[linux-lvm] mapping blocks to PVs?

[linux-lvm] mapping blocks to PVs?

[Andres Salomon]

I've got a LV that keeps corrupting, made up of several IDE disks.  I've
already discovered one of the drives to be flaking out, after various
read tests on the device gave BadCRC/IRQ timeout errors.  Read tests on
the other drives succeeded; however, I'm still getting stuff like:

journal_bmap: journal block not found at offset 7180 on lvm(58,0)

and

/dev/site/lvol1: recovering journal
JFS: bad block at offset 7182

How can I figure out which PV is holding block 7180 and 7182?  

-- 
"I think a lot of the basis of the open source movement comes from
  procrastinating students..."
	-- Andrew Tridgell <http://www.linux-mag.com/2001-07/tridgell_04.html>

Re: [linux-lvm] mapping blocks to PVs?

[Andreas Dilger]

On Feb 23, 2002  16:45 -0500, Andres Salomon wrote:
> I've got a LV that keeps corrupting, made up of several IDE disks.  I've
> already discovered one of the drives to be flaking out, after various
> read tests on the device gave BadCRC/IRQ timeout errors.  Read tests on
> the other drives succeeded; however, I'm still getting stuff like:
> 
> journal_bmap: journal block not found at offset 7180 on lvm(58,0)
> 
> and
> 
> /dev/site/lvol1: recovering journal
> JFS: bad block at offset 7182
> 
> How can I figure out which PV is holding block 7180 and 7182?  

I think "lvdisplay -v" or "pvdisplay -v" or similar will show the _sector_
numbers for each PE.  Note that for ext3, the block numbers reported are
4kB blocks (usually, check with dumpe2fs -h), so you are probably looking
for sectors 7180*8 and 7182*8.

Cheers, Andreas
--
Andreas Dilger
http://sourceforge.net/projects/ext2resize/
http://www-mddsp.enel.ucalgary.ca/People/adilger/

Re: [linux-lvm] mapping blocks to PVs?

[Andres Salomon]

Ok, assuming the block/sector stuff is correct:

   --- Physical extents ---
   PE    LV                        LE      Disk sector
   00000 /dev/site/lvol1           00000   33144
   00001 /dev/site/lvol1           00001   65912

badblocks(2) is not picking up those blocks as being bad; this means I
need to manually remove the first LE from this LV.  I assume I can
do something similar to `pvmove -n /dev/site/lvol1:0 /dev/hdf1 /dev/hdb1`
(hdf1 is the one causing problems, but I can't simply remove the PV from the
VG w/out deleting lots of stuff).  However, how do I lvreduce from the
beginning of the LV?  Is that possible?

On Sun, Feb 24, 2002 at 12:01:43AM -0700, Andreas Dilger wrote:
> 
> On Feb 23, 2002  16:45 -0500, Andres Salomon wrote:
> > I've got a LV that keeps corrupting, made up of several IDE disks.  I've
> > already discovered one of the drives to be flaking out, after various
> > read tests on the device gave BadCRC/IRQ timeout errors.  Read tests on
> > the other drives succeeded; however, I'm still getting stuff like:
> > 
> > journal_bmap: journal block not found at offset 7180 on lvm(58,0)
> > 
> > and
> > 
> > /dev/site/lvol1: recovering journal
> > JFS: bad block at offset 7182
> > 
> > How can I figure out which PV is holding block 7180 and 7182?  
> 
> I think "lvdisplay -v" or "pvdisplay -v" or similar will show the _sector_
> numbers for each PE.  Note that for ext3, the block numbers reported are
> 4kB blocks (usually, check with dumpe2fs -h), so you are probably looking
> for sectors 7180*8 and 7182*8.
> 
> Cheers, Andreas
> --
> Andreas Dilger
> http://sourceforge.net/projects/ext2resize/
> http://www-mddsp.enel.ucalgary.ca/People/adilger/
> 
> 
> _______________________________________________
> linux-lvm mailing list
> linux-lvm@sistina.com
> http://lists.sistina.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html

-- 
"I think a lot of the basis of the open source movement comes from
  procrastinating students..."
	-- Andrew Tridgell <http://www.linux-mag.com/2001-07/tridgell_04.html>

Re: [linux-lvm] mapping blocks to PVs?

