RHEL4 install format problem on Dell PowerEdge w/PERC RAID

[Rick Stevens]

On Mon, 2006-07-24 at 12:34 -0400, Thomas Walter wrote:
> On Mon, 17 Jul 2006, Rick Stevens wrote:
> 
> > On Mon, 2006-07-17 at 16:09 -0400, Thomas Walter wrote:
> > > I issued the dd command (below) in rescue mode and tried the RHEL4 update 
> > > 3 install again and got the same result....... format dies at 99% with a 
> > > screenfull of error messages scrolling buy.
> > 
> > Ugh!  That's not good.  Since it's happening toward the end of the
> > format, it sounds like you have either dissimilar drives in the RAID
> > (all drives in a hardware RAID should be the same make and model) or one
> > or more of the drives in the RAID are dying.
> > 
> > By the way, we prefer bottom posting here (post your responses after
> > what you're responding to).  It makes following the logic of the
> > messages easier.
> > 
> > > I was unable to determine if RAID was running in degraded mode. All boot 
> > > messages seemed to indicate everything related to RAID is OK.
> > 
> > You need to get into the BIOS of the controller to determine that.  A
> > hardware RAID simply presents a "logical" drive to the OS.  The OS is
> > not aware of the RAID at all--all it sees is a drive.
> > 
> > If your 2400 uses the PercRAID common to Dell machines, hit "CTRL-M"
> > at the appropriate prompt after doing a reset.  Then go into "Objects"
> > and select the virtual drive created by the RAID.  Verify that it's in
> > an "optimal" state.  If not, you really should resolve that first.  You
> > should also look at all the drives involved and verify that they're the
> > same make and model.
> > 
> > If the RAID controller says it's healthy, then bring up the rescue mode
> > again and try using the badblocks utility on the drive:
> > 
> > 	# badblocks /dev/sda
> 
> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
> I did a badblocks -w /dev/sda  and there were no error messages. I was 
> able to find my RH9 install CDs and just tried them. The disk format was 
> successful and RPMS are installing! Could this problem have to do with 
> LVM? But exactly where?
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

I'm glad to hear things are working.  As to if it was LVM, it's
possible.  The "badblocks -w" does a destructive test on the drive which
wipes out all of the formatting (LVM, MBR, partition table, etc.).  As
to what specifically was causing the issue, that's rather hard to say.
It could have been LVM.

If you were upgrading from a system that used a 2.4 kernel (e.g. Fedora
Core 1, RHEL3 and anything before them) to something that uses kernel
2.6 and up (FC2 and up, RHEL4 and up), then it may actually be likely
that LVM was the culprit.  The 2.4 kernels used LVM1, 2.6 uses LVM2 and
the two (LVM1 and LVM2) are not compatible, but I would have expected
the installer to fix that or at least called your attention to it (you
can't convert from LVM1 to LVM2 without wiping out the data).

----------------------------------------------------------------------
- Rick Stevens, Senior Systems Engineer     rstevens at vitalstream.com -
- VitalStream, Inc.                       http://www.vitalstream.com -
-                                                                    -
-     If Bill Gates got a dime for every time Windows crashes...     -
-              ...oh, wait.  He does.  THAT explains it!             -
----------------------------------------------------------------------