little problem with reiserfs and bad blocks
I have a drive that failed fsck after a power outage. I find out it reported bad blocks, and so fsck can't do it's thing. Now that drive is data, the bulk of it replaceable with a lot of time. (70 gigs, what can I say, I'm a pack rat) But anyways. I would like to be able to get this drive usable and salvage what I can. Using the badblock command, it reports 4 blocks bad. I have yet to find anything usefull on bad blocks, etc with reiserfs. Heck, I think if I could just force mount it, I should get most of it right? Any help here? I know ext2/3, fat, fat32, hpfs, ntfs, all have ways to deal with this, but I have yet to find anyone in linux w/ reiserfs.
On Fri, 2004-01-30 at 11:11, c_nelson77 wrote:
I have a drive that failed fsck after a power outage. I find out it reported bad blocks, and so fsck can't do it's thing. Now that drive is data, the bulk of it replaceable with a lot of time. (70 gigs, what can I say, I'm a pack rat)
But anyways. I would like to be able to get this drive usable and salvage what I can. Using the badblock command, it reports 4 blocks bad. I have yet to find anything usefull on bad blocks, etc with reiserfs. Heck, I think if I could just force mount it, I should get most of it right?
Any help here? I know ext2/3, fat, fat32, hpfs, ntfs, all have ways to deal with this, but I have yet to find anyone in linux w/ reiserfs.
Have you tried running 'resierfsck --rebuild-tree' ? -- Anders Karlsson <anders@trudheim.com> Trudheim Technology Limited
IIRC, reiserfs assigns reserved blocks to any bad blocks it may find, and from then on ignores the bad blocks. This may work as long as there are enough good blocks in reserve. Once there are too many bad blocks it is probably time to swap the HD out. I would try to force mount the partition and get off what you can. I did this with my /dev/hda which recently ended up toast and got the majority of my files saved. You may want to look into dd_rescue (http://www.garloff.de/kurt/linux/ddrescue/ ) if you have problems mounting and reading the partition. Best regards, Alex. On Fri, 30 Jan 2004, Anders Karlsson wrote:
On Fri, 2004-01-30 at 11:11, c_nelson77 wrote:
I have a drive that failed fsck after a power outage. I find out it reported bad blocks, and so fsck can't do it's thing. Now that drive is data, the bulk of it replaceable with a lot of time. (70 gigs, what can I say, I'm a pack rat)
But anyways. I would like to be able to get this drive usable and salvage what I can. Using the badblock command, it reports 4 blocks bad. I have yet to find anything usefull on bad blocks, etc with reiserfs. Heck, I think if I could just force mount it, I should get most of it right?
Any help here? I know ext2/3, fat, fat32, hpfs, ntfs, all have ways to deal with this, but I have yet to find anyone in linux w/ reiserfs.
Have you tried running 'resierfsck --rebuild-tree' ?
-- Anders Karlsson <anders@trudheim.com> Trudheim Technology Limited
Hi, On Friday 30 January 2004 16:07, Alex Angerhofer wrote:
probably time to swap the HD out. I would try to force mount the partition and get off what you can. I did this with my /dev/hda which recently ended
I recently experienced an unrecoverable error with another journaling fs (jfs in my case). Also on a big HD. I went out and bought a new one (bigger) Mount the old partition/HD read only and copy it all to a new partition Badblocks tend to grow in my experience. On modern HD's teh HD's firmware replaces bad blocks by hidden spare ones (you will hear the head make quick full disk movements when reading/writing to them). If badblocks finds any, it means the disk has gone bad and is out of spare blocks. This tends to worsen with time. Replace now, and see if it still has warranty on it. If it is out of it, some makes have a tool on their website (windows/dos only, Maxtor is one example) that can lowelevel test a hd and rebuild (wipes disk) a HD skipping bad blocks and makeing new spares (iirc) But still, the safest is to replace it. BB, Arjen
Yes I have, it keeps failing because of the bad blocks. On Friday 30 January 2004 04:34, Anders Karlsson wrote:
On Fri, 2004-01-30 at 11:11, c_nelson77 wrote:
I have a drive that failed fsck after a power outage. I find out it reported bad blocks, and so fsck can't do it's thing. Now that drive is data, the bulk of it replaceable with a lot of time. (70 gigs, what can I say, I'm a pack rat)
But anyways. I would like to be able to get this drive usable and salvage what I can. Using the badblock command, it reports 4 blocks bad. I have yet to find anything usefull on bad blocks, etc with reiserfs. Heck, I think if I could just force mount it, I should get most of it right?
Any help here? I know ext2/3, fat, fat32, hpfs, ntfs, all have ways to deal with this, but I have yet to find anyone in linux w/ reiserfs.
Have you tried running 'resierfsck --rebuild-tree' ?
The Friday 2004-01-30 at 04:11 -0700, c_nelson77 wrote:
Any help here? I know ext2/3, fat, fat32, hpfs, ntfs, all have ways to deal with this, but I have yet to find anyone in linux w/ reiserfs.
That's right, reiserfs can not handle badblocks. ext273 can. There is a way to do it manually, but I don't like it (fsck will remove the hack). However, possibly your HD firmware can move badbloks elsewere, transparent to any OS. This is enabled with hdparm, and the log/status can be seen with smartctl. Relocation of bad sectors happens when trying to write to a bad sector (but not on reading). -- Cheers, Carlos Robinson
On Fri, 30 Jan 2004 04:11:50 -0700 c_nelson77 <c_nelson77@comcast.net> wrote:
I have a drive that failed fsck after a power outage. I find out it reported bad blocks, and so fsck can't do it's thing.
Well, you can do it, but it is very convoluted: http://www.namesys.com/bad-block-handling.html Charles -- "Computers may be stupid, but they're always obedient. Well, almost always." -- Larry Wall (Open Sources, 1999 O'Reilly and Associates)
participants (6)
-
Alex Angerhofer -
Anders Karlsson -
Arjen Runsink -
c_nelson77 -
Carlos E. R. -
Charles Philip Chan