From: a a <you...@ya...> - 2012-08-17 09:32:31
|
I have a samsung f3 drive that smartctl reports a bad block # 1 Short offline Completed: read failure 90% 9179 1422422113 # 2 Short offline Completed: read failure 90% 9179 1422422113 # 3 Extended offline Completed: read failure 40% 9178 1422422113 # 4 Short offline Completed without error 00% 9175 - # 5 Short offline Completed: read failure 90% 9175 1422422113 # 6 Extended offline Completed: read failure 40% 9165 1422422113 # 7 Short offline Completed: read failure 90% 9161 26587969 -- The first bad block 26587969 the dd trick worked fine: /dev/sdb1 63 1758667679 879333808+ fd Linux raid autodetect dd if=/dev/zero of=/dev/sdb1 bs=4096 count=1 seek=3323488 -- However I am unable to clear the second bad block (dd does not hang with a sense error and the write does not remap it). This disk is used in a linux software raid 1 configuration. - The question is how to do I clear this error so smartctl can continue the scan ? (I'm 100% sure this block is not covering important data ) |
From: Andreas M. <an...@us...> - 2012-09-10 07:10:13
|
Hi, On Mon, Sep 10, 2012 at 06:17:10AM +0000, sma...@li... wrote: > The first bad block 26587969 the dd trick worked fine: > /dev/sdb1????????????? 63? 1758667679?? 879333808+? fd? Linux raid autodetect > dd if=/dev/zero of=/dev/sdb1 bs=4096 count=1 seek=3323488 > -- > However I am unable to clear the second bad block (dd does not hang with a sense error and the write does not remap it). This disk is used in a linux software raid 1 configuration. > - > The question is how to do I clear this error so smartctl can continue the scan ? > (I'm 100% sure this block is not covering important data ) > ? Ouch, sounds fun, a bad sector which does not get sufficiently recognized as "bad" to initiate remapping yet causes SMART scans to freak out. I think tutorials about advanced uses of *GNU* ddrescue (gddrescue) (e.g. docs talking about ddrescue fill mode etc.) may cover this item, too. Or possibly the very nice sg3-tools toolset has some lowlevel help for this which is usable by ATA-based disks, too (but of course be careful with certain parts of these tools!). HTH, Andreas Mohr |
From: a a <you...@ya...> - 2012-09-10 15:07:35
|
Andreas -- Thank you for your response; what I ended up doing was failing the disk (it was in a raid 1 configuration and the error was showing up everytime the mdadm checked the array); ran badblock in destructive mode (write checks) which found no badblocks; rechecked the array with smartctl (it turned up clean) and then readded it to the array; after a long rebuild build (less than a day) all was clean :) - Naturally if this had been a non-raided disk I would have need to do something alternative. --- On Mon, 9/10/12, Andreas Mohr <an...@us...> wrote: From: Andreas Mohr <an...@us...> Subject: Re: [smartmontools-support] how to clear bad block if block is readable ? To: sma...@li... Cc: "a a" <you...@ya...> Date: Monday, September 10, 2012, 3:10 AM Hi, On Mon, Sep 10, 2012 at 06:17:10AM +0000, sma...@li... wrote: > The first bad block 26587969 the dd trick worked fine: > /dev/sdb1????????????? 63? 1758667679?? 879333808+? fd? Linux raid autodetect > dd if=/dev/zero of=/dev/sdb1 bs=4096 count=1 seek=3323488 > -- > However I am unable to clear the second bad block (dd does not hang with a sense error and the write does not remap it). This disk is used in a linux software raid 1 configuration. > - > The question is how to do I clear this error so smartctl can continue the scan ? > (I'm 100% sure this block is not covering important data ) > ? Ouch, sounds fun, a bad sector which does not get sufficiently recognized as "bad" to initiate remapping yet causes SMART scans to freak out. I think tutorials about advanced uses of *GNU* ddrescue (gddrescue) (e.g. docs talking about ddrescue fill mode etc.) may cover this item, too. Or possibly the very nice sg3-tools toolset has some lowlevel help for this which is usable by ATA-based disks, too (but of course be careful with certain parts of these tools!). HTH, Andreas Mohr |
From: Alex S. <ml...@os...> - 2012-09-10 11:16:22
|
Some my recommendation is to move this drive to a trashcan. If it is in the RAID1 and another drive is good - just disconnect it from the RAID, and do dd if=/dev/zero of=/dev/sdX bs=1M to overwrite it completely. May be it will help. On 08/17/2012 11:32 AM, a a wrote: > > The question is how to do I clear this error so smartctl can continue > the scan ? > (I'm 100% sure this block is not covering important data ) > |