harddisk crash -> needs surgery-> end_request I/O error, dev 03:01 sector 6928
Hello, Basically our suse is not booting anymore with a kernel panic: VFS: Unable to mount root fs on 03:01 It is suse 9.0 with 2.4 kernel the disk partitions hda1 5GB / reiserfs hda2 250MB swap hda3 1GB windows partition This is the complete error: hda:dma_intr:status = 0x51{DriveReady SeekComplete Error} hda:dma_intr:status = 0x40{Uncorrectable Error} LBAsect=7002,sector=6928 end_request:I/O error,dev 03:01(hda),sector 6928 journal-1212:REPLAY FAILURE fsck required! buffer write failure replay failure, unable to mount sh-2022:reiserfs_read_super:unable to initialize journal space kernel panic:VFS:Unable to mount root fs on 03:01 I have tried reiserfsck with all options that make sense: --fix_fixable --rebuild-tree --scan-whole-partition Every time is boils down to I/O error on block 806 I have also tried to mount it read only and with force but it doesn't work. Google does not come up with realy usefull stuff. One suggested it had to do with his memory. I ran a memtest, everything OK I also booted the kernel with ide=nodma because I thaught it had to do with udma but that made no difference either Is there still a way to recover that disk? Even if it means with a hexeditor (I am willing learn, who can initiate me into the deep secrets of filesystems) This disk gets backed up every night, but we got stroke by a lightening during the day cutting of all power. When rebooting this was the result. Guess what, the sales guy's just saved a bunch of important stuff on it and they want I back :-( Thanx for anyone who can provide tips. Sincerely, Filip -- *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* * System Engineer, Verzekeringen NV * * www.verzekeringen.be * * Oostkaai 23 B-2170 Merksem * * 03/6416673 - 0477/340942 * *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
On Thursday 05 August 2004 11:44 pm, Filip Sergeys wrote:
This disk gets backed up every night, but we got stroke by a lightening during the day cutting of all power. When rebooting this was the result. Guess what, the sales guy's just saved a bunch of important stuff on it and they want I back :-(
Hey, Tell those sales guys this was a message from GOD. With lightning involved you are lucky to have a machine at all. The drive manufacturer might have utilities on their web site that can try to repair this. I know hitachi has some and so does seagate (seatools). But realistically, the best bet is run down to the computer store and get another drive, re-install, and then start copying your data back on (unless the lightning took the backup device too). Then you can fiddle around trying to recover the data. I've known systems to mount drives as long as it wasn't the boot or root drive, but realistically, unless you want to go to heroic extremes and send it to Ontrack or something, it seems unlikely. -- _____________________________________ John Andersen
On Fri, 2004-08-06 at 10:11, John Andersen wrote:
On Thursday 05 August 2004 11:44 pm, Filip Sergeys wrote:
This disk gets backed up every night, but we got stroke by a lightening during the day cutting of all power. When rebooting this was the result. Guess what, the sales guy's just saved a bunch of important stuff on it and they want I back :-(
Hey, Tell those sales guys this was a message from GOD.
Funny, but not exactly what I had hoped for ;)
With lightning involved you are lucky to have a machine at all.
The drive manufacturer might have utilities on their web site that can try to repair this. I know hitachi has some and so does seagate (seatools). But realistically, the best bet is run down to the computer store and get another drive, re-install, and then start copying your data back on (unless the lightning took the backup device too).
Done and they are 98% happy, except for that piece of data :) Thanx for pointing me to seatools, I will try that.
Then you can fiddle around trying to recover the data. Fiddle around: suggestions for tools, commandline arguments, strategies, ...? Is there a way to get into the heart of the filesystem?
I've known
systems to mount drives as long as it wasn't the boot or root drive, but realistically, unless you want to go to heroic extremes and send it to Ontrack or something, it seems unlikely.
-- _____________________________________ John Andersen
-- *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* * System Engineer, Verzekeringen NV * * www.verzekeringen.be * * Oostkaai 23 B-2170 Merksem * * 03/6416673 - 0477/340942 * *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
On Friday 06 August 2004 12:43 am, Filip Sergeys wrote:
Fiddle around: suggestions for tools, commandline arguments, strategies, ...? Is there a way to get into the heart of the filesystem?
