video on demandexample:
1. creating two 20MB volumes to serve as loadshare destinations
2. creating two loadshare mirrors on vserver grvsnfs1
3. initalize-ls the mirrors
4. create a schedule with a 1 minute interval
5. create two mountpoints on linux an mount the ls-mirrors
6. create a new volume in the vserver and watch the 1 minute update delay
1.
gr::> vol create -vserver grvsnfs1 -volume rootls -aggregate gr01_aggr1 -size 20MB -type DP
gr::> vol create -vserver grvsnfs1 -volume rootls1 -aggregate gr02_aggr1 -size 20MB -type DP
2.
gr::> snapmirror create -source-cluster gr -source-vserver grvsnfs1 -source-volume root_vol -destination-cluster gr -destination-vserver grvsnfs1 -destination-volume rootls -type ls
gr::> snapmirror create -source-cluster gr -source-vserver grvsnfs1 -source-volume root_vol -destination-cluster gr -destination-vserver grvsnfs1 -destination-volume rootls1 -type ls
3.
gr::> snapmirror initialize-ls-set -source-path gr://grvsnfs1/root_vol -foreground true
4.
gr::> job schedule interval create -name 1minute -minutes 1
gr::> snapmirror modify -destination-path kp://intdest/rootls1 -schedule 1minute
*taken that nfs is setup correctly and there are 2 lifs 1 on each of the 2 nodes
you can mount the loadshare mirror on both interfaces
lifs 192.168.4.85
192.168.4.88
5.
linux: mkdir /mnt/rootls
mkdir /mnt/rootls1
mount 192.168.4.85:/ /mnt/rootls
mount 192.168.4.88:/ /mnt/rootls1
6.
gr::> vol create -vserver grvsnfs1 -volume vol4 -aggregate gr02_aggr1 -size 100m -state online -type RW
gr::> vol mount -vserver grvsnfs1 -volume vol4 -junction-path /vol4 (volume mount)
Notice: Volume vol4 now has a mount point from volume root_vol. The load sharing (LS) mirrors of volume
root_vol are scheduled to be updated at 4/4/2013 15:39:53. Volume vol4 will not be visible in the
global namespace until the LS mirrors of volume root_vol have been updated.
linux:
while true
do
ls /mnt/rootls;ls /mnt/rootls1
sleep 10
done
*vol4 will pop up after about a minute.