English Community

Datacenter SystemsBladeCenter / Flex Systems
All Forum Topics
Options

3 Posts

07-04-2020

IN

7 Signins

65 Page Views

  • Posts: 3
  • Registered: ‎07-04-2020
  • Location: IN
  • Views: 65
  • Message 1 of 4

vLAG

2020-07-04, 15:01 PM

LACP (active-active teaming) with ESXi and PureFlex Chassis required vLAG, however do the blade internal ports (INT) from IO Modules also needs LACP?

 

Solved! See the solution
Reply
Options

104 Posts

06-03-2020

US

81 Signins

875 Page Views

  • Posts: 104
  • Registered: ‎06-03-2020
  • Location: US
  • Views: 875
  • Message 2 of 4

Re:vLAG

2020-07-05, 4:39 AM

For LACP to work it must be configured on both sides of the connection. So if you want to run LACP in ESX you must configure both vLAG and LACP on the Flex switches INT ports to that host. Keep in mind that ESX only supports LACP aggregations on the distributed vSwitch, not the standard vSwitch. The ESX standard vSwitch supports static aggregation (using teaming type "route based on ip hash"), and in that case, you would need to configure vLAG and static aggregation ("portchannel") on the Flex switches INT ports to that host.

 

Hope this helps.

 

Thanks, Matt

0 person found this solution to be helpful.

This helped me too

Reply
Options

104 Posts

06-03-2020

US

81 Signins

875 Page Views

  • Posts: 104
  • Registered: ‎06-03-2020
  • Location: US
  • Views: 875
  • Message 3 of 4

Re:vLAG

2020-07-05, 22:54 PM

I will add a few more comments on this, to ensure we are all in sync. 

 

If all you want is to have vLAG/LACP facing the EXT upstream ports, you can do that, and not have any aggregation facing the INT ports. This is strictly decided on what teaming type you have on the ESX hosts in the Flex. If the ESX host vSwitches are configured for the default teaming type ("route based on originating virtual port ID"), than that is what is known as a "Switch Independent Mode" of teaming, and you do not need to have anything special configured on the switches (other than allowing the proper VLANs and such). 

 

In the default mode of ESX vSwitch teaming, ESX will perform a per-VM load balance on the NICs (assuming they are both set to active), where the first VM to come up will use the first NIC, and keep using that NIC for all communications until it fails. And the second VM to come up will use the next available NIC, and so on, and round robin NIC use such that any given VM latches on to a single NIC at a time, and uses that NIC as long as it is good. And then can failover to another NIC in the vSwitch, if the NIC or switch they are using fails.

 

In summary, just because you have vLAG configured on a switch, does not mean all ports have to be part of an aggregation. If they need aggregation will depend on what the other side of each connection is doing. And in the case of ESX, the default teaming type does not use an aggregation. 

 

It is very common in Flex to use vLAG and LACP aggregations for the EXT upstream path, but NOT use any form of aggregation on the INT ports (when the default ESX vSwitch teaming is in use). But if you do choose a teaming type in ESX that is aggregation based, then you must use aggregation on the INT ports facing that server.

 

Hope this helps... But let me know if it just made it muddier. :-)

 

Thanks, Matt

 

 

Reply
Options

3 Posts

07-04-2020

IN

7 Signins

65 Page Views

  • Posts: 3
  • Registered: ‎07-04-2020
  • Location: IN
  • Views: 65
  • Message 4 of 4

Re:vLAG

2020-07-06, 14:53 PM

Hi,

 

Thanking you for your time.

 

 

 

As rightly said it is uncommon to use internal ports as LACP/vLAG. However in my case used internal ports also as LACP/vLAG as such used 'route based ip hash' from VMware.  

 

Reply
Forum Home

Community Guidelines

Please review our Guidelines before posting.

Learn More

Check out current deals!

Go Shop
X

Save

X

Delete