Well, yes, you could get the specs for reiserfs and start bit by bit, but first you have to get the data off the drive, and THAT does not sound very promising... As I said, I recently had an drive from which I could not boot because of damage. I was able to mount it, (even tho it complained) and read certain portions of it. But if I ever even LOOKED at the bad sectors, the thing would lock up and finally time out. In the end I salvaged about half of the needed data off of the drive. The damage was in the file system for /bin and /sbin and the boot sector was un-readable as well. User data was quite a ways further down the drive. -- _____________________________________ John Andersen
On Fri, 2004-08-06 at 11:09, John Andersen wrote: > On Friday 06 August 2004 12:43 am, Filip Sergeys wrote: > > Fiddle around: suggestions for tools, commandline arguments, strategies, > > ...? > > Is there a way to get into the heart of the filesystem? > > Well, yes, you could get the specs for reiserfs and start bit by bit, > but first you have to get the data off the drive, and THAT does not > sound very promising... > > As I said, I recently had an drive from which I could not boot > because of damage. I was able to mount it, (even tho it complained) > and read certain portions of it. But if I ever even LOOKED at the > bad sectors, the thing would lock up and finally time out. > I have no idea yet what is on the bad sectors. So that will be a trial and error to read off the parts that are not damaged. I basically only need the /home dirs. Did you use specific mount options? > In the end I salvaged about half of the needed data off > of the drive. > The damage was in the file system for /bin and /sbin and > the boot sector was un-readable as well. User data was > quite a ways further down the drive. > > -- > _____________________________________ > John Andersen -- *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* * System Engineer, Verzekeringen NV * * www.verzekeringen.be * * Oostkaai 23 B-2170 Merksem * * 03/6416673 - 0477/340942 * *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
Am Freitag, 6. August 2004 09:44 schrieb Filip Sergeys:
journal-1212:REPLAY FAILURE fsck required! buffer write failure replay failure, unable to mount Since it's the journal replay faling, try mounting with "nolog" option. There also is "debugreiserfs(8)". Might want to check that out. HTH Dan -- buddha 2.6.4-54.5-default 9:05am an 19:31, 1 Benutzer,
Fredag den 6. august 2004 09:44 skrev Filip Sergeys:
Hello,
Basically our suse is not booting anymore with a kernel panic: VFS: Unable to mount root fs on 03:01
It is suse 9.0 with 2.4 kernel the disk partitions hda1 5GB / reiserfs hda2 250MB swap hda3 1GB windows partition
This is the complete error:
hda:dma_intr:status = 0x51{DriveReady SeekComplete Error} hda:dma_intr:status = 0x40{Uncorrectable Error} LBAsect=7002,sector=6928 end_request:I/O error,dev 03:01(hda),sector 6928 journal-1212:REPLAY FAILURE fsck required! buffer write failure replay failure, unable to mount sh-2022:reiserfs_read_super:unable to initialize journal space kernel panic:VFS:Unable to mount root fs on 03:01
I have tried reiserfsck with all options that make sense: --fix_fixable --rebuild-tree --scan-whole-partition Every time is boils down to I/O error on block 806
I have also tried to mount it read only and with force but it doesn't work.
Maybee the options of the boot CD can help you out - it did that for me :-)
Google does not come up with realy usefull stuff. One suggested it had to do with his memory. I ran a memtest, everything OK
I also booted the kernel with ide=nodma because I thaught it had to do with udma but that made no difference either
Is there still a way to recover that disk? Even if it means with a hexeditor (I am willing learn, who can initiate me into the deep secrets of filesystems) This disk gets backed up every night, but we got stroke by a lightening during the day cutting of all power. When rebooting this was the result. Guess what, the sales guy's just saved a bunch of important stuff on it and they want I back :-(
Thanx for anyone who can provide tips.
