Hello experts,
I have installed SLES 11 SP4 on power using the following scenario
we have multipathing and we have 2 LUNS
one LUN is on data center A and one LUN is on data center B
on each LUN, i made two partitions on install via expert partitioner
/dev/mapper/LUN1_part1 - 512 MB - ext3 - /boot
/dev/mapper/LUN1_part2 - 40,5 GB - no fs - not mounted
/dev/mapper/LUN2_part1 - 512 MB - no fs - not mounted
/dev/mapper/LUN2_part2 - 40,5 GB - no fs - not mounted
then I made a software raid1 over mdadm on install time via expert partitioner
/dev/md0 - on devices /dev/mapper/LUN1_part2 and /dev/mapper/LUN2_part2
then I created a volume group vg_rootfs on /dev/md0
then I created the following logical volumes on vg_rootfs
lv_rootfs - 5 GB - ext3 - /
lv_home - 5 GB - ext3 - /home
lv_opt - 5 GB - ext3 - /opt
lv_usr - 5 GB - ext3 - /usr
lv_var - 5 GB - ext3 - /var
lv_tmp - 5 GB - ext3 - /tmp
lv_var_log - 3 GB - ext3 - /var/log
lv_swap - 2 GB - swap - swap
after installation I first wanted to test the failover by taking one LUN away from the system. So far i have NOT replicated the /boot sector via sfdisk to /dev/mapper/LUN2_part1, but I think for this scenario this should not be necessary
when I take /dev/mapper/LUN1 away from the system (where the /boot sector ist), everything works fine. I can still write files to the file system and i can still do commands in my ssh session. Then I gave the system /dev/mapper/LUN1 back. Now I wanted to test to take away the other LUN and just reboot the system - this is where I fail.
when I take away /dev/mapper/LUN2, where the /boot sector is currently NOT replicated, ssh freezes and i cannot state any commands.
How to fix this?
Best regards