[Heinz J . Mauelshagen]

On Sun, Feb 24, 2002 at 05:58:32PM -0500, Andres Salomon wrote:
> Ok, assuming the block/sector stuff is correct:
> 
>    --- Physical extents ---
>    PE    LV                        LE      Disk sector
>    00000 /dev/site/lvol1           00000   33144
>    00001 /dev/site/lvol1           00001   65912
> 
> badblocks(2) is not picking up those blocks as being bad; this means I
> need to manually remove the first LE from this LV.  I assume I can
> do something similar to `pvmove -n /dev/site/lvol1:0 /dev/hdf1 /dev/hdb1`

Yes.

> (hdf1 is the one causing problems, but I can't simply remove the PV from the
> VG w/out deleting lots of stuff).  However, how do I lvreduce from the
> beginning of the LV?  Is that possible?

No, just from the end.

Regards,
Heinz    -- The LVM Guy --

> 
> On Sun, Feb 24, 2002 at 12:01:43AM -0700, Andreas Dilger wrote:
> > 
> > On Feb 23, 2002  16:45 -0500, Andres Salomon wrote:
> > > I've got a LV that keeps corrupting, made up of several IDE disks.  I've
> > > already discovered one of the drives to be flaking out, after various
> > > read tests on the device gave BadCRC/IRQ timeout errors.  Read tests on
> > > the other drives succeeded; however, I'm still getting stuff like:
> > > 
> > > journal_bmap: journal block not found at offset 7180 on lvm(58,0)
> > > 
> > > and
> > > 
> > > /dev/site/lvol1: recovering journal
> > > JFS: bad block at offset 7182
> > > 
> > > How can I figure out which PV is holding block 7180 and 7182?  
> > 
> > I think "lvdisplay -v" or "pvdisplay -v" or similar will show the _sector_
> > numbers for each PE.  Note that for ext3, the block numbers reported are
> > 4kB blocks (usually, check with dumpe2fs -h), so you are probably looking
> > for sectors 7180*8 and 7182*8.
> > 
> > Cheers, Andreas
> > --
> > Andreas Dilger
> > http://sourceforge.net/projects/ext2resize/
> > http://www-mddsp.enel.ucalgary.ca/People/adilger/
> > 
> > 
> > _______________________________________________
> > linux-lvm mailing list
> > linux-lvm@sistina.com
> > http://lists.sistina.com/mailman/listinfo/linux-lvm
> > read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html
> 
> -- 
> "I think a lot of the basis of the open source movement comes from
>   procrastinating students..."
> 	-- Andrew Tridgell <http://www.linux-mag.com/2001-07/tridgell_04.html>
> 
> _______________________________________________
> linux-lvm mailing list
> linux-lvm@sistina.com
> http://lists.sistina.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html

*** Software bugs are stupid.
    Nevertheless it needs not so stupid people to solve them ***

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Heinz Mauelshagen                                 Sistina Software Inc.
Senior Consultant/Developer                       Am Sonnenhang 11
                                                  56242 Marienrachdorf
                                                  Germany
Mauelshagen@Sistina.com                           +49 2626 141200
                                                       FAX 924446
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Re: [linux-lvm] mapping blocks to PVs?

[Andreas Dilger]

On Feb 24, 2002  17:58 -0500, Andres Salomon wrote:
> Ok, assuming the block/sector stuff is correct:
> 
>    --- Physical extents ---
>    PE    LV                        LE      Disk sector
>    00000 /dev/site/lvol1           00000   33144
>    00001 /dev/site/lvol1           00001   65912
> 
> badblocks(2) is not picking up those blocks as being bad; this means I
> need to manually remove the first LE from this LV.  I assume I can
> do something similar to `pvmove -n /dev/site/lvol1:0 /dev/hdf1 /dev/hdb1`
> (hdf1 is the one causing problems, but I can't simply remove the PV from the
> VG w/out deleting lots of stuff).  However, how do I lvreduce from the
> beginning of the LV?  Is that possible?

