VMware Cloud Community
leedudmish
Contributor
Contributor

Auto Assign / CLI DVS port group to LAG

Hi all,

We have a new cluster and 8 hosts which is configured with a dvs switch - each host has had a lag created on the uplink side of the DVS to the physical network infrastructure (4 uplinks teamed). All is working ok however what we have found is that any new Distributed port groups pushed out to the DVS switch is not assigning the port group to the LAG uplinks instead pointing to the original dvs uplinks of the dvs switch(1-8) of which no physical adaptors are bound as they have been moved to the LAG.

Does anyone have a similar set-up and is there a way to ensure port groups assigned to the dvs switch are automatically added to the lag uplinks rather than the standard DVuplinks or a cli command way of driving this change as currently each port group pushed to the host will require manual intervention by the engineer to assign the port groups uplinks from the dv uplinks to the lag under teaming and failover options of the port group moving the normal dvs uplinks 1-8 to unused and then moving the lag from unused to the active uplink.

Because we are looking at an 8 node cluster we would have to visit each port group and assign the 8 different node lags we will have over to the port group each time...its not that we cant do it but just curious if there is a way of setting a policy or cli