Fwd: The Redhat linux reboot puts me into a maintenance shell because disks are discovered and name differently.

[unix syzadmin]

---------- Forwarded message ----------
From: unix syzadmin <unixsyzadmin at gmail.com>
Date: Thu, Mar 15, 2012 at 1:02 AM
Subject: Re: The Redhat linux reboot puts me into a maintenance shell
because disks are discovered and name differently.
To: Barry Brimer <lists at brimer.org>


This problem was resolved by using "rdblacklist=lpfc" in the kernel line
for /etc/grub.conf and rebuilding the initramfs using dracut.
The theory is when initramfs is used during boot phase the lpfc module is
not used; so none of SAN disks attached to HBA are discovered.


** excerpt from /etc/grub.conf ***
title RHEL 6.1 (2.6.32-131.21.1.el6.x86_64)
root (hd0,0)
 kernel /vmlinuz-2.6.32-131.21.1.el6.x86_64 ro root=/dev/rootvg/root.vol
rd_LVM_VG=rootvg rd_NO_LUKS rd_NO_MD rd_NO_DM rdblacklist=lpfc
LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 KEYBOARDTYPE=pc KEYTABLE=us
crashkernel=auto
 initrd /initramfs-2.6.32-131.21.1.el6.x86_64.img

** end **
# dracut --force /boot/initramfs-$(uname -r).img $(uname -r)


On Tue, Feb 28, 2012 at 8:05 PM, unix syzadmin <unixsyzadmin at gmail.com>wrote:

> Thanks,
>
> I am running RHEL 6.1.
>
> # more /etc/redhat-release
> Red Hat Enterprise Linux Server release 6.1 (Santiago)
>
> # uname -r
> 2.6.32-131.21.1.el6.x86_64
>
> ***** Some excerpts from /etc/fstab & fdisk  ****
>
> /dev/mapper/rootvg-root.vol / ext4 defaults 1 1
> UUID=9ac96796-d14a-4013-8e5b-a83da60dc95c /boot ext4 defaults 1 2
> /dev/mapper/rootvg-home.vol /home ext4 defaults 1 2
> /dev/mapper/rootvg-opt.vol /opt ext4 defaults 1 2
> /dev/mapper/rootvg-sysinfo.vol /sysinfo ext4 defaults 1 2
> /dev/mapper/rootvg-tmp.vol /tmp ext4 defaults 1 2
> /dev/mapper/rootvg-usr.vol /usr ext4 defaults 1 2
> /dev/mapper/rootvg-var.vol /var ext4 defaults 1 2
> /dev/mapper/rootvg-swap.vol swap swap defaults 0 0
> tmpfs /dev/shm tmpfs defaults 0 0
> devpts /dev/pts devpts gid=5,mode=620 0 0
> sysfs /sys sysfs defaults 0 0
> proc /proc proc defaults 0 0
>
> # fdisk -l /dev/sda
>
> Disk /dev/sda: 598.9 GB, 598879502336 bytes
> 255 heads, 63 sectors/track, 72809 cylinders
> Units = cylinders of 16065 * 512 = 8225280 bytes
> Sector size (logical/physical): 512 bytes / 512 bytes
> I/O size (minimum/optimal): 512 bytes / 512 bytes
> Disk identifier: 0x000957bb
>
>    Device Boot      Start         End      Blocks   Id  System
> /dev/sda1   *           1          66      524288   83  Linux
> Partition 1 does not end on cylinder boundary.
> /dev/sda2              66       72810   584317952   8e  Linux LVM
>
>
> **** End ****
>
> Are you suggesting to use logical volume / file-system UUID's instead of
> the device mapper names in the /etc/fstab?
> What about the root & swap LV/file-sytem mentioned in /etc/grub.conf? Can
> we use UUID there as well?
>
>
> # grep kernel /etc/grub.conf
> kernel /vmlinuz-2.6.32-131.21.1.el6.x86_64 ro
> root=/dev/mapper/rootvg-root.vol rd_LVM_LV=rootvg/root.vol
> rd_LVM_LV=rootvg/swap.vol rd_NO_LUKS rd_NO_MD rd_NO_DM LANG=en_US.UTF-8
> SYSFONT=latarcyrheb-sun16 KEYBOARDTYPE=pc KEYTABLE=us crashkernel=auto
>
>
>
> Thanks,
>
> On Tue, Feb 28, 2012 at 7:21 PM, Barry Brimer <lists at brimer.org> wrote:
>
>> > The problem; I guess is that the system is discovering & naming the
>> disks
>> > in different order everytime the server is rebooted.  For example the
>> > /dev/sda which represents the raid5 device created on local internal SAS
>> > disk is discovered as /dev/sdew ... and I guess this is the reason I am
>> > being dropped into the maintenance shell.  I tell this because after I
>> > login to the maintenance shell I see that the /dev/sda represents a 500G
>> > SAN disk instead of a 600G internal disk.  The internal disk is named
>> > /dev/sdew and  if I disconnect the fibre cables; the server boots just
>> > fine.
>> >
>> > Is there a way to make sure that the disks are named consistently across
>> > reboots?
>> > Please suggest or point me in the right direction.
>>
>> Your /etc/fstab and /etc/grub.conf (really /boot/grub/grub.conf) should
>> use
>> LABELs to identify filesystems, not devices.  RHEL 6 also allows you to
>> specify
>> filesystems by UUID.  I don't know if that works in RHEL 5 or not.
>>
>
>