Version 4.7

 

Noticed my server hanging up one day and restarted it. When it came back online I noticed that there was an orange dot for the parity, and red for disk 5. In umenu, the parity disk shows up as invalid and disk 5 shows up as disk_dsbl. I'm guessing disk 5 is kaput and needs to be swapped out? I've attached the log below

syslog-2013-11-19_1.txt

Oh, yes - disk5 with a couple of read errors.

Wouldn't be that bad if there wasn't the invalid parity also.

This is probably the worst case scenario.

I'm very interested in the troubleshooting options now.

You'd better open an official support ticket with limetech.

 

Did you save the log file before restarting the hanging server?

 

Thinking loud of a solution:

You probably have to set parity as valid and do a rebuild of disk5.

After that you check if the content of disk5 is good or corrupted.

A set of checksums would come in handy now.

I would definitely NOT "set parity as valid" ==> that would rebuild disk5 with known errors !!

 

I'd simply replace both the parity drive and disk5, and redo the parity sync so you have a good, protected array; then copy the data that was on disk5 back to the array from your backups.

 

If you don't have good backups, you can try copying the data from disk5 from another computer to the array AFTER you've reconstituted a good array.  If you're using a Windows computer for this, you can use the free Linux Reader to read the data.

http://www.diskinternals.com/linux-reader/

 

Another possibility is to use Reiserfsck to see if you can restore the contents of disk5 => but that's not as good as getting "known good" data from your backups.

 

Yes Gary, backups...

Let's presuppose there are none.

 

The question is, how deep can parity be corrupted?

If you do both - rebuild disk5 with known corrupted parity and copy data from the defective disk5 you might be able to recover all of your data.

With the "externally recovered" disk5 you can compare the restored disk5 and you might find the missing files from the "externally recovered" disk5 in a good shape on the rebuild.

 

You won't loose more if you rebuild on a "known false" parity.

 

 

 

 

It's certainly true you could do a New Config and tell UnRAID that parity is good;  but the first thing UnRAID does after you start an array like that is a parity check ... so you'll need to stop that check as quickly as possible if you want to rebuild disk5 based on whatever data happens to be good on the parity disk.

 

I agree that if you rebuild onto a new disk; and then compare that with what data you can read from the failed disk5, that any matches are likely good.    What's far less clear is what you could/should do with mismatches.  Assuming you have neither backups nor checksums of your files;  you'll likely simply lose that data.

 

It's certainly true you could do a New Config and tell UnRAID that parity is good;  but the first thing UnRAID does after you start an array like that is a parity check ... so you'll need to stop that check as quickly as possible if you want to rebuild disk5 based on whatever data happens to be good on the parity disk.

Not a new config!

Just set the parity valid. There can't be a parity check with a "missing" disk5 that has to be rebuild.

But setting the parity valid is probably not possible with the options we have in the GUI.

 

What's far less clear is what you could/should do with mismatches.

You will have to check them by hand if they're good or not.

Just to be clear, these are the steps I need to do (i do not have good backups):

 

1. Remove both drives.

2. Add new parity drive and a new disk 5.

3. Start parity sync

4. Attempt to copy files from "disabled" disk 5 data back on from another computer?

 

Is it possible that there may be a cabling issue, or does the log indicate the physical connections are working, but the drives are indeed defective? And how would I go about starting a ticket with lime-tech? I only see a contact form on the main page. Is that it?

[email protected]

 

Point at this thread.

Ok I tried to add 2 drives at once, and I get the error:

 

"Stopped. Invalid configuration. Too many wrong and/or missing disks!"

 

I also see the old drive in italicized font underneath the new parity drive. I can assign the data disk and parity drives but I can't do anything with the array. So i tried disconnecting the new parity drive, and hooking the old one back up. I have an orange orb next to the old parity drive and the new disk 5 data drive. I tell it to format the unformated drive (disk 5) but then it doesn't do anything and im stuck on this screen with the attached picture below

 

Ok I tried to add 2 drives at once, and I get the error:

 

"Stopped. Invalid configuration. Too many wrong and/or missing disks!"

 

I also see the old drive in italicized font underneath the new parity drive. I can assign the data disk and parity drives but I can't do anything with the array. So i tried disconnecting the new parity drive, and hooking the old one back up. I have an orange orb next to the old parity drive and the new disk 5 data drive. I tell it to format the unformated drive (disk 5) but then it doesn't do anything and im stuck on this screen with the attached picture below

 

NOT the right thing to do unless you've given up on reconstructing your failed drive #5.  IF your plan is to give up on a reconstruction (since you don't have good parity to do this with), and you simply try to copy what you can from the old, failed, disk;  then you need to do a NEW CONFIG and set the drives as you want -- THEN it will start fine and do a new parity sync.

 

If, however, you want to try to force the bad parity to be used anyway to reconstruct disk 5 (which I do not recommend; but that's what Fireball3 suggested and you implied you might want to do);  then what you need to do is a New Config with all the original assignments -- checking the "Parity is Good" box; and then IMMEDIATELY stop the parity check that starts when you Start the array.    Then Stop the array;  unassign disk5;  Start the array (so it shows as missing); and then Stop the array and assign a new disk to the disk5 slot; and then Start the array again ... it will now do a rebuild onto the new disk5.    I DO NOT recommend this process ... you'll end up with a "known bad" disk5 => and since you don't have backups to compare it against, I'm not sure what you gain from it.

 

I'd just do what I had suggested -- do a New Config with both a new disk5 and a new parity drive;  let the new parity sync be computed;  and then, with a known-good array, just copy whatever data you can read from the old disk to the good array.

 

Ok I tried to add 2 drives at once, and I get the error:

 

"Stopped. Invalid configuration. Too many wrong and/or missing disks!"

 

I also see the old drive in italicized font underneath the new parity drive. I can assign the data disk and parity drives but I can't do anything with the array. So i tried disconnecting the new parity drive, and hooking the old one back up. I have an orange orb next to the old parity drive and the new disk 5 data drive. I tell it to format the unformated drive (disk 5) but then it doesn't do anything and im stuck on this screen with the attached picture below

 

NOT the right thing to do unless you've given up on reconstructing your failed drive #5.  IF your plan is to give up on a reconstruction (since you don't have good parity to do this with), and you simply try to copy what you can from the old, failed, disk;  then you need to do a NEW CONFIG and set the drives as you want -- THEN it will start fine and do a new parity sync.

 

If, however, you want to try to force the bad parity to be used anyway to reconstruct disk 5 (which I do not recommend; but that's what Fireball3 suggested and you implied you might want to do);  then what you need to do is a New Config with all the original assignments -- checking the "Parity is Good" box; and then IMMEDIATELY stop the parity check that starts when you Start the array.    Then Stop the array;  unassign disk5;  Start the array (so it shows as missing); and then Stop the array and assign a new disk to the disk5 slot; and then Start the array again ... it will now do a rebuild onto the new disk5.    I DO NOT recommend this process ... you'll end up with a "known bad" disk5 => and since you don't have backups to compare it against, I'm not sure what you gain from it.

 

I'd just do what I had suggested -- do a New Config with both a new disk5 and a new parity drive;  let the new parity sync be computed;  and then, with a known-good array, just copy whatever data you can read from the old disk to the good array.

 

Ok I do want to just scrap Disk 5 and the Parity Drive, but I'm unsure about how to go about starting a new configuration. Do I physically replace both drives at once? Before starting the new configuration?

See here: http://lime-technology.com/forum/index.php?topic=30009.0