BladeSystem Virtual Connect
cancel
Showing results for 
Search instead for 
Did you mean: 

Virtual Connect Fibre Channel (VC-FC) Login Re-Distribution

chuckk281
Trusted Contributor

Virtual Connect Fibre Channel (VC-FC) Login Re-Distribution

Aniekan had a Virtual Connect - Fibre Channel login question:

 

**************************

 

What are the requirements for login distribution and why can’t I enable automatic login distribution on a fabric?

I want to ensure that if an uplink port on the VC-FC goes down that server ports connected to it get redistributed to other remaining uplink ports on the module.

 

Customer has VC 8Gb 20-Port FC modules with HP BLc QLogic 8Gb FC HBA cards on BL860c and BL460c G7 blades. FW versions are:

 

OA – 3.21

VC-E – 3.15

VC-FC – 1.41

 

Two ports on each VC-FC module are connected to an external SAN switch with NPIV enabled on the ports. (Customer is using two different blades in a switch enclosure.)

 

In the “Define SAN Fabric” screen of VCM, the default (and only) choice is MANUAL login distribution.

 

Docs such as the VC CLI user guide (http://bizsupport2.austin.hp.com/bc/docs/support/SupportManual/c02255342/c02255342.pdf) suggests that it’s not support on the specific module:

 

The HP 4Gb VC-FC Module, HP Virtual Connect 4Gb FC Module, HP VC 8Gb 20-Port FC Module and HP VC 8Gb 24-Port Module support only manual login redistribution.

 

Other docs (http://bizsupport1.austin.hp.com/bc/docs/support/SupportManual/c02516247/c02516247.pdf) go further by saying it is only supported on Virtual Connect FlexFabric:

 

The default is manual. The automatic option is only available on FlexFabric modules

 

What are the requirements for automatic login distribution? Is there an option that needs to be set on the SAN switch?

Actual experiences appreciated.

 

****************************

 

Vincent replied with the answers:

 

****************************

 

As you found out, automatic login redistribution can be enabled only with FlexFabric modules, it does not depend on anything else.

However, the use case is not what you indicated (failover): in case a FC uplink fails, logins that were using it are automatically redistributed on the remaining uplink(s) regardless of automatic/manual login redistribution. It’s when the failed uplink comes *back* that the behaviour changes: with automatic login redistribution, some logins are automatically moved to the restored uplink after a delay, with manual login redistribution you need to click a button to make it happen.

 

*******************

Any other experiences on this? Comments?