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

Primary and Secondary explained....

 
duhaas
Advisor

Primary and Secondary explained....

I have two virtual connects in slots 1 and 2. I've created a network called production thats made up of the first 4 ports on each virtual connect. The ports 1-4 in bay 1 are all marked as linked/active, and ports 1-4 in bay 2 are all maked linked/standby. I just want to understand whats meant by active/standby. Does it mean that the ports in bay 2 dont have any network communication going through them??? I have always been stuck on understanding this.
4 REPLIES
Calvin Staples
Frequent Advisor

Primary and Secondary explained....

That is correct. The VCs are preventing loops so half of the ports are in standby. If you create two production networks; prod1 for ports on VC1 and prod2 for ports on VC 2 and connect NIC1 to prod1 and NIC2 to prod2 you would see both networks are active. In this scenario you would want to use the "Smart Link" option when defining the networks to assure failover of your teamed NICs on the server(s) should all the ports on one VC fail. (Note: If you use the TLB with fault tolerance on your team you will have 2 GB transmit and 1 GB receive throughput. SLB is not supported with the VCs.) Doing the network this way will provide 8 GB total throughput versus 4 GB that you would have using the one network. I hope this helps.
duhaas
Advisor

Primary and Secondary explained....

Thanks for the response, so on the backend switch, I would only want to make sure the fours ports that would make up each network are part of their own port channel on the switch side? Another question I have is, if you look @ the port mapping doc for a half height blade, according to it, the second onboard nic uses the 2nd IO slot in the chassis, so if all those ports in my 2nd virtual connect are in standby, how is it talking on the network???

Primary and Secondary explained....

Only "uplink ports" that resides on 1 VC module can be active to your backbone. All other ports on other VC-modules connected to the same network (in your situation "production") are linked/standby. The configuration state of the ports must be set to "AUTO", which allows trunking or LACP. If 1 active link will fail, VC selects the path (read Trunk or LACP-group) with highest bandwidth. This causes a failover to VC-module 2. These 4 links will become linked/active and the remaining links from VC-modules 1 become linked/standby. This works within your full VC-domain.
duhaas
Advisor

Primary and Secondary explained....

Thanks for the detailed response, I apprecaite all the help. So the idea of eth ports 3 and 4 on a full height blade mapping to ports on vc2 isnt true and the networking for those ports actually occurs on vc1 in the config i have? It seems to me to make best use of ports is to trunk ports in sets of two or four on each seperate vc module, making all ports active?