NIC Teaming to seperate blade switches (no IRF or stacking)

[lcaccamo]

We have a host on a blade (esxi 5.5 with Ent + licensing). The chassis has two interconnect switches in it. The host has one standard vSwitch with two pNICs uplinks used for management and vm networks. These are both 1Gbps. The two interconnect switches are not stacked or configured with IRF as they don't support such functionality. Since the two pNIC uplinks are on separate switches I cannot create an LACP on them. Is it safe to say that there is now way, in this configuration, to utilize more than 1Gbps throughput? I believe I am forced to go with an active/standby teaming configure since the uplinks cannot be aggregated.

Am I right?

[brunofernandez1]

yes you are right. You can use a max. of 1GBit troughput for each vmnic. BUT you can use both vmnics in active mode.

For example:

if vm1 is connected to vmnic1, you can use 1gbit/s

and if vm2 is connected to vmnic2, you can use also 1gbit/s

so each VM can have only a max of 1gbit/s but in Total 2gibit/s

I ope that you understand what I wan't to say :smileygrin:

[lcaccamo]

When I set VM network uplinks both to active active, then I have intermittent packet loss to downstream VMs. As soon as I change it to Active / Standby, the issues stop.

[brunofernandez1]

what network load balancing mode to you have active?

you should have this one selected without LACP: Route based on originating Virtual Port

please have o look on this blog: http://www.vmwarearena.com/2014/02/vsphere-distributed-switch-part-19.html

[adamwiso]

What kind of switches(?) are they not stacked?