Unraid has disabled one of my data disks, and it is showing 4 write errors.

 

I use shares to dump my data to unraid, so I don't know if I have written files to the drive since it was disabled, and I don't know when it became disabled.

 

What do I do?  Do I need to buy a replacement drive and have unraid re-build onto it?  Do have have it re-build onto the existing disk?

In the meantime is all my data at risk?

Check your cables, and get a SMART report on the drive. It's likely that you need a new drive. Yes, your data is at risk. You can force unRAID to trust the disk, but don't do that unless you KNOW the disk is good.

How do I run a syslog, and a smart report?

 

 

Unraid has disabled one of my data disks, and it is showing 4 write errors.

 

I use shares to dump my data to unraid, so I don't know if I have written files to the drive since it was disabled, and I don't know when it became disabled.

 

What do I do? 

Post a copy of your system log.  Instructions under "Troubleshooting" in the wiki.

Do I need to buy a replacement drive and have unraid re-build onto it?

Depends.  If it is simply a loose cable to the existing drive, it can be corrected and the data re-constructed onto the existing drive.

  Do have have it re-build onto the existing disk?

No, it can be onto a replacement drive.

In the meantime is all my data at risk?

It is at risk of a second concurrent disk failure.

 

DO NOT use the "restore" button to set a new disk configuration, DO NOT use the init config command to set a new disk configuration.  Both of those would set a new disk configuration WITHOUT the failed disk or its data.  Either of those will cause you to throw away your own data.

 

Do

Post a syslog, before you stop the array a down power down to check the connections.

Then, when you power up, get a SMART report on the drive.  If it responds, odds are you'll be able to re-construct onto it.

 

To re-construct onto it,

Stop the array

Un-Assign the disk on the devices page

Start the array with it un-assigned

Stop the array once more

Re-Assign the disk on the devices page

Start the array by pressing "Start"

Yes, your data is at risk. You can force unRAID to trust the disk, but don't do that unless you KNOW the disk is good.

DO NOT FORCE THE ARRAY to trust the disk unless you want to lose all the data you might have written to it since it failed.  You've been warned.

 

Joe L

Joe, thank you.

 

I am working on the syslog and smart reports now.

 

 

If I do have to replace the drive, I would want to buy a 2tb drive, but my parity is only 1.5tb.  Could I replace the parity drive with the 2tb, then re-build the disabled 1.5tb with the good 1.5tb that used to be parity?  Or would this not work?  What would the steps be for this?

I have attached a partial syslog file, I had to cut out a bunch of older entries due to the file size restriction for posting.

 

partial_syslog-2010-10-20.txt

I have attached a partial syslog file, I had to cut out a bunch of older entries due to the file size restriction for posting.

 

 

Zip the entire syslog file and post it or use pastebin.  It will compress a lot because of the repeated content.  The older stuff you cut off was where the write failures occurred.

The disk has been out of service for a while, so I'd just stop the array, unassign it, start the array without it assigned, stop the array once more, re-assign it, then start the array (by pressing "Start")  and let it re-construct onto it.  (You might need to check the box under "Start" to enable it)

 

If the cabling was loose, hopefully you fixed it.  The SMART report looked pretty normal.  There are 14 re-allocated sectors, but that is not too bad.

 

Let the drive re-construct.

I have attched the full syslog, zipped.

The server rebooted due to power outage in the last couple weeks, so if the drive errors were before that, the information will be gone.

 

 

I ran the smart report against the wrong drive, forgot the change the sd*

I ran the short smart test, then ran the smart report against the correct drive, it is attached.

 

Let me know if this changes your recommendations above, thanks Joe!

 

syslog-2010-10-20.zip

smartNov2010.txt

If I do have to buy a hard drive to replace the disabled one, I would like to buy a 2TB.

 

But my parity drive is only 1.5T.

 

Is there any issue with:

1.  I have a disabled data drive

2.  Pull 1.5T Parity drive out

3.  Re-build parity on new 2TB drive

4.  Pull disabled drive out, replace with old 1.5T parity

5.  Rebuild 1.5T data

 

Would I follow the "swap-disable" instructions from the unraid wiki:

  1. Stop the array.

  2. Power down the unit.

  3. Replace the parity hard disk with a new bigger one.

  4. Replace the failed hard disk with you old parity disk.

  5. Power up the unit.

  6. Start the array.

If you have a disabled disk then you must use the swap-disabled method.  You would not be able to follow the first set of steps as it would be impossible to calculate parity with a disabled disk.

Joe, did you have a chance to review my full (zipped) syslog, and the proper smart report?  The first smart report I posted was for my parity drive, not my disabled data drive.

 

Once you review please let me know if I should re-construct or buy new.

 

Thank you Joe, greatly appreciated.

 

Rory

bump

14 reallocated sectors is trivial if the number does not increase with repeated parity checks.  Typically there is a pool of several thousand spare.

 

The actual failure must have been in one of the prior syslogs.  (They are rotated out once they grow to a certain size.) You might look to see if you have a syslog.1 or syslog.2 file in /var/log.  The older of the two will probably have the initial error.

 

I see no reason you should not attempt to re-construct the failed drive.

Thank you for the reply Joe.

 

I ended up buying a 2TB today, and am currently performing the swap-disable.

Unraid picked up the swap-disable, and is copying parity as I type, 21% done.

 

2 questions:

1.  So now I have a 1.5TB drive that might still be good, what tests/programs can I run against it to find out if its any good still?  Preferably windows-based tests/programs.

2.  I bought a 2TB Seagate 5900 with the cc34 firmware, and tried to upgrade the 35, but after re-booting into the firmware update program, it found the hard drive and c34 firmware, but flashed a message that it could not or would not update, then booted windows back up.  Anything I should be concerned with?

1.  So now I have a 1.5TB drive that might still be good, what tests/programs can I run against it to find out if its any good still?  Preferably windows-based tests/programs.

I like SpinRite but it is DOS based not Windows based.  It is also not free so maybe someone else can suggest a free alternative.

Looking through Seagate's website they have SeaTools which they want customers to run before sending back drives under warranty, so I will start there.

 

Thanks everyone, the server re-copied parity successfully, and is not re-building the data drive, should be ready in about 7 hours.