Actually, judging by the error, I think that this error is in the layout
of the journal itself instead of a disk error.  If it was a disk error
you would also be getting IDE errors, and badblocks would detect it.
The "bad block at offset" message happens when it is doing a bmap on the
journal, and hasn't even started the read.

I would suggest just removing the journal (tune2fs -O^has_journal <LV>) and
then re-creating it (tune2fs -j <LV>) and the problem should go away.

Cheers, Andreas

> On Sun, Feb 24, 2002 at 12:01:43AM -0700, Andreas Dilger wrote:
> > On Feb 23, 2002  16:45 -0500, Andres Salomon wrote:
> > > I've got a LV that keeps corrupting, made up of several IDE disks.  I've
> > > already discovered one of the drives to be flaking out, after various
> > > read tests on the device gave BadCRC/IRQ timeout errors.  Read tests on
> > > the other drives succeeded; however, I'm still getting stuff like:
> > > 
> > > journal_bmap: journal block not found at offset 7180 on lvm(58,0)
> > > 
> > > and
> > > 
> > > /dev/site/lvol1: recovering journal
> > > JFS: bad block at offset 7182
--
Andreas Dilger
http://sourceforge.net/projects/ext2resize/
http://www-mddsp.enel.ucalgary.ca/People/adilger/

Re: [linux-lvm] mapping blocks to PVs?

[Andres Salomon]

That did the trick, thank you :)

On Sun, Feb 24, 2002 at 09:59:21PM -0700, Andreas Dilger wrote:
> 
> On Feb 24, 2002  17:58 -0500, Andres Salomon wrote:
> > Ok, assuming the block/sector stuff is correct:
> > 
> >    --- Physical extents ---
> >    PE    LV                        LE      Disk sector
> >    00000 /dev/site/lvol1           00000   33144
> >    00001 /dev/site/lvol1           00001   65912
> > 
> > badblocks(2) is not picking up those blocks as being bad; this means I
> > need to manually remove the first LE from this LV.  I assume I can
> > do something similar to `pvmove -n /dev/site/lvol1:0 /dev/hdf1 /dev/hdb1`
> > (hdf1 is the one causing problems, but I can't simply remove the PV from the
> > VG w/out deleting lots of stuff).  However, how do I lvreduce from the
> > beginning of the LV?  Is that possible?
> 
> Actually, judging by the error, I think that this error is in the layout
> of the journal itself instead of a disk error.  If it was a disk error
> you would also be getting IDE errors, and badblocks would detect it.
> The "bad block at offset" message happens when it is doing a bmap on the
> journal, and hasn't even started the read.
> 
> I would suggest just removing the journal (tune2fs -O^has_journal <LV>) and
> then re-creating it (tune2fs -j <LV>) and the problem should go away.
> 
> Cheers, Andreas
> 
> > On Sun, Feb 24, 2002 at 12:01:43AM -0700, Andreas Dilger wrote:
> > > On Feb 23, 2002  16:45 -0500, Andres Salomon wrote:
> > > > I've got a LV that keeps corrupting, made up of several IDE disks.  I've
> > > > already discovered one of the drives to be flaking out, after various
> > > > read tests on the device gave BadCRC/IRQ timeout errors.  Read tests on
> > > > the other drives succeeded; however, I'm still getting stuff like:
> > > > 
> > > > journal_bmap: journal block not found at offset 7180 on lvm(58,0)
> > > > 
> > > > and
> > > > 
> > > > /dev/site/lvol1: recovering journal
> > > > JFS: bad block at offset 7182
> --
> Andreas Dilger
> http://sourceforge.net/projects/ext2resize/
> http://www-mddsp.enel.ucalgary.ca/People/adilger/
> 
> 
> _______________________________________________
> linux-lvm mailing list
> linux-lvm@sistina.com
> http://lists.sistina.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html

-- 
"I think a lot of the basis of the open source movement comes from
  procrastinating students..."
	-- Andrew Tridgell <http://www.linux-mag.com/2001-07/tridgell_04.html>