For a dot1q trunk your port-channel interface needs to reflect mode trunk. You might want to enable spanning-tree on the port-channel interface as well.
VLAN tagged interfaces should be made on the top level of the vif as aborzenkov points out. He's also correct in that NetApp doesn't support untagged (native) vlans. Assuming your native vlan is probably 1 or at least isn't 650 or 652, you don't need to worry about it.
So, for vlans 650 and 652 you'll need to add the following to the rc files before the ifconfig statements:
vlan create svif0 650 652
Furthermore, you'll then need to apply the IP addresses to the vlan tagged interfaces.
ifconfig svif0-650 ....
ifconfig svif0-652 ....
You won't be able to do this with the current configuration in place. Either you need to do a:
ifconfig svif0 0.0.0.0 down
or you need to edit the /etc/rc and /etc/hosts files and reboot or cf takeover/giveback in each direction. If you're going to use `hostname` in the ifconfig statement, you'll want to edit /etc/hosts to have hostname bos-equnas-1a-svif0-650 and hostname bos-equnas-1a-svif0-652 and similar for the other controller. Make sure you then also append the vlan id to in the ifconfig statements.
Finally, the reason I advise customers to put vlan tagged interfaces into place for even a single vlan is that it's non-disruptive to add additional vlans once the dot1q trunk is setup. If you want change from access mode to dot1q non-disruptively, you can use cf takeover and giveback to do so for everything but CIFS - that's if you don't make any mistakes in /etc/rc. My opinion is do the hard work and testing up front before it goes into production. Every additional vlan after that will be following the form you setup for the first vlan.
