Data loss after upg...
 
Notifications
Clear all

Data loss after upgrade

4 Posts
2 Users
0 Likes
389 Views
(@chocoholic)
Posts: 7
Member
Topic starter
 

I'm not sure if this is the best forum for this mess. Yes or no, I hope you'll bear with me:

Rebuilding 42TB work drive

 

1) We've an 18TB RAID5 (12.8TB in use, large video source files) and a 42TB RAID5 (17.45TB in use, tons of FCP libraries, our work drive), each of them built using paired OWC 2-slot toaster docks (Thunderbolt 2), each RAID5 set connected to its own Tbolt bus on the iMac Pro. Both have run happy as clams, if clams could run, for months.

2) Did the upgrade from Mojave to Catalina.

3) And, perhaps but I hope not significantly, both RAID5s and a Time Machine were mounted during the upgrade (not my doing — I’m on the other side of the Atlantic!).

4) Some really dumb things were done during the upgrade which forced multiple unstable restarts. Only then were the the drives unmounted and disconnected. Ultimately, the Mac needing to be wiped clean and have Big Sur installed, then data restored from Time Machine.

5) Post-upgrade to Big Sur, my colleague doesn’t recall being asked to move from SoftRAID 5.8.4 to 6.0.5 (remember, I’m on the other side of the Atlantic). He didn’t do it.

6) 18TB RAID5 post-Big Sur install: once mounted, he found there was some data loss, we now have dozens of folders in different path levels that are empty, while the folders themselves are there. Also, some folders now have permission-denied symbols, we don’t know whether they’re empty, and attempts to change permissions have failed. Disk Utility didn't like what it saw on the first run, but the second run worked. (Unfortunately, my colleague didn't capture the messages from the first run.) Ran Disk Warrior after Disk Utility (wouldn't have been my approach), it gives a clean bill of health.

7) 42TB RAID5 post-Big Sur install: the FCP libraries kept crashing the program, then the volume wouldn’t mount, then we asked SoftRAID to mount it and it saw a drive as out of sync and has been rebuilding for approx. 24hrs.

Questions:

a) There’s no “time remaining” indication on the rebuilding, oughtn’t there be?

b) As we can’t yet know what we may find amiss on the 42TB until it finishes, the only real question I can pose is re: the 18TB:  Do you have any thoughts as to why we lost data, and how might we go about retrieving it? Nothing’s been written to it other than whatever directory work has been done. I’ve found Disk Drill useful for a file here and there in the past, but don’t know whether it’s recommended for a situation like this.

 

Thank you so much!!

Rebuilding 42TB work drive

 

 
Posted : 18/06/2021 11:38 am
(@softraid-support)
Posts: 8044
Member Admin
 

Do you have Disk Warrior? I would be cautious before "replacing" the directory, but I would certainly scan the volume and see if it can get the missing files back. there is a preview option at the end, and I highly encourage you to browse the pseudo volume before clicking replace (it is undoable).

This is a Finder type issue, the RAID can handle what happened. So I do not know what caused the file damages.

The volume is not rebuilding, as it is not mounted. I think Disk Warrior woudl have not had much problem before all the FCP work, I am a little concerned now.

 

Disk Drill may be able to recover files, but I think first, Disk Warrior is a better attempt.

 

 
Posted : 18/06/2021 1:50 pm
(@chocoholic)
Posts: 7
Member
Topic starter
 

I'm a little confused: why would the Volume tile say it's rebuilding if, as you say, it isn't rebuilding?

When we try to remount it, it doesn't appear on the Desktop.

When we restart, we get this:

mounting message

…but even then the volume doesn't appear on the Desktop.

Thus far, nothing we've done rebuilding-wise has resulted in the disk access indicators flickering.

DiskWarrior did find the volume, and fixed thousands of entries, but we haven't replaced the directory yet, pending closer examination.

 

 
Posted : 18/06/2021 3:34 pm
(@softraid-support)
Posts: 8044
Member Admin
 

@chocoholic

If you run SoftRAID again, you may see a disk is missing. If it is very slow to power up, you may get this, but it will eventually show up. SoftRAID waits a 30 seconds with the other disks, before putting that message up.

Rebulding means it is in rebuild mode, however the volume must mount.

SoftRAID publishes the disk for mounting, but the directory is probably damaged, so it does not actually mount. Glad you are examing the volume. I would open the Preview, and look for sample files, open a few, and examine the "rescued items" folder or damaged items if any. Those are the biggest risks.

Hopefully Disk Warrior found everything!

Once you replace the directory, the volume can rebuild as normal.

Last, I would change optimization to workstation, so it rebuilds faster (higher priority)

 
Posted : 18/06/2021 3:44 pm
Share:
close
open