Unsolved

This post is more than 5 years old

7023

March 2nd, 2010 11:00

AIX VIOs with NPIV

I am working on configuring a redundant AIX VIO with NPIV.  Here is the setup:

Pair of AIX VIO servers.  Each server has a physical dual port HBA (5735) with the ports connected to redundant fabrics.

We will be configuring 4 virtual fibre channel adapters for each VIO Client(2 on each VIO server, 1 for each HBA port).  Each virtual fibre channel adapter creates 2 WWPNs.

Is this the correct process:

1. Zone all 8 WWPNs to the target storage (4 on one fabric to a storage port and 4 on the other fabric to a redundant storage port).

2. Mask any LUNs on the target storage array to all 8 WWPNs.

3. ODM and Powerpath installed on each client.


I think I am going to go into WWPN overload.  8 WWPNs for each client.  I have 14 clients just to start...  *sigh*   Thought this was supposed to be easier...

AJE

4 Operator

 • 

2.1K Posts

March 2nd, 2010 13:00

Hmmmm..... sounds like you might need to get some answers from IBM around that. I can't understand how assigning a second virtual WWPN would help with "mobility" or client failover. Then again, I''m no expert on VIO or NPIV (yet anyway).

If no one here can help you, please let us know what you find out in the end

March 2nd, 2010 13:00

The 2 WWPNs per virtual adapter come straight from the RedBook.  What they dont explain very well is the zoning for the WWPNs.  All it mentions is the second WWPN enables the mobility of the client for failover, but doesn't go much deeper...

AJE

4 Operator

 • 

2.1K Posts

March 2nd, 2010 13:00

I'm not running NPIV for anything in our shop yet, but I have a basic understanding of the technology. Almost everything you wrote sounds right to me, but I have one question (that could make a big difference.

You state that when you create the 4 virtual HBAs for each VIO Client it will create "2 WWPNs" for each one (for a total of 8 WWPMs per VIO Client). Why would each virtual HBA create 2 WWPNs instead of just one? Even if this is true (for whatever reason) you could just ignore one of the WWPNs from each and still only work with 4 WWPNs per VIO Client. Sort of like putting in a dual port HBA when you don't have to. Doesn't mean you have to use both ports, right?

1 Message

April 30th, 2010 03:00

A bit of a johnny come lately on this one but nevermind...

For each NPIV connection there are two WWPNs created. Only the first WWPN should be used for SAN zoning and storage allocation.

When you create the LPAR the WWPNs are allocated and using the lssyscfg cmmand you can get them from the HMC and script all zong etc.

April 30th, 2010 06:00

You will need to zone and allocate both WWN for each NPIV connection, if you intend to have the guest LPAR failover to another VIO.  Since this would greatly increase the zoning and allocation tasks, we chose to use the virtual scsi passthrough capabilities of the VIO instead of NPIV.

Aaron

March 10th, 2011 23:00

Hi,

You should only Zone the Client WWN.

Just remind those advice for NPIV configuration [Form IBM Presentation]:

Things to consider

  • WWPN pair is generated EACH time you create a VFC. NEVER is re-created or re-used. Just like a real HBA.
  • If you create a new VFC, you get a NEW pair of WWPNs.
  • Save the partition profile with VFCs in it. Make a copy, don’t delete a profile with a VFCin it.
  • Make sure the partition profile is backed up for local and disaster recovery! Otherwise you’ll have to create new VFCs and map to them during a recovery.
  • Target Storage SUBSYSTEM must be zoned and visible from source and destination systems for LPM to work.
  • Active/Passive storage controllers must BOTH be in the SAN zone for LPM to work
  • Do NOT include the VIOS physical 8G adapter WWPNs in the zone
  • You should NOT see any NPIV LUNs in the VIOS
  • Load multi-path code in the client LPAR, NOT in the VIOS
  • Monitor VIOS CPU and Memory – NPIV impact is unclear to me at this time
  • No ‘passthru’ tunables in VIOS
No Events found!

Top