Disaster Recovery scripts

[Donald C Nelson]

Does anyone have any information on Disaster Recovery scripts that would allow me to recreate my VG's and LV's at a DR site without having to manually re-create them?

[Rita C Workman]

Well I use MC/SG (Continental Cluster) to fail over from site to site...

Rgrds,
Rita

[Donald C Nelson]

Thanks for the reply but I won't be using MC/Service Guard, but just bringing my ignite tapes and system info from the home server to the hot site.

[Patrick Wallek]

There is really no way to automate that. When recreating LVs and VGs at a DR site you most likely will not know the device files for the disks ahead of time. You may also have different disk sizes which will influence the number of disks per VG and possibly other VG parameters.

For VG creation you could just write a simple script and just add the values for the disk device files once you are at the DR site.

LV creation can be easily scripted as well, since that just depends on the VGs. Just do a series of lvcreate and newfs commands and you are set for the LVs.

[Ken Penland_1]

some prep work would need to be done...here is what we do:

Whenever there is a change to your volume group info:
strings /etc/lvmtab | grep -v vg00 | grep dev | grep -v dsk | cut -f3 -d"/" > /midtier/vgs

then:
for i in `cat /midtier/vgs`
do
/usr/sbin/vgexport -m /midtier/${i}mapfile -p -s -v $i
sleep 2
done

this will create a mapfile of all your lvols n stuff...
on the new box:

for i in `cat /midtier/vgs`
do
mkdir /dev/$i
mknod /dev/$i/group c 64 0x0${count}0000
echo "ok..importing $i"
/usr/sbin/vgimport -m /midtier/${i}mapfile -s -v $i
/usr/sbin/vgchange -a y $i
count=`expr $count + 1`
sleep 3
done

[Donald C Nelson]

I will go with the my automated scripts plus responses from the forum.