05-17-2012 08:59 AM
I apologize if these questions have been asked before but I have limited time to search the forum. We have a NetApp MetroCluster and are scheduled to upgrade our Brocade 5100 switch firmware from 6.1.1.a to 6.3.2b2 which is supported on the MC support matrix. I know we have to do a two step upgrade (using 6.2.2b as a step), and we will also need to implement single loop storage zones with 6.3.2b2 (which I should have all the commands down for), but was wondering the following:
I know I need to to do one switch - one fabric at a time, so should I:
1. Do SW1 (both step upgrades), add the new zoning configs, then move on to SW3? If so, does my saved config file need to be loaded after each upgrade or will it be persistent?
2. Do SW1 to 6.2.2b, then SW3, then SW2, then SW4 and after these complete repeat the process with 6.3.2b2. Once again, will my switch configs remain persistent after each step?
Please drop a line if you have any answers to these questions, and thanks in advance.
05-18-2012 01:34 AM
Hi, these new zone configs - are they some kind of TI zone or just normal zones?
Why not do them before you do the upgrades then?
I suppose one reason would be that you don't want to zone in the metro-cluster while you're on an older FabOS.
What is this saved config file that you want to load? Is it the file from 'configdownload'?
Your config should stay after all FabOS upgrades. You only need to add zones once per fabric, and the zones in one fabric are hopefully not the same as in the other fabric (different wwns).
Option 2 of yours sounds a bit safer. If you want you could even wait a bit at 6.2.2b (why not 6.2.2f? - the latest) and see that everything is stable.
05-18-2012 07:12 AM
No, these aren't TI zones. Prior to 6.3.1c, with our MetroCluster, only two zones were needed - one that only had the heartbeat connection between heads, and the other that had all other FC ports (connections to from the head and each disk shelf loop). After 6.3.1c, multiple zones need to be created as each disk loop port along with the filer head fc connectivity ports have to have their own zone for example:
zone create "STOR_1_6", "1,1;1,2;3,1;3,2;1,6" <-- this zone has port 1&2 from sw1 & sw3 (fc connections from the head), along with disk loop port 6.
I can create the zones, but cannot implement them until we are at 6.3.2b2 (which is certified by NetApp and on their MetroCluster support matrix). I unfortunately cannot stay at 6.2.2f as it is not on the support matrix.
That is indeed the 'configdownload' saved config file I was referencing. I was just wondering if it would be persistent during our upgrades.
Thanks for adding your input. I think I have a pretty decent grasp on how I am going to proceed, but thought I would bounce it off this forum.
05-18-2012 07:49 AM
Ah, interesting to see how these are configured!
About the configdownloads, indeed between some versions the configdownload is not compatible (I believe it had something to do with virtual fabrics). But the configdownload should not be necessary to run, unless something really bad happens.