Sincerely,
Filip
-- *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* * System Engineer, Verzekeringen NV * * www.verzekeringen.be * * Oostkaai 23 B-2170 Merksem * * 03/6416673 - 0477/340942 * *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
On Fri, 2004-08-06 at 13:23, Johan Nielsen wrote:
Fredag den 6. august 2004 09:44 skrev Filip Sergeys:
Hello,
Basically our suse is not booting anymore with a kernel panic: VFS: Unable to mount root fs on 03:01
It is suse 9.0 with 2.4 kernel the disk partitions hda1 5GB / reiserfs hda2 250MB swap hda3 1GB windows partition
This is the complete error:
hda:dma_intr:status = 0x51{DriveReady SeekComplete Error} hda:dma_intr:status = 0x40{Uncorrectable Error} LBAsect=7002,sector=6928 end_request:I/O error,dev 03:01(hda),sector 6928 journal-1212:REPLAY FAILURE fsck required! buffer write failure replay failure, unable to mount sh-2022:reiserfs_read_super:unable to initialize journal space kernel panic:VFS:Unable to mount root fs on 03:01
I have tried reiserfsck with all options that make sense: --fix_fixable --rebuild-tree --scan-whole-partition Every time is boils down to I/O error on block 806
I have also tried to mount it read only and with force but it doesn't work.
Maybee the options of the boot CD can help you out - it did that for me :-)
I tryed "Repair system" before, but didn't solve. Basically because the reiserfsck done during the repair fails.
Google does not come up with realy usefull stuff. One suggested it had to do with his memory. I ran a memtest, everything OK
I also booted the kernel with ide=nodma because I thaught it had to do with udma but that made no difference either
Is there still a way to recover that disk? Even if it means with a hexeditor (I am willing learn, who can initiate me into the deep secrets of filesystems) This disk gets backed up every night, but we got stroke by a lightening during the day cutting of all power. When rebooting this was the result. Guess what, the sales guy's just saved a bunch of important stuff on it and they want I back :-(
Thanx for anyone who can provide tips.
Sincerely,
Filip
-- *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* * System Engineer, Verzekeringen NV * * www.verzekeringen.be * * Oostkaai 23 B-2170 Merksem * * 03/6416673 - 0477/340942 * *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
-- Check the headers for your unsubscription address For additional commands send e-mail to suse-linux-e-help@suse.com Also check the archives at http://lists.suse.com Please read the FAQs: suse-linux-e-faq@suse.com
-- *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* * System Engineer, Verzekeringen NV * * www.verzekeringen.be * * Oostkaai 23 B-2170 Merksem * * 03/6416673 - 0477/340942 * *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
On 6 Aug 2004, Filip Sergeys wrote:
Hello,
Basically our suse is not booting anymore with a kernel panic: VFS: Unable to mount root fs on 03:01
... Is there still a way to recover that disk? Even if it means with a hexeditor (I am willing learn, who can initiate me into the deep secrets of filesystems) This disk gets backed up every night, but we got stroke by a lightening during the day cutting of all power. When rebooting this was the result. Guess what, the sales guy's just saved a bunch of important stuff on it and they want I back :-(
Have you tried ddrescue, yet? It is on the DVD. The rpm is called ddrescue while the command is called dd_rescue. Type 'dd_rescue -h' for usage info. Good luck, Alex.
On Fri, 2004-08-06 at 13:50, Alex Angerhofer wrote:
On 6 Aug 2004, Filip Sergeys wrote:
Hello,
Basically our suse is not booting anymore with a kernel panic: VFS: Unable to mount root fs on 03:01
... Is there still a way to recover that disk? Even if it means with a hexeditor (I am willing learn, who can initiate me into the deep secrets of filesystems) This disk gets backed up every night, but we got stroke by a lightening during the day cutting of all power. When rebooting this was the result. Guess what, the sales guy's just saved a bunch of important stuff on it and they want I back :-(
Have you tried ddrescue, yet? It is on the DVD. The rpm is called ddrescue while the command is called dd_rescue. Type 'dd_rescue -h' for usage info.
