't run it since the 2 months ago because the offsite server needs some expansion and funds are tight.  

Definitely some lessons to be learned from this, overall I think I did a decent job but overlooked some critical things and will certainly address them all going forward once recovered.  I think a policy of new usb backups every 1st of the month as well as any time drive changes happen would be key.  Obviously I will make sure to store them off the array as well.  The offsite backups were why I thought that wasn't the worst idea, but that is also assuming the frequency of the backups increases.    LOTS to address after this is resolved.  I'll consider the plugin but I was trying to avoid being cloud dependent.  I've heard concerns of potentially sensitive data being stored unencrypted in the Connect Plugin and I didn't look into that at all to find out how valid those concerns are, not yet anyway.

Edited December 15, 20232 yr by serversonfire

If you know the assignments, suggest you do a new config and re-assign all the devices, including the pool.

1 minute ago, JorgeB said:

If you know the assignments, suggest you do a new config and re-assign all the devices, including the pool.

 

Thank you, sorry I am so cautious about it.  Does cache 1 vs cache 2 matter at all?  Those are the only 2 I don't have labeled since they are on the motherboard.  I have them setup as a mirror setup.  

 

I will try that as soon as I can get home this afternoon.

16 minutes ago, serversonfire said:

Does cache 1 vs cache 2 matter at all?

Nope.

8 hours ago, JorgeB said:

Nope.

 

Thank you very much, got me back up and running.  Syncing parity now.  Had some weird issues with a few containers not able to update due to missing templates, but I was able to restore those from backup and then update successfully. 

I seem to be ok now.  Anything to do with parity besides let the sync finish?  Thank you again you were extremely helpful and I appreciate you guiding me through my mess. 

Now I need to do a proper USB backup of the new drive and replace it with a more reliable USB 2.0 drive in the near future.    I'll look into the plugin as well.

8 hours ago, serversonfire said:

Anything to do with parity besides let the sync finish?

That should be it, you could have selected "parity is already valid", assuming it was.