Re: FC3 can't find LVM volumes, boot fails, kernel panics

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



I have some more information.  Also, I'm no longer sure that this
isn't some sort of mkinitrd(8) problem, and may be related to another
thread in this mailing list about the ext3 modules not being in the
SMP kernel.

I'm now thinking I have a similar problem, except my missing module
appears to be "megaraid".  This is HARDWARE RAID, not software RAID,
so I'm not exactly sure why that module is needed (rather than the
generic scsi_sd module).

Even booting the rescue disk and trying to re-run the mkinitrd failed,
saying it couldn't resolve the module dependency for megaraid.


To debug what was going on I manually extracted the initrd.  First,
booting the rescue CD.

  rescue$  cd /mnt/sysimage/boot
  rescue$  mkdir tmpimg
  rescue$  cd tmpimg
  rescue$  gunzip -c ../initrd-2.6.9-1.667.img | cpio -idmv

At that point I edited the "init" file to,

  1. Comment out the line "setquiet"
  2. Change the "lvm vgscan" to "lvm vgscan -v"

Edit the etc/lvm/lvm.conf file and turned verbose=1 and level=3 in the
log() function.

Repackage up a new image.  First mv the original one out of the way to save it.

  rescue$  find . | cpio --quiet -c -o >../newinitrd
  rescue$  cd ..
  rescue$  gzip -9 < newinitrd > initrd-2.6.9-1.667.img

REBOOT, non-rescue

Then on reboot I get a lot more out.
It gets to loading "aic7xxx.ko" and pauses a long time.  Then a bunch
of output goes by, most of which I loose.  The ending looks something
like this:

  vgmknodes     setting global/locking_type to 1
  vgmknodes     setting global/loging_dir to /var/log/lvm
  vgmknodes     File-based locking enabled.
  vgmknodes   Finding all logical volumes
  vgmknodes  No volume groups found
  vgmknodes      Dumping persistent devide cache to /etc/lvm/.cache
  Creating root device
  Mounting root filesystem
  ....
  (eventually kernel panic because no / filesystem)


In rescue mode again I probe the /proc/scsi and find that the hardware
RAID (which is a Dell PowerEdge Expandable RAID using a PERC 4/DC
controller) shows up as being

   MegaRAID Model: LD0 RAID0 39760R Rev 3.41
   MegaRAID Model: LD1 RAID0 39760R Rev 3.41

The hardware RAID is configured to have two logical drives, each in RAID-0.
The SCSI device mappings are

Bus 1 Dev 6 SCSI-ID 0 LUN 0
  /dev/sda1 - Dell Utility partition (type 0xde)
  /dev/sda2 - /boot partition (type 0x83)
  /dev/sda3 - LVM partition for VG "vg00" (type 0x8e)

Bus 1 Dev 6 SCSI-ID 1 LUN 0
  /dev/sdb1 - LVM partition for VG "vg00" (type 0x8e)

So this most certainly looks like a megaraid module issue.   But yet
the rescue disc boots and finds it just fine.

How can I repair this?

-- 
Deron Meranda


[Index of Archives]     [Current Fedora Users]     [Fedora Desktop]     [Fedora SELinux]     [Yosemite News]     [Yosemite Photos]     [KDE Users]     [Fedora Tools]     [Fedora Docs]

  Powered by Linux