More liebert fun

We've been trying to get Liebert's MuliLink to work the way we want it to. And to do that we've needed to purchase three separate license keys. This was done in three separate orders since it wasn't clear what keys we needed. We shouldn't have to go through this again since we've not bought them all.

Anyway.

Today we got the MultiLink Advanced disk. So I cheer, and pop the CD in. Yep, installers for all their platforms. Whee. So I upload the Linux one to the linux server currently behaving as our UPS<->Universe interface and try to upgrade.

Nope. Seems I needed a specific file in a specific place. The file is in the ldconfig info, but not un /usr/lib/ where it has to be. So I create a symlink for it to get the installer to work. It gets past that point, extracts the internal archive, and bombs elsewhere. Grumbling, I see if I can get my hands on the internal archive to see what kind of manual voodoo I can do. And what do I see, but this directory:

/tmp/ML/licenses/

with a file named "license.adv". Oh ho! go I. You see, "license.key" allows you to install MultiLink elsewhere and have this install give info to those other installs (the "<->" bit of the interface, i.e. needed). And "license.net" allows you to shut those installs down (sold in license batches). So I presumed that "license.adv" was that unlocked the "Advanced" features that include the very useful bit known as e-mail notifications. So I suck that down to my workstation.

Then I notice something odd in the names of the installer. It seems the installer on the CD is build #7, and what I have already installed (downloaded from Liebert) is build #9. So my software is newer than what they have now. Huh.

So I launch the viewer, and go into the license upgrade screen, and feed it the license.adv file. It takes it. I restart the viewer, and all the advanced features are unlocked. No need to try and get the thing to install! Yay!

Advanced will be very useful on our Novell cluster, since we need to issue a "cluster down" command if a UPS-initiated shutdown is called for. Otherwise, all the cluster resources would try to fail over to nodes that are also trying to go down, and that can't be good. So we'll be able to give it a script. We finally, finally, have what we wanted back in August of last year.