I will try dd_rhelp. It uses dd_rescue but has a few improvements to save you time (that is what it says at least) It starts at arbitrary places on the disc and if it encounters bad blocks, it jumps to another place an works its way back in reversed direction I'll let you know what the success rate is! Thanx for the tip. Regards, Filip
Good luck, Alex.
-- Check the headers for your unsubscription address For additional commands send e-mail to suse-linux-e-help@suse.com Also check the archives at http://lists.suse.com Please read the FAQs: suse-linux-e-faq@suse.com
-- *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* * System Engineer, Verzekeringen NV * * www.verzekeringen.be * * Oostkaai 23 B-2170 Merksem * * 03/6416673 - 0477/340942 * *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
On Friday 06 August 2004 07:44, Filip Sergeys wrote:
Basically our suse is not booting anymore with a kernel panic: VFS: Unable to mount root fs on 03:01
~ maybe, the boot ram-disk is not loading the reiserfs module . . . perhaps, worth mounting the /boot partition, then, re-running < mkinitrd > finally, re-running < lilo > ? -- best wishes ____________ sent on Linux ____________
On Friday 06 August 2004 09:44, Filip Sergeys wrote:
Hello,
Basically our suse is not booting anymore with a kernel panic: VFS: Unable to mount root fs on 03:01
It is suse 9.0 with 2.4 kernel the disk partitions hda1 5GB / reiserfs hda2 250MB swap hda3 1GB windows partition
This is the complete error:
hda:dma_intr:status = 0x51{DriveReady SeekComplete Error} hda:dma_intr:status = 0x40{Uncorrectable Error} LBAsect=7002,sector=6928 end_request:I/O error,dev 03:01(hda),sector 6928 journal-1212:REPLAY FAILURE fsck required! buffer write failure replay failure, unable to mount sh-2022:reiserfs_read_super:unable to initialize journal space kernel panic:VFS:Unable to mount root fs on 03:01
I have tried reiserfsck with all options that make sense: --fix_fixable --rebuild-tree --scan-whole-partition Every time is boils down to I/O error on block 806
I have also tried to mount it read only and with force but it doesn't work.
Are there any bad blocks on the disk? Try 'man badblocks' for more info. And there is the utility spinrite. A commercial product, $89, 30-day money-back guarantee, http://www.grc.com/, http://www.linuxjournal.com/article.php?sid=7684 Cheers, Leen
The Friday 2004-08-06 at 09:44 +0200, Filip Sergeys wrote:
hda:dma_intr:status = 0x40{Uncorrectable Error} LBAsect=7002,sector=6928 end_request:I/O error,dev 03:01(hda),sector 6928
You have a bad sector, not readable or writable. As already mentioned, you may try dd_rescue/dd_rhelp (mount the partition RO, booting from another disk). The HD "might" be reusable; bad sectors are not so rare (years ago HDs came with a list of bad sectors printed in paper), and modern HDs can remap those sectors to somewhere else on space reserved on the same disk by the manufacturer. But this feature triggers when writing to a bad sector, not when reading. The reiserfs is different from other fs (amongst many other things) in that it can not do that remapping itself, like, for example, does ext2 (or even fat), so the program 'badblocks' is of no use here - except for confirming you do have bad sectors. Your HD manufacturer will probably have an utility to verify the HD; usually, it is a 1.4Mb file for creating a boot floppy. Also, in Linux, you have 'smartctl' (package smartmontools) that not only allows reading smart values (Analysis and Reporting Technology), but can also init and read results of the HD self testing facility.
This disk gets backed up every night, but we got stroke by a lightening during the day cutting of all power. When rebooting this was the result. Guess what, the sales guy's just saved a bunch of important stuff on it and they want I back :-(
If you absolutely need that data, and dd_rescue doesn't help, the reiserfs people have support for a fee: chances are that the data you need is not on the bad sector. Also, there are companies that make their living of retrieving unrecoverable data - well, almost unrecoverable, I guess ;-) -- Cheers, Carlos Robinson
participants (8)
-
Alex Angerhofer -
Carlos E. R. -
Dan Am -
Filip Sergeys -
Johan Nielsen -
John Andersen -
Leendert Meyer -
pinto