Hi,

The following is my configuration.

 

 

I noticed that my Parity Drive as well as my Disk1 Drive has bad sectors. Please see the smart results for the same.

Parity Disk Smart Results :-

 

Disk 1 Smart Results :-

 

I have also attached the diagnostic reports. It has been quite a while since I have done a parity check.

 

The following is what I would like to accomplish.

 

1. I want to swap disk1 with a new 2TB disk which I have precleared. Since both the disk1 and parity drives have bad sectors, I am not sure how to proceed on this.

2. I would also like to utilize this opportunity to convert the file system to XFS. I am hesitant to do it because of the bad sector on the parity.

3. I would like some recommendations on the state of both the hard drives for future use. I mean, something like parity disk is ok to use after doing a preclear and rebuild or throw it away and get a new disk for parity etc. Same for the current disk1 too.

 

On a side note, I have backups for the entire data in Disk1. 

 

Thanks in advance. 

 

 

 

 

tower-diagnostics-20171214-0002.zip

If both have pending sectors there's no way to do it without data loss, run an extended SMART test on each before doing anything else.

Thanks. I have initiated it. While the test for Parity drive is in progress, the test for the disk1 drive seems to have failed.

 

Quote

 

Errors occurred - Check SMART report

 

I have also attached the smart report. Please let me know if I am doing this correctly. 

tower-smart-20171214-0109.zip

If it is of any help, the following is the self test history of the disk failing the test.

 

Quote

Num  Test_Description    Status                  Remaining  LifeTime(hours)  LBA_of_first_error
# 1  Extended offline    Completed: read failure       90%      2114         1536289559
# 2  Extended offline    Completed: read failure       90%      2114         1536289559
# 3  Extended offline    Completed: read failure       90%      2114         1536289559
# 4  Short offline       Completed: read failure       20%      2011         1536289559
# 5  Short offline       Completed: read failure       20%      1816         1536289559

 

If you know which files on the data disk that is broken and you have a backup of them, then you might be able to short-term repair the data disk by copying the backup file to the disk. This will give the data disk a chance to map in spare sectors instead of the broken sectors. So you might be able to get current-pending and offline-uncorrectable zeroed.

 

That could then allow the RAID to build a new parity disk.

 

Or you might be able to short-term repair the parity drive too by having the machine scan and repair the parity - but that is a step you don't want to do until you have managed an extended SMART scan of all data drives with zero pending/uncorrectable sectors.

 

The following link contains information how to figure out from broken sector which file i affected:

https://www.smartmontools.org/browser/trunk/www/badblockhowto.xml#reiserfs_ex

17 hours ago, johnnie.black said:

If both have pending sectors there's no way to do it without data loss, run an extended SMART test on each before doing anything else.

 

I did run the extended SMART test on both the disks. The test failed for disk1 with the following error.

Num  Test_Description    Status                  Remaining  LifeTime(hours)  LBA_of_first_error
# 1  Extended offline    Completed: read failure       90%      2114         1536289559
# 2  Extended offline    Completed: read failure       90%      2114         1536289559
# 3  Extended offline    Completed: read failure       90%      2114         1536289559
# 4  Short offline       Completed: read failure       20%      2011         1536289559
# 5  Short offline       Completed: read failure       20%      1816         1536289559

 

I have attached the smart reports for both the disks.

 

 

17 hours ago, pwm said:

If you know which files on the data disk that is broken and you have a backup of them, then you might be able to short-term repair the data disk by copying the backup file to the disk. This will give the data disk a chance to map in spare sectors instead of the broken sectors. So you might be able to get current-pending and offline-uncorrectable zeroed.

 

That could then allow the RAID to build a new parity disk.

 

Or you might be able to short-term repair the parity drive too by having the machine scan and repair the parity - but that is a step you don't want to do until you have managed an extended SMART scan of all data drives with zero pending/uncorrectable sectors.

 

The following link contains information how to figure out from broken sector which file i affected:

https://www.smartmontools.org/browser/trunk/www/badblockhowto.xml#reiserfs_ex

Since the smart test failed by just reporting the error at this block - 1536289559, I tried the instructions to figure out the files in the broken sector. It returns the following.

 

root@Tower:~# debugreiserfs /dev/sde1 |grep '^Blocksize'
debugreiserfs 3.6.24

Blocksize: 4096

root@Tower:~# sfdisk -luS /dev/sde
Disk /dev/sde: 931.5 GiB, 1000204886016 bytes, 1953525168 sectors
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0x3f64c1aa

Device     Boot Start        End    Sectors   Size Id Type
/dev/sde1          63 1953525167 1953525105 931.5G 83 Linux

root@Tower:~# echo "(1536289559-63)*512/4096" |bc -l
192036187.00000000000000000000


root@Tower:~# debugreiserfs -1 192036187 /dev/sde1
debugreiserfs 3.6.24

192036187 is free in ondisk bitmap
Looks like unformatted

The result looks more confusing. Is it because there is no file at that sector or something else in the calculation is wrong due to which the result indicates unformatted.

 

Please let me know the next steps.

 

Since part of my objective was also to convert the file systems, does the following plan makes sense?

1. Unassign Parity

2. Copy Disk2 Files to the new hard drive.

3. Swap disk name assignments so that the new hard drive becomes Disk2 and existing disk2 is made available for round two of copy

4. Copy Disk1 files to the previous disk2 drive. Verify the files with the backups to find any missing or corrupt files

5. Rebuild parity on a new drive.

 

 

 

 

ST2000DL003-9VT166_5YD3XCND-20171214-1256.txt

SAMSUNG_HD103SI_S1XGJ90S612592-20171214-1256.txt

Edited December 14, 20178 yr by sreeju
corrected punctuation

Parity disk passed the extended SMART test, do it's good at least for now, if you have a spare you should be able to successfully replace disk1, if your spare is larger than parity you can use the parity swap procedure.

Thanks @johnnie.blackand @pwm for your help. Since I had backups, I didn't rebuild the disks. I used the spare to convert the other disks to XFS and then copied the files affected by the bad disk to the spare. I then rebuilt the parity to a new disk.