Zone config for MSA2012fc (G1)

[Paul_637]

I'm trying to understand the requirements for the above given its "active/active but not quite active/active as we know it" configuration so I need to establish exactly which WWN's to use on each of the two (redundant) SAN fabrics.

I've looked at the "How to configure soft-zones" document at http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?objectID=c01692989 but this does more to confuse than clarify.

It shows the output of "show port-wwn" on the MSA which clearly shows the WWNs it holds, but both of the SAN Switch configs (Blue & Red in the example)refer to the same pair of WWNs, referring to ports B0 & A0 only with no reference to B1 & A1.

So, what's the correct way to create an alias for a single MSA2012fc with dual controllers?

Any pointers gratefully appreciated!

Paul

Is this correct?

[Sheldon Smith]

Damn good question. I work with EVAs, but the cfgshow outputs are definitely hosed: The same ports would NOT be on both fabrics! :)

I suspect the drawing is correct: A0/B1 on one fabric, and A1/B0 on the other.

[Taurian]

Hi Paul,

If your MSA is wired according to the example, then same port cannot show up on both fabrics, as the fabrics are not even connected to each other.

A0 and B1 should be on first fabric, A1 and B0 should be on the second.

I suggest you post us the output of cfgshow from both switches so that we can compare.

Regards,
Taurian.

[Taurian]

Hi Paul,

If you are referring to the cfgshow output in the example on your link, then look closely. All three WWN for cfgshow on SW1 (Blue Fabric) and different from the three WWN for cfgshow on SW2 (Red Fabric). Compare initial numbers for the last two ports. They are unique on both switches.

>> but both of the SAN Switch configs (Blue & Red in the example)refer to the same pair of WWNs, referring to ports B0 & A0 only with no reference to B1 & A1.

In the example, the alias created on switches contain one WWN from each controller. Both switches contain exactly the same entries. This is due to the fact that it is the floating WWN. This is useful during failover scenarios.

You can see the failover example given at the link. It is quite self explanatory.

Regards,
Taurian.

[Torsten.]

Be careful, the WWN's are different in the first quad of digits (e.g. 2170-xxx... vs. 2070-xxx...). Normally you expect the difference at the end of the WWN.

[Paul_637]

@Taurian/Torsten: Indeed, the examples in the switchshow output are different, but I'm referring to the zoning config part of the example which is identical for both zones from what I can see.

So, A0 (20:78) & B0 (20:70) are the *only* WWNs that appear in either switch.

For my config therefore (and in a correctly cabled setup) should I only ever zone with the WWNs from A0/B0 ?f

Looking at the example failure of Controller A it does look this way, but why don't they just make it clear & state it as fact rather than forcing us to deduce it??

Paul

[Paul_637]

...and now I've ensured that things are connected as they should be, looking at the switch ports (via the GUI) then both the 20:78 & 20:70 ports are the ones that are presented to both switches!

Will try zoning 'em up & see what I break... :)

[Torsten.]

It should not be so hard, just include the servers HBA and one port from each array controller into a single zone - do this on both SANs.

[Paul_637]

It's not so much that it's hard, it's the fact that two of the four WWNs associated with the MSA appear within the switch (with a switch-show) so those are the ones you'd automatically assume should be configured within the alias/zone for that particular switch, but that's not the case! Only one of them should actually appear in the zone config, the other two basically aren't used at all (so why have them??).

This is very much contrary to my previous experience configuring switches for EVA's...

Anyhow, I've now configured the zones & things are working. Tomorrow I'll initiate some failovers and see if that trend continues!