One thing that maybe needs to be addressed if its not too late is shutting down safely. We have already seen a couple of instances of automatic correcting parity checks happening here. Does OP know that you must stop the array before shutting down?

 

It might also be worthwhile to go into Settings - Disk Settings and set Enable auto start to No.

 

P. S. if OP has already shutdown, safely or otherwise, he could manually edit disk.cfg to disable auto start.

 

Well, in regards to "safely shutdown".

I think that if the array is not stopped first, the shutdown or reboot buttons are greyed-out, so it is impossible to shut down without stopping it.

But giving it some thought...."the array must be stopped first" I used "shutdown -h now" to shutdown the array when it was invisible in my network...taken from http://lime-technology.com/wiki/index.php?title=Configuration_Tutorial#Setting_a_static_IP_address

I do not think that command will stop it and then shut it down....therefore, not a safe shutdown. Am I right?

 

I am not ready yet to start with PHASE 3. I am still traumatized  lol

You are correct.  The "shutdown"command does NOT cleanly stop the array.  You must STOP the array first, then shutdown.

Instruction on stopping the array from the command line are here:

http://lime-technology.com/wiki/index.php/Console#To_cleanly_Stop_the_array_from_the_command_line

 

razr: Here's probably the easiest way to get on with things.

 

Let's do this:

1. Stop array, got to Utils/New Config and run that utility.  This will completely clear out the disk config.

2. Go back to Main and assign disk1, disk2, disk3, and cache the way you had them.  DO NOT assign parity, just leave unassigned.

3. Now Start the array.  You should be able to browse the disk shares on your network.

 

Take a look at disk1 and let's see how much got recovered.  Depending on how messed up this file system is will dictate probably how to proceed to restore your server back to something useful again.

 

OK guys, I am ready. I already copied everything from disk1 to an external HD.

Disks are in the array..shut down...

Should I still do as TOM proposes here??

 

Please let me know...thank you.

razr: Here's probably the easiest way to get on with things.

 

Let's do this:

1. Stop array, got to Utils/New Config and run that utility.  This will completely clear out the disk config.

2. Go back to Main and assign disk1, disk2, disk3, and cache the way you had them.  DO NOT assign parity, just leave unassigned.

3. Now Start the array.  You should be able to browse the disk shares on your network.

 

Take a look at disk1 and let's see how much got recovered.  Depending on how messed up this file system is will dictate probably how to proceed to restore your server back to something useful again.

 

OK guys, I am ready. I already copied everything from disk1 to an external HD.

Disks are in the array..shut down...

Should I still do as TOM proposes here??

 

Please let me know...thank you.

 

All or most of your data was recovered? You have verified?

I would say most, or the most important data. I might find later that I am missing something.

The thing is that it is not like I was not doing back-ups.

I store in unraid :

business folder----it was in sync with a pc

Music-----it was in sync with a pc

Movies----too many movies----impossible to sync

Pictures and family videos---used to be synced, bought a computer a few months ago----forgot to sync--

I looked through the pictures, and I am 90% sure there is nothing missing.

 

I would say we are good to go 

Ideally, I would like to try to reconstruct disk1 (without messing up disks2 and 3)

Yes, rebuild the disk.

Yes, rebuild the disk.

 

Steps please..... I dont want to make mistakes

 

I have not done this 1. Stop array, got to Utils/New Config and run that utility.  This will completely clear out the disk config.

Shall I?

 

SO just assign all disks back to the array in the right order, start the array, format disk1 (or preclear before?)

Start parity check?

 

 

Post a screenshot of unRAID main.

Here you go

Just want to make sure I understand where you are now.

 

Your "drive1" that you corrupted is currently assigned as your cache drive?

Just want to make sure I understand where you are now.

 

Your "drive1" that you corrupted is currently assigned as your cache drive?

 

Ups sorry, wrong pic...

...

Pictures and family videos---used to be synced, bought a computer a few months ago----forgot to sync--

I looked through the pictures, and I am 90% sure there is nothing missing.

...

So do you now have more than one copy of these?

Yes, disk 1 + all important data is on a separate drive. But I would not like to loose the movies that are stored in disks 2 and 3....if possible 

Anyone?

Not able to help now. Too busy a work. Someone else should be able to walk you through the steps to get a smart report on disk2 and, if it looks ok, reestablish the array.

You can Start the array now and let it rebuild Disk1 -- it won't have any impact on disks 2 or 3.

