Brocade’s flavor of NPIV is called Access Gateway. It’s a way to dumb down the switch and make it more of a pass-through device. When AG is enabled, the switch makes much less routing or switching decisions, and passes all the traffic to an upstream switch. The upstream switch ports switch to F ports, and the “egress” ports on the NPIV switch become N ports.
Trunking doesn’t really work, in AG mode. The switch will load balance between uplinks, and fail traffic to another link in the event of a link failure, but port bandwidth aggregation isn’t done.
You can’t cascade a switch off of a switch that has AG enabled, so it’s only good for edge switches.
I don’t think it works with hard zoning, but I’m not going to test it.
The whole switch becomes part of the AD of the upstream switch port.
The upshot for us is that the switch doesn’t care about fabric changes. It doesn’t recalculate all of the available routes ever time there is a fabric rebuild, which should reduce the interruption to the fabric every time there’s a rebuild. We had problems with errors on ISLs causing a fabric rebuild, and with so many edge switches it was causing issues for the hosts.
To enable AG, first switch to the AD that the upstream port, and this switch, is a member of. In my case it was AD1. And, your upstream port has to have AG enabled (per port on the upstream switch), my 48Ks all came with it enabled on every port.
Check the AG mode:
sanswa:AD1:root> ag --modeshow Access Gateway mode is NOT enabled
Enable AG mode:
he10enc1sanswa:AD1:root> switchdisable he10enc1sanswa:AD1:root> ag --modeenable WARNING: Access Gateway mode changes the standard behavior of the switch. Please check Access Gateway Administrator's Guide before proceeding. Enabling agmode will remove all the configuration data on the switch including zoning configuration and security database. Please backup your configuration using configupload. This operation will reboot the switch. Do you want to continue? (yes, y, no, n): [no] yes Access Gateway mode was enabled successfully Switch is being rebooted...
The switch does a fastboot, which will take several minutes.
After the switch is up, the upstream port looks like:
... 29 2 13 011d00 id N4 Online F-Port 3 NPIV public ...
SanSW2:AD1:admin> portshow 2/13 portName: portHealth: HEALTHY Authentication: None portDisableReason: None portCFlags: 0x1 portFlags: 0x24b03 PRESENT ACTIVE F_PORT G_PORT U_PORT NPIV LOGICAL_ONLINE LOGIN NOELP LED ACCEPT FLOGI portType: 10.0 portState: 1 Online portPhys: 6 In_Sync portScn: 32 F_Port port generation number: 92 portId: 011d00 portIfId: 43220012 portWwn: 20:1d:00:05:1e:36:3f:8a portWwn of device(s) connected: 21:01:00:1b:32:a9:78:e4 21:01:00:1b:32:b0:9a:42 20:00:00:05:1e:88:71:c2 Distance: normal portSpeed: N4Gbps
Notice the WWNs listed. One is the N port on the switch. The other two are hosts connected to that switch.