Can I somehow force a preclear on a previously assigned drive?  The drive had a sector fault and I want to run a few cycles of preclear to see if it will clear the fault, but preclear sees this drive as still assigned to the array, and won't allow it to be precleared.  I want to try this before I go out and buy a replacement drive.

 

 

 

 

Can I somehow force a preclear on a previously assigned drive?  The drive had a sector fault and I want to run a few cycles of preclear to see if it will clear the fault, but preclear sees this drive as still assigned to the array, and won't allow it to be precleared.  I want to try this before I go out and buy a replacement drive.

 

The only way I know of to remove a drive from the array is to run New Config from the GUI or initconfig from telnet. This will wipe all your drive assignments.

 

You will want to screenshot your current drive placements, then stop the array and run one of the commands above. You can then go back to the main GUI page and manually reassign the disks back (excluding the one you want removed from the array). Once you start up the array though it will want to redo your parity build (you can select valid parity to avoid this, but I am not sure of the risk).

 

Start the array with the drive unassigned.

dgaschk,

  Tried doing that, but when running preclear_disk.sh, it comes back, "disk is already assigned to array"  "no preclear action will be performed". even though the GUI says that it is not assigned.  I saw in other posts, that the super.dat file probably still has the serial number recorded so is still recognized by the array?

 

bkastner,

    Was hesitant to perform the new config steps because this disk was originally part of the array, and wasn't sure if I did the new config with now only 2 disks: parity and the working disk, if there was potential to loose the data. Have not performed any parity check yet since this disk went faulty, until I know I've reconstructed the array properly. Have duplicated all the "shares" that resided on the faulty disk, so I don't care about that data, just the existing data on the 2 good disks.

 

Thanks.

You have 3 drives + Parity?  Or just 3 drives?

 

What does smart say about the failed drive?

bkastner,

    Was hesitant to perform the new config steps because this disk was originally part of the array, and wasn't sure if I did the new config with now only 2 disks: parity and the working disk, if there was potential to loose the data. Have not performed any parity check yet since this disk went faulty, until I know I've reconstructed the array properly. Have duplicated all the "shares" that resided on the faulty disk, so I don't care about that data, just the existing data on the 2 good disks.

 

Thanks.

 

If there is data on the faulty disk then doing a new config will prevent you from getting it back, but I don't think you can re-preclear it without this step (as you've seen). If you want to recover data from the disk your best option is to go buy a disk as big or larger, preclear it and then add it to the array as a replacement for the faulty disk. You can then re-preclear the faulty disk and if you want you should be able to add it as a new drive after.

 

The only other option I can think of is to boot up a new USB key with UnRAID and then you could preclear the disk because there will be no legacy data. Once you've precleared it and are happy you can swap back the old USB key with your actual config and either re-add the disk to the array or add a new one.

 

The downside of this solution is your array is completely offline until you boot back up with your original USB drive.

 

Also, you want to make 1000% sure you preclear the right disk while on the USB drive, or else you are totally pooched.

 

If you are only concerned about the data on the one good data drive now then you could safely run new config and just reassign the parity & data drive (screen shot the config before so you get the drives in the right place). The above comments are only applicable if you want the chance to recover data from the questionable drive.

 

Bkastner, took your advice and started up a fresh install of UNRAID and preclearing suspect drive.  Seems to have stuck in a loop after 12 hours, on Step 2 - Copying zeros to remainder of disk.  Progress shows the following:

192842+5 records in

192842+5 records out

404422705152 bytes (404 GB) copied, 2824035 s, 143 MB/s

Wrote 404,422,705,152 bytes ot of 2,000,398,934,016 bytes (20% Done)

 

These lines keep refreshing with no change in status amounts.

 

Thoughts?

 

 

 

Chris, 3 drives: 2 data, 1 parity, all 2TB Hitachi 7K3000.  Smart came back with 1 IDNF error, but no abnormal results in the attributes section of the report, and both assessment tests pass without errors.

Bkastner, took your advice and started up a fresh install of UNRAID and preclearing suspect drive.  Seems to have stuck in a loop after 12 hours, on Step 2 - Copying zeros to remainder of disk.  Progress shows the following:

192842+5 records in

192842+5 records out

404422705152 bytes (404 GB) copied, 2824035 s, 143 MB/s

Wrote 404,422,705,152 bytes ot of 2,000,398,934,016 bytes (20% Done)

 

These lines keep refreshing with no change in status amounts.

 

Thoughts?

 

 

 

Chris, 3 drives: 2 data, 1 parity, all 2TB Hitachi 7K3000.  Smart came back with 1 IDNF error, but no abnormal results in the attributes section of the report, and both assessment tests pass without errors.

 

I am unclear if the drive that reported the IDNF error is the one you are preclearing and having issues with. If you look up IDNF:

 

Address Not Found Error. The Identify Drive Command has not received an acceptable response from the drive. This may be due to a defect. Ensure that you are using the latest version of diagnostic utility and that your cable is in good working condition. Retest. Replace the drive if the error repeats.

 

I would maybe see if Hitachi has disk diagnostic tools to check the drive out. It definitely looks like there is a disk issue. The Hitachi tools may be able to mark the sector as bad, and the rest of the disk is okay, or else, if it's under warranty you may want to RMA it to be safe.