Hello,

 

I have this weird issue with unRaid since yesterday. I wanted to delete a specific folder but then unRaid instantly crashes.

While investigating the issue I could pinpoint the problem to one specific HDD. I don't know if it affects other files on that HDD since I only tried the specific folder I want to delete.

The crashes are consistent.

After each crash Check Parity runs and finds many errors on that HDD, but when it finiches it says the Parity is fine.

I also tried to run a File System check on that specific HDD which led to another instant crash.

What do you suggest I should do?

More than I want to delete this folder I need to know the rest of the data on this HDD is not corrupted, or worse, that all my data could be corrupted.

 

Thank you

 

P.S: I run 5.0.6 Pro

Need zipped syslog and smart report for the problem disk. Screenshot of the main page of the GUI might be helpful as well.

 

A hardware list might be helpful as well. PSU model, motherboard, HBA's, list of all HDD's in the system, etc.

 

What do you mean by crash? Do you get a message on the console? Does the tower reboot itself? Describe the crash itself.

Need zipped syslog and smart report for the problem disk. Screenshot of the main page of the GUI might be helpful as well.

 

A hardware list might be helpful as well. PSU model, motherboard, HBA's, list of all HDD's in the system, etc.

 

What do you mean by crash? Do you get a message on the console? Does the tower reboot itself? Describe the crash itself.

 

When it comes to Linux I am a real newbie.

I also run Unmenu on top. How do I export the info you need to files? Or do I just copy/paste from the screens I see in Unmenu?

 

By crash I mean a real crash. Suddenly the NAS is unavailable on the network. Stalled. Only a long press on the Power button to shut down and boot is helping.

How do I export the info you need to files?

http://lime-technology.com/forum/index.php?topic=9880.0

Here are the Syslog and Smart report on the problematic HDD.

Syslog.txt

Smart_report_on_Disk_7.txt

Forgot the main screen capture

More info.

I could rename the folder I want to delete but even when I tried deleting it from a Telnet session (as opposed to Windows) Unraid crashed/stalled. It is now unresponsive which means another hard reboot from the Power button.

Very weird.

Here are the Syslog and Smart report on the problematic HDD.

That disk looks very sick!

 

It has a lot of reallocated sectors.  This is not necessarily an issue as long as the value stays constant, but is often an indication that the disk is not in perfect condition.

 

I am extremely surprised that the disk has not been 'red-balled' by unRAID.    You may find that this happens if you try stopping/restarting the array.

 

More worrying are the pending sectors and uncorrectable errors.    These indicate that there are parts of the disk that cannot be read correctly.  It is possible that completely wiping the disk might clear these conditions but I would not trust the disk.  In its current state with sectors not being read correctly it would stop any other disk being rebuilt successfully if it failed.

 

The first recommended action should be to copy all the data that you can from that disk to another one to secure it.  Then I would suggest that as soon as possible you should get a replacement disk and let unRAID rebuild the contents onto the new disk.

I understand. I would like to copy all the data to another hard drive then I wish to wipe the disc and rebuild it with the use of parity.  How do I do that?

I understand. I would like to copy all the data to another hard drive then I wish to wipe the disc and rebuild it with the use of parity.  How do I do that?

The rebuild process is:

• Stop the array
  
• Change the disk to unassigned
  
• Start the array.  It should now show as missing.  At this time you will still be able to access its data as it will be being emulated by the combination of the other drives plus parity.
  
• Assign the disk back again.  You will be told that starting the array will initiate a rebuild
  
• Start the array to begin the rebuild[/li
  

However if you are trying to rebuild onto the same disk I would think it is extremely likely the rebuild will fail looking at that SMART report.  If that happens you need to have another disk to rebuild onto.

 

I think there is also a good chance of there being file system corruption on that disk (a rebuild will not fix this).  Running a reiserfsck check on the disk afterwards is the way to fix file system corruption.

I understand. I would like to copy all the data to another hard drive then I wish to wipe the disc and rebuild it with the use of parity.  How do I do that?

 

Personally, I would be ordering a new hard disk to replace that one.  Run a couple of preclear cycles on the new disk before using it in your array.  I would also stop writing to the array until you get the new disk installed and rebuilt. 

 

After you get the array working again, you can try to run a preclear cycle on that old disk and see what happens.  With that number of sectors in trouble, I would never put it back in service in a server. 

I am taking all your advice into account and will have this HDD replaced (it is still under warranty).

In the meantime I am copying all of its data to other HDDs in my Unraid, which means that the failing HDD can be wiped out afterwards.

How do I:

- Wipe it clean?

- Run some checks on it?

- Run Parity on the remainder of HDDs without rebuilding the data on the failing HDD, leaving it empty?

 

If only I could delete all the data on this HDD the rebuild via Parity would leave it empty as I want but I cannot delete it as it crashes the Unraid. Is there any other way?

I am taking all your advice into account and will have this HDD replaced (it is still under warranty).

In the meantime I am copying all of its data to other HDDs in my Unraid, which means that the failing HDD can be wiped out afterwards.

How do I:

- Wipe it clean?

- Run some checks on it?

- Run Parity on the remainder of HDDs without rebuilding the data on the failing HDD, leaving it empty?

 

If only I could delete all the data on this HDD the rebuild via Parity would leave it empty as I want but I cannot delete it as it crashes the Unraid. Is there any other way?

 

If you physically remove the drive from the array, The drive will be emulated using the other data drives and the parity drive. 

 

I assume that you want to wipe it clean because you don't want to send a drive in for warranty with data on it.  The easiest way to run the preclear script on it.  The script is available here:

 

      http://lime-technology.com/forum/index.php?topic=2817.0

 

 

(There used to be a number of free programs available on the Internet but I could find any with a quick google search.)  Personally, I won't do anything with this old drive until I had a new one installed and running.  (You are correct in worrying about the data on this drive. 90+% of the files are probably still good!)  For that reason alone, I would hang onto it until I had the new drive working and I could see if there are any problems with the rebuilt.

