HP Proliant Server adapter teaming

rsutoratosu

Platinum Member
Feb 18, 2011
2,716
4
81
So we have an argument in the office.. maybe someone with more knowledge would know this.

As far as I know, when you team the adapter for load balancing, it does not require any additional hardware... the software creates a virtual IP and assigns 2 ip to the 2 individual nic cards.

The cisco guys in the office is arguing that you need special team aware switches in order for the adapter teaming to work properly..

Any truth in that ? we feel that they're trying to get us to buy more cisco stuff.. but I can't be 100% sure

hp-team1.jpg
 

theevilsharpie

Platinum Member
Nov 2, 2009
2,322
14
81
The cisco guys in the office is arguing that you need special team aware switches in order for the adapter teaming to work properly..

This is partially true. You need to have a switch that supports LACP in order to do bi-directional link aggregation.

If you're using teaming strictly for failover, then you don't need anything special on the switch.
 

Enigma102083

Member
Dec 25, 2009
147
0
0
Cisco guys are right. Your switch needs to be aware of what's going on if you're using it for load balance. Granted it doesn't HAVE to be Cisco gear to do it, but none the less they are right.
 

Lifted

Diamond Member
Nov 30, 2004
5,748
2
0
For strictly failover (active/passive) the switch(es) don't need to be aware.

For load balancing (active/active) on a single switch, the switch must support link aggregation.

For load balancing + failover (each server port connected to a different switch or blade), the switches need to be in a stacked (Cisco) / virtual chassis (Juniper) configuration.
 

Emulex

Diamond Member
Jan 28, 2001
9,759
1
71
Very similar to how vsphere does active/active - but there is no true load balancing, you need something smart like openflow to handle vpn port routing on a per connection basis.

Easier to just move up to faster networking.
 

drebo

Diamond Member
Feb 24, 2006
7,034
1
81
In this case, you don't need switches as the NICs aren't doing LACP. In this case, the HP network drivers use round-robin MAC addresses when responding to ARP requests for that IP address. In that manner, they can aggregate without causing MAC flapping on the switch.

It's not as good a solution as LACP, but it is functional.

I typically avoid using it.