Use of LACP for Nutanix Network Configuration | Nutanix Community
Skip to main content
Solved

Use of LACP for Nutanix Network Configuration

  • November 4, 2014
  • 3 replies
  • 7645 views

Forum|alt.badge.img+7
Hello folks,

What are the advantages and disadvantages for use of LACP to improve east/west traffic and prevent unnecessary north/south traffic? Are there any documents on it's use for a Nutanix installation? Is the use of LACP very common for Nutanix installations? Are there folks kicking themselves for not implementing LACP.

Thank you for your opinions.

Best answer by DonnieBrasco

The use of link aggregation such as LAG, LACP and potentially other link aggregation technologies is a hypervisor and network switch configuration consideration. A Nutanix cluster can work with and benefit from the configuration of link aggregation on the hypervisor and physical switch.As a best practice, link aggregation on the physical interfaces of a Nutanix node should be limited to same-sized interfaces. That is, do not include a 1 GbE interface with a 10 GbE interface. If you must, configure the 1 GbE interface as a standby interface. Most Nutanix nodes include two 10 GbE interfaces in the standard configuration and as such you should not need to use the 1GB interfaces.Note that the use of link aggregation must not be extended to vSwitchNutanix,

For more details about LACP support for Nutanix clusters, you can refer to below mentioned KB article.

https://portal.nutanix.com/#/page/kbs/details?targetId=kA0600000008f8TCAQ
View original
Did this topic help you find an answer to your question?
This topic has been closed for comments

3 replies

DonnieBrasco
Nutanix Employee
Forum|alt.badge.img+18
  • Nutanix Employee
  • 82 replies
  • Answer
  • November 4, 2014
The use of link aggregation such as LAG, LACP and potentially other link aggregation technologies is a hypervisor and network switch configuration consideration. A Nutanix cluster can work with and benefit from the configuration of link aggregation on the hypervisor and physical switch.As a best practice, link aggregation on the physical interfaces of a Nutanix node should be limited to same-sized interfaces. That is, do not include a 1 GbE interface with a 10 GbE interface. If you must, configure the 1 GbE interface as a standby interface. Most Nutanix nodes include two 10 GbE interfaces in the standard configuration and as such you should not need to use the 1GB interfaces.Note that the use of link aggregation must not be extended to vSwitchNutanix,

For more details about LACP support for Nutanix clusters, you can refer to below mentioned KB article.

https://portal.nutanix.com/#/page/kbs/details?targetId=kA0600000008f8TCAQ

Forum|alt.badge.img+7
  • Author
  • Adventurer
  • 3 replies
  • November 4, 2014
Thank you for your response. This will help us greatly in our decision on the use of LACP/LAG. I appreciate your time.

Forum|alt.badge.img+14
  • Trailblazer
  • 30 replies
  • November 4, 2014
Are you using LACP /w vSphere VDS? If so, the Nutanix best practice is actually to use LBT (load-based teaming) of the physical adapters. Josh Odgers has a good write up here, I also asked the LACP question previously. The major benefit of not using LACP is that it simplifies your datacenter switching config to use trunk ports and LBT.

http://next.nutanix.com/t5/Hypervisor/VMware-vNetworking-NIC-Teaming-amp-Failover-Recommendation/m-p/1296#U1296