Recently a couple of 2TB drives started showing up as read-only. I decided to replace them with 3 and 4 TB drives, and then check the old drives and maybe preclear and add them back to the array if they are ok. Today I replaced disk10, a Hitachi 2TB, with a Seagate 4 TB drive. My Parity drive is a 4 TB WD Black drive.

The new Seagate drive was recognised immediately.

On the Main page I try to check the box and start the array, to rebuild the data. However, instead of starting the process, unraid returns me to the same screen again with the array stopped ad infinitum. Is there anything I can do? Any clues as to what might be happening? Syslog attached.

Is unraid rejecting the new disk because it possibly appears to be larger than the parity, for example? (It shows 4Tb for both).

syslog-2013-05-31_part_1.txt

Part two of syslog

syslog-2013-05-31_part_2.txt

In unmenu's MyMain the disk 10 has HPA? next to it, and shows up as Unknown 4T.

In unmenu's MyMain the disk 10 has HPA? next to it, and shows up as Unknown 4T.

When myMain was written, 4TB drives did not exist, and therefore it just does not know the expected size of a 4TB drive.  You can ignore that HPA? as it is just a warning.

 

Many have had issues when performing disk replacements with SimpleFeatures active.  Disable it and then try again.  At least one of its plugin add-ons, if not it itself, have caused basic features in unRAID to fail in the past.  (Not sure if they were ever fixed).  In one case, it broke the

sgdisk

command. (needed to partition a newly added disk with a GPT partition)

 

Joe L.

On the Main page I try to check the box and start the array, to rebuild the data. However, instead of starting the process, unraid returns me to the same screen again with the array stopped ad infinitum. Is there anything I can do? Any clues as to what might be happening? Syslog attached.

Didn't look at the syslog, but I have 2 thoughts.

 

Do you have any add ons installed? I would temporarily revert back to completely stock and try again.

 

Are you positive that your controller card chipset fully supports larger than 2.2TB disks? I'd search the forum for other people with the same hardware as you to confirm.

Recently a couple of 2TB drives started showing up as read-only. I decided to replace them with 3 and 4 TB drives, and then check the old drives and maybe preclear and add them back to the array if they are ok.

Re-constructing a corrupted file-system onto a new disk will result in a new larger disk with an equally corrupted file-system.

 

Your intention is great, but your method wrong.  Fix the file-system corruption first, then upgrade the disk.

Recently a couple of 2TB drives started showing up as read-only. I decided to replace them with 3 and 4 TB drives, and then check the old drives and maybe preclear and add them back to the array if they are ok.

Re-constructing a corrupted file-system onto a new disk will result in a new larger disk with an equally corrupted file-system.

 

Your intention is great, but your method wrong.  Fix the file-system corruption first, then upgrade the disk.

 

Thanks, JoeL. I am thinking that since:

The drive is installed physically,

The drive it replaces was about 98% full...

 

That I should do it this way around:

Build the data onto the new drive.

Fix the filesystem.

 

That way, I'll have a back up on the original drive of the files, in case fixing the filesytem results in some kind of disaster. What do you think? I am happy to take your advice and replace the original drive first if you think that would be a better and safer way to do things.

Did you happen to find and read this thread?

 

    http://lime-technology.com/forum/index.php?topic=27221.0

 

 

 

Are you positive that your controller card chipset fully supports larger than 2.2TB disks? I'd search the forum for other people with the same hardware as you to confirm.

 

My parity drive is 4TB, and I am running a 3 TB drive as well in the arr. No issues. Thanks for the suggestion.

Did you happen to find and read this thread?

 

    http://lime-technology.com/forum/index.php?topic=27221.0

 

Thanks. I could try preclearing, but I was hoping to have to avoid almost a week of preclearing and rebuilding data (I think it'll take about 5 days to do all that for a 4 TB drive in my system).

 

I don't think I have simple features turned on, and I don't know how to go back to stock unraid. Is there a way to disable the added packages without uninstalling them?

 

Did you happen to find and read this thread?

 

    http://lime-technology.com/forum/index.php?topic=27221.0

 

Thanks. I could try preclearing, but I was hoping to have to avoid almost a week of preclearing and rebuilding data (I think it'll take about 5 days to do all that for a 4 TB drive in my system).

 

I don't think I have simple features turned on, and I don't know how to go back to stock unraid. Is there a way to disable the added packages without uninstalling them?

 

 

One preclear cycle on a 4TB drive will take about 36 hours. It's worth it to at least do one cycle.

Did you happen to find and read this thread?

 

    http://lime-technology.com/forum/index.php?topic=27221.0

 

Thanks. I could try preclearing, but I was hoping to have to avoid almost a week of preclearing and rebuilding data (I think it'll take about 5 days to do all that for a 4 TB drive in my system).

 

I don't think I have simple features turned on, and I don't know how to go back to stock unraid. Is there a way to disable the added packages without uninstalling them?

 

see my sig.

see my sig.

 

Thanks very much! I am currently preclearing the drive (5% post read right now). This info will be useful in the (near) future.