 

Here's the issue:  Your earliest mistake (booting to the wrong flash drive) wiped out part of disk1, due to the automatic parity rebuild that was rebuilding parity on that disk;  and your subsequent actions resulted in an automated parity check on the correct 4TB disk that was rebuilding parity on it from the WRONG data on disk1.    Since then, you also made some changes to disk1 by running reiserfsck.

 

So ... whether or not a rebuild will result in any more recoverable data depends on WHICH disk -- disk1 or the parity disk -- had the most data messed up by those two actions.    But it certainly won't hurt to do it.  After you've rebuilt disk1, you'll need to redo the reiserfsck recovery on it ... exactly the same set of recovery steps you did before ... and then see whether or not you can recover more data than the first time.

 

So ... the steps you want to do now are:

(1)  Start the array and let Disk1 rebuild

(2)  Repeat the recover steps with reiserfsck that you did before -- with all of the same precautions.

 

Once you've done that, you're ready to do a New Config, and let a new parity sync be done, so you'll be back to a parity-protected array again.  But that's a couple days away at this point.

 

 

... it is not like I was not doing back-ups.

 

Except, of course, that a backup that's not up-to-date isn't very useful (as you've discovered in the past week)

 

 

Movies----too many movies----impossible to sync

 

You may not consider the movies worth the couple hundred $$ it would cost to back them up; but it's certainly not "impossible".    I have 40+TB of data on 2 UnRAID servers ... ALL of it is ALWAYS backed up.    I know folks with over 100TB of data that they keep backed up.

 

 

Pictures and family videos---used to be synced, bought a computer a few months ago----forgot to sync--

 

Fortunately you recovered most (or all) of these.  But as you've learned all-too-well, if you don't back something up, you can lose it.  I always tell folks they don't really have to backup their data ... only that data they're going to be upset about losing.    Always assume your system is going to catastrophically fail at midnight ... and be sure anything you're going to be upset about losing that night is backed up      Never skip a backup ... doing so is like deciding to take ONE trip to the store without fastening your seat belt -- Murphy's law almost guarantees that's when you'll have an accident !!

 

 

Ideally, I would like to try to reconstruct disk1 (without messing up disks2 and 3)

 

Good idea -- it will, as I outlined above, let you then attempt recovery on Disk1 again, and you may be able to recover a bit more data from it.  [i suggested doing this a few days ago -- in Reply #125]

 

(1)  Start the array and let Disk1 rebuild

(2)  Repeat the recover steps with reiserfsck that you did before -- with all of the same precautions.

 

Once you've done that, you're ready to do a New Config, and let a new parity sync be done, so you'll be back to a parity-protected array again.

If the rebuild completes correctly, and all subsequent reiserfsck actions are done on the md* device, the array will already be parity protected, and there is no need to do a new config. A non-correcting parity check to be absolutely positive that the rebuild was successful would be all that was needed to be sure the array is parity protected.

 

Any time a drive is rebuilt from parity a non-correcting parity check is the way to go, just in case the rebuild wasn't good. You trusted parity to be accurate to rebuild the drive, you surely don't want to "correct" parity if the freshly rebuilt drive fails.

Agree r.e. doing a non-correcting check after a rebuild.  That's one of the rare times I recommend non-correcting check.

 

I wasn't aware you could do the reiserfsck corrections and maintain parity ... but that's definitely a good idea.

 

So ... do the rebuild;  follow with a non-correcting parity check;  and then, assuming that shows zero sync errors, do your reiserfsck corrections and you'll have a parity-protected array again.

 

Note that you'll want to look at Disk1 at that point and see if it's got more or less of the good data you had already recovered => if it's less, you'll want to delete anything that's not correct on it and copy it from your previously recovered files.

 

I wasn't aware you could do the reiserfsck corrections and maintain parity

If you do the corrections on the md* device, parity is maintained. If you write to the sd*1 device, parity is not maintained. Never operate directly on the /dev/sd*1 devices unless you are a linux guru, or are told the exact syntax to use in your situation by a guru.

Razr - how's it going? I have been swamped at work but can help this weekend if your array is not back assembled.

Hello everyone.

My apologies, I have been really busy at work and hadn't had time to check back with you.

After I recovered and saved to an external drive most, if not all the data running reiserfsck, I added the disk back to the array, formatted, ran parity, and copied the data back to the array. Ran parity again and all is been good.

Thank you all for your help.

Excellent!