 

If the preclear runs to completion, that will be nothing but zeros on it.  If the script fails to run to completion, the disk will probably appear to be empty anyway to most people.  (The NSA could probably recover something in almost any case if they ever got their hands on it!  If you are worried about that type of data recovery effort, you should be smashing the disk with a twenty pound sledge hammer.)

To make it less confusing, can I:

 

- Remove the bad HDD from the Raid?

- Then rebuild parity WITHOUT the bad HDD?

- Format the bad HDD and check it OUTSIDE the raid ?

- Only then reinsert the bad HDD to the Raid and run parity again with an empty HDD (formerly bad HDD) ?

To make it less confusing, can I:

 

- Remove the bad HDD from the Raid?

- Then rebuild parity WITHOUT the bad HDD?

- Format the bad HDD and check it OUTSIDE the raid ?

- Only then reinsert the bad HDD to the Raid and run parity again with an empty HDD (formerly bad HDD) ?

 

You do NOT want to touch parity in any way, shape, or form UNTIL you have rebuilt the data from Disk 7 on a new disk!!!!  The instructions by itimpi for rebuilding provide you the means to access all of the data currently on disk 7 using emulation.  Just stop after the third step. 

 

Any type of formatting on the present disk will rend the present data unreadable using conventional tools.  (That data will (probably) still be physically there but the directory (folder) structure will be gone.)    If you want to try to see what is readable on the disk, here are two programs that provide that type of service in a Window environment:

 

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

    http://yareg.akucom.de/

 

I have not personally used either one of them but others on the forum have. 

To make it less confusing, can I:

 

- Remove the bad HDD from the Raid?

- Then rebuild parity WITHOUT the bad HDD?

- Format the bad HDD and check it OUTSIDE the raid ?

- Only then reinsert the bad HDD to the Raid and run parity again with an empty HDD (formerly bad HDD) ?

 

You do NOT want to touch parity in any way, shape, or form UNTIL you have rebuilt the data from Disk 7 on a new disk!!!!  The instructions by itimpi for rebuilding provide you the means to access all of the data currently on disk 7 using emulation.  Just stop after the third step. 

 

Any type of formatting on the present disk will rend the present data unreadable using conventional tools.  (That data will (probably) still be physically there but the directory (folder) structure will be gone.)    If you want to try to see what is readable on the disk, here are two programs that provide that type of service in a Window environment:

 

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

    http://yareg.akucom.de/

 

I have not personally used either one of them but others on the forum have.

 

Maybe I didn't expressed myself clearly.

Since I am copying all the data from Disk7 on another Disk and just don't care about Disk7 anymore. So I can remove it from the raid. But after that I would like the Parity to run anew on the remaining HDDs and validate itself, without Disk7 being replaced.

Is that possible?

Think of it as if I wanted to simply remove a HDD from my raid and rerun parity afterwards.

I believe what you want to do can be done with the "New Config' utility in the 'Utils' tab.

 

EDIT: click on the tab for more information.

 

 

I believe what you want to do can be done with the "New Config' utility in the 'Utils' tab.

 

EDIT: click on the tab for more information.

 

That's exactly it! Thank you.

 

Hello again,

 

This is becoming weird.

I have replaced the failing disk with a new one. Parity has rebuilt the content and validated it. Within the reconstructed content there is this folder that started all this thread.

And still, if I try to delete it, whether it is from Windows, or from a telnet session, from the virtual logical volume or from the actual physical disk where it resides, it invariably crashes my whole server, stalling it.

How can I investigates these files in that folder?

Hello again,

 

This is becoming weird.

I have replaced the failing disk with a new one. Parity has rebuilt the content and validated it. Within the reconstructed content there is this folder that started all this thread.

And still, if I try to delete it, whether it is from Windows, or from a telnet session, from the virtual logical volume or from the actual physical disk where it resides, it invariably crashes my whole server, stalling it.

How can I investigates these files in that folder?

Have you run reiserfsck against the disk since rebuilding it?    If there was any file system corruption then the rebuild would not have fixed that as a rebuild merely reconstructs the disk exactly as is was before the rebuild (rebuild works at the physical sector level).  You HAVE to run reiserfsck to fix file system corruption.

Hello again,

 

This is becoming weird.

I have replaced the failing disk with a new one. Parity has rebuilt the content and validated it. Within the reconstructed content there is this folder that started all this thread.

And still, if I try to delete it, whether it is from Windows, or from a telnet session, from the virtual logical volume or from the actual physical disk where it resides, it invariably crashes my whole server, stalling it.

How can I investigates these files in that folder?

Have you run reiserfsck against the disk since rebuilding it?    If there was any file system corruption then the rebuild would not have fixed that as a rebuild merely reconstructs the disk exactly as is was before the rebuild (rebuild works at the physical sector level).  You HAVE to run reiserfsck to fix file system corruption.

 

Noted.

THX

Now I have a second drive that shows many errors in the console, but when I run reiserfsck --check on that drive there are 0 errors.

What can it be?

Note that parity is valid with that drive included.

Now I have a second drive that shows many errors in the console, but when I run reiserfsck --check on that drive there are 0 errors.

What can it be?

Note that parity is valid with that drive included.

Errors on the main GUI are read or write errors, read if the drive is green, read and write if the drive has been red-balled. There is no mechanism in place that I know of to automatically notify of file system level errors. Parity has no concept of files, it only recreates a drive in total, with whatever file system or errors that may be on the drive when the drive is red balled.

 

I'd post a smart report for the drive that is showing errors, and a zipped current syslog would be helpful as well.