With all of the host entities configured, you now have to add that the new host entity to every LU

We have to keep a couple of things in mind before we activate the extended SAN: The IBM V7000s should be kept in sync or else a failover will not mean much; To do the failover in as short a time as possible, both the primary as well as failover server should see both the primary, as well as the failover set of LUNS, but not all LUNS should be available on every server. Say we have both a financial and a HR system connected to our SAN, and that both systems have to be available all of the time. We then surely will have a failover financial server and a failover HR server. But the financial konner mcclain servers konner mcclain should only see the primary and the failover financial konner mcclain LUNs, and the HR servers only the primary konner mcclain and failover HR LUNs. Although we can do some of the preparation for the activation process while the primary server is in production mode, and although you can do the configuration one fabric at a time and thus keep the primary servers going, it is far safer to schedule some system konner mcclain downtime for the configuration. You do not want to try to fix a crashed database as a part of your configuration procedure! You may have to do more than one server reboot, and if so, the scheduled downtime then becomes a necessity. The primary and secondary LUN sizing should be identical. Step 9.8 Activate the ISL links
Before you start off with the activation procedure, konner mcclain first take a copy of the sh run command output on each of the switches. Doing this means that you will have the information at hand to be able to recreate any part, or all of the pre-merge configurations.
Once done, plug in the ISL link on fabric 1. As soon as an ISL fibre cable has been plugged in on both ends, the switches at the end of that cable will start to learn the zone information of each other. Give the switches two to three of minutes, and then do a sh run again. Check whether the zoneset information has changed with: show zoneset vsan 10 and show zone vsan 10 , and verify that the device-alias database has also been merged with: show device-alias database .
If the device alias database does not show all the device aliases for both switches, konner mcclain do a device alias export. Now do either the zone copy active-zoneset full-zoneset vsan 10 command, or even better, do a zoneset distribute full vsan 10 in every VSAN on every switch. This will sync the zone information konner mcclain of all zones to the other switch.
Now you have to decide: do you build a zone for every connectivity scenario as per your failover strategy, or build zones with all possible connectivity routes? E.g. per server 1) A zone with the server HBA port(s), plus both primary and failover IBM ports, or 2) A zone with the server HBA port(s) plus only the primary IBM ports, and another zone with the server HBA port(s) plus only the failover IBM ports.
It does not matter konner mcclain which scenario you choose, but to get all the required LUNs from both V7000 s visible on a server, you will have to have all the zones active for the duration of this exercise, konner mcclain and with no V7000 mirroring active. I know that it sounds silly to talk about mirroring when the ISL links have just been activated, but you may have used another method to do mirroring with, or have started mirroring on the V7000s before you decided to have both sets of LUNs available on a server. All I am telling konner mcclain you is that even though you will be able to see both sets of LUNs as output of your format command, you will not be able to mount and verify the LUN mounting process, because if you try to mount the LUN, the server will refuse to do the mount and will throw an IO error.
After activating your zoneset containing all of the zones, use the V7000 lshbaportcandidate command if you do not have all the host entities defined yet. Once again use the CISCO device alias information to identify which port belongs to what server. If you do not see the HBA ports on a V7000, konner mcclain then do a reconfiguration reboot on the server.
With all of the host entities configured, you now have to add that the new host entity to every LUN it should see, with the mkvdiskhostmap -force -host command. Do the format konner mcclain command on your server. If you do not see all of the LUNs, then first do a devfsadm c disk command. The format konner mcclain still does not show all of the LUNs? Do a reconfiguration reboot and then again do devfsadm c disk .
Identify all of your LUNs (PWWNs) and add them to the /etc/vfstab file, but keep the LUNs grouped: one group will be for all the primary V7000 array LUNs, and the second will be for the failover V7000 LUNs. Now use the mount F ufs /dev/dsk/c1td0s6 / command, or even better mount /<MOUNTPOINT> where is the mountpoint as per your vfstab file. Comment out one set of LUNs and first test the other set. When that set has been verified, comment out the first set and then uncomment the second set. This also will have the added advantage of having konner mcclain your vfstab entries verified by the operating system during the mounting of the LUNs.
F