I  precleared the 4 TB drive and worryingly, the drive showed close to the threshold values for a few of the parameters, but was otherwise ok. I assigned the drive and statred to rebuild the data onto it. Unfortunately there was a write error right off the bat and the drive redballed. Upon thinking about it, it made sense if the corrupt file system is being copied onto the drive to start with, there will be a write error from the beginning. I guess this is what Joe L was trying to warn me about. I've since copied the missing drive's data using parity info to a 2TB NTFS drive on my PC.

 

My thoughts on how to proceed are:

• To remove the missing disk 10 altogether from the array.
  
• Rebuild parity with missing drive.
  
• Preclear a drive and add it to the array.
  
• Copy the data from the missing drive (from NTFS backup) to the new drive.

 

Then for the other read-only drive in the array:

 

• Copy all the data to an off array drive on my PC.
  
• Try to fix the ReiserFS on the drive.
  
• If all is well and no data is lost - hurrah!
  
• If there are problems, but the filesystem is fixed, I'll just copy all the data back onto the drive form the backup, or at least the files that were lost during the fixing of the ResierFS.
  
• 
  

I assigned the drive and statred to rebuild the data onto it. Unfortunately there was a write error right off the bat and the drive redballed. Upon thinking about it, it made sense if the corrupt file system is being copied onto the drive to start with, there will be a write error from the beginning.

No - the write error applied to the new drive.  A corrupt file system does not cause a  write error - it simply gets faithfully copied - and the resulting disk also has a corrupt file system that will need a run of something like reiserfsck to fix it.

I assigned the drive and statred to rebuild the data onto it. Unfortunately there was a write error right off the bat and the drive redballed. Upon thinking about it, it made sense if the corrupt file system is being copied onto the drive to start with, there will be a write error from the beginning.

No - the write error applied to the new drive.  A corrupt file system does not cause a  write error - it simply gets faithfully copied - and the resulting disk also has a corrupt file system that will need a run of something like reiserfsck to fix it.

 

Correct. This write error has nothing to do with the corrupt file system. You're confusing the various layers of the protocol stack. This error occurred at the physical layer. The file system is a higher level abstraction.

I assigned the drive and statred to rebuild the data onto it. Unfortunately there was a write error right off the bat and the drive redballed. Upon thinking about it, it made sense if the corrupt file system is being copied onto the drive to start with, there will be a write error from the beginning.

No - the write error applied to the new drive.  A corrupt file system does not cause a  write error - it simply gets faithfully copied - and the resulting disk also has a corrupt file system that will need a run of something like reiserfsck to fix it.

 

Correct. This write error has nothing to do with the corrupt file system. You're confusing the various layers of the protocol stack. This error occurred at the physical layer. The file system is a higher level abstraction.

 

Thanks very much, both of you. Unfortunately I don't have another 4 TB drive to replace it with, so I will use a 3 TB one. The drive I was originally replacing was a 2 TB one.

Then repair the fs, and then RMA the 4 TB drive. I'd sworn off Seagates a long time ago, but decided to give this one a go. So far my best success has been with WD blacks and anything Samsung/Hitachi/Toshiba (except for the Samsung firmware fiasco a while back). I'm sure all drives have their issues, and I've just been unlucky with Seagates and WD Greens.

I'm not sure whether to start another thread or not, but the problems continue. Brief summary:

 

I replaced a 2 TB drive with a precleared 4 TB drive. The 4 TB drive gave a write error in the first few seconds of the data rebuild. I replaced it with new a 3 TB drive while I wait for the RMA. Unraid will not let me rebuild the 2 TB of data onto the 3 TB drive because it is replacing a 4TB drive. Essentially I'm replacing a 2TB drive with a 3 TB, but that momentary introduction of the faulty 4 TB drive has screwed things up. Fortunately I backed up all the files of the original 2 TB drive onto an NTFS volume beforehand. What's the best way to proceed now? Options:

 

1) Uninstall the 3 TB drive, tell unraid to rewrite parity with remaining drives, preclear 3 TB, add it to the array, and then copy the 2TB's data back to it across the network from the NTFS drive.

2) Wait for RMA'd 4 TB drive, which could be 2 weeks, leaving the other drives unprotected. Don't really want to do this, as my risk taking has a bad track record.

3) Reinstall the 2 Tb drive (which has a suspect ReiserFS), force unRaid to accept the new configuration. Rebuild parity, fix the ReiserFS of the 2 TB drive, and then replace one of my 1.5 TB drives with the 3 TB drive. I am not sure whether unraid may have written to the 'missing' disk 10 (originally the 2 TB drive) since it was removed. I doubt it because the virtual drive is being recreated from parity, and should therefore have writes disabled (because the ReiserFS was corrupt on the original disk at the time of its removal.)

4) Something else.

 

I'm leaning towards 1. It'll have my other drives protected sooner, won't require me to run a reiserfsck on the original 2 TB drive, and gives me two copies of the missing data - one on the removed 2 TB disk 10, and one on the NTFS volume. Any recommendations appreciated.