LBA Sector error won't go away

Hi all,

I have an HDR running 1.02.32 CF 2.19 and the SMART test log is showing some errors on 197 and 198. I have booted into maintenance mode and run the fix-disk diagnostic. When it eventually gets to the relevant sector, it asks about repairing the sector, to which I answer yes, but it keeps on failing to correct the error - see excerpt from the fix-disk log.

Code:
LBA: 3017807304 is on partition /dev/sdb2, start: 2104576, bad sector offset: 3015702728
LBA 3017807304 maps to file system block 376962841 on /dev/sdb2
Block 376962841 is not in use
Running select disk self test
Error at LBA 3017807304
LBA: 3017807304 is on partition /dev/sdb2, start: 2104576, bad sector offset: 3015702728
LBA 3017807304 maps to file system block 376962841 on /dev/sdb2
Block 376962841 is not in use
Running select disk self test
Error at LBA 3017807304
LBA: 3017807304 is on partition /dev/sdb2, start: 2104576, bad sector offset: 3015702728
LBA 3017807304 maps to file system block 376962841 on /dev/sdb2
Block 376962841 is not in use
Running select disk self test

I have tried the hdparm --repair sector with the --yes-i-know-what-i-am-doing and it says succeeded but the error is still there. If I manually run the SMART short test, it succeeds. The long test fails at 30% remaining.

Any ideas how I can fix these errors? It's been a pain having the HDR in maintenance mode.

Note: originally it was telling me there was a file associated with that sector, which I then deleted from the command line.

Help!!!! ;)
 
What exactly does the smart log say when the long test fails?
(Option 4 from the telnet menu, "Check self-test progress" will tell you).

What sector number did you provide to the hdparm utility?
 
Hi af123,

Code:
Please select option: 4
 
# 1  Extended offline    Completed: read failure      30%      1393        3017807304
 
Press return to continue:

As I have run the test several times, I have several of those in the smart log when I issue smartctl -a /dev/sda and look back through the list.

I used hdparm --repair-sector 3017807304

Thanks
 
No - the --repair-sector command should have forced the disk to reallocate the sector. Does it appear to work correctly?
 
Thanks af123. The disk appears to be working fine. When the fix-disk was trying to repair the sector, it kept saying success and repeating the error/fix loop over and over. Perhaps I'll try a manual bad-blocks command now it's been a few days.
 
Not had much time to play recently. Have set the device off on another smartctl -t long /dev/sda test. While reviewing the smartctl -a /dev/sda log, I have seen this at the bottom for quite a while. Is this something that is preventing fixing the LBA error?

Code:
SMART Selective self-test log data structure revision number 1
SPAN    MIN_LBA    MAX_LBA  CURRENT_TEST_STATUS
    1  3017807304  3907029167  Not_testing
    2          0          0  Not_testing
    3          0          0  Not_testing
    4          0          0  Not_testing
    5          0          0  Not_testing
Selective self-test flags (0x0):
  After scanning selected spans, do NOT read-scan remainder of disk.
If Selective self-test is pending on power-up, resume after 0 minute delay.

Thanks
 
Ah, OK. Thanks xyz321.

Never got the checkdisk with badblocks to complete - had to abort after a several hours. May have another go soon.
 
fix-disk will not help if the "hdparm --repair-sector" command does not fix the error (it just runs the same command). Possible solutions might be:
  • Use fix-disk with the -x -c options to force it to run a file system check which will mark and subsequently ignore the bad blocks. There is no guarantee that this will stop any stuttering playback since the fault will remain and could affect reads from other areas on the disk.
  • Try the nuke option after backing up the disk. This is the hdparm security erase function which sends a command to the disk which will cause it to wipe itself and may fix the problem - no guarantees.
  • Replace the disk.
 
Thanks for the info xyz321. I do not have an stuttering playback from the disk - I deleted the file that was identified as using that block. I was trying to fix the disk errors reported on the smart test, and obviously can't while the smart test stops at that bad block.

I have started an e2fsck -vv -c -p /dev/sda2 after setting up a swap area manually. I don't know if the -p has made it silent (I see no output from the -vv), as I ran e2fsck -vv -c on sda1 and 3 and had output. I thought the -p option would be a good thing for running overnight. By the way, f2fsck doesn't list a -x option - is that something for the fix-disk?

Thanks.
 
Back
Top