Topic: inode failures... what next?

[John Lewis]

Looks like I have a failing HD, or bad blocks, on my SME 5.03 server.  Any suggestions as to what my next actions should be?

Will going to runlevel 1, and running fsck do the trick?  Is there anything else I can do to recover the files from this bad block?

I have the following in my /var/log/messages:

Jan  1 04:02:06 gluon kernel: hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
Jan  1 04:02:06 gluon kernel: hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=38049025, sector=37486680
Jan  1 04:02:06 gluon kernel: end_request: I/O error, dev 03:06 (hda), sector 37486680
Jan  1 04:02:06 gluon kernel: EXT2-fs error (device ide0(3,6)): ext2_write_inode: unable to read inode block - inode=2333955, block=4685835
Jan  1 04:02:11 gluon kernel: hda: dma_intr: status=0x51 { DriveReady SeekComplete Error }
Jan  1 04:02:11 gluon kernel: hda: dma_intr: error=0x40 { UncorrectableError }, LBAsect=38049025, sector=37486680
Jan  1 04:02:11 gluon kernel: end_request: I/O error, dev 03:06 (hda), sector 37486680
Jan  1 04:02:11 gluon kernel: EXT2-fs error (device ide0(3,6)): ext2_write_inode: unable to read inode block - inode=2333956, block=4685835
Jan  1 04:02:11 gluon kernel: EXT2-fs error (device ide0(3,6)): ext2_write_inode: unable to read inode block - inode=2333954, block=4685835
Jan  1 04:02:11 gluon kernel: EXT2-fs error (device ide0(3,6)): ext2_write_inode: unable to read inode block - inode=2333953, block=4685835

[Julian Luton]

Dear John,
I had this on a RedHat 6 box a couple of years back and running fsck on bootup did the trick.
My problem was caused by a powercut. Do you know how yours started. If it was switched off without being properly shutdown then this is recoverable.
However if it went down during normal operation then you probably have a failing hard drive and should backup and replace it ASAP - you may be able to get fsck to flag up and block out the bad blocks but it'll only happen again until you get serious data loss.
All the best,
Julian

[Rob Hillis]

John Lewis wrote:

> Looks like I have a failing HD, or bad blocks, on my SME 5.03
> server.  Any suggestions as to what my next actions should be?

If you suspect your hard drive's reliability then BACK UP IMMEDIATELY.  I'd be looking to replace that hard drive ASAP.

> Will going to runlevel 1, and running fsck do the trick?

*Maybe* in the short-term, until the next bad block pops up.

> Is
> there anything else I can do to recover the files from this
> bad block?

Depends on whether it's truly a bad block, or whether it's just marginal.  If it's marginal, then repeated attempts to copy the file will probably recover the file.

[John Lewis]

Thanks for the replies.  Nope, this did not happen because of a powercut.  The server is on an UPS...

What makes this bad is that since this inode event, the tape backups have failed to backup anything.

Multiple fsck attempts did not fix the issue, so I'm currently embarking on backing up to desktop, and restoring to a new hard drive...

When I can reformat the old drive in peace, I'll see what happens.

Frustrating as the drive is a recent replacement from IBM of a earlier failed drive...

Seems like HD's are going the way of floppy drives.  Cheaper and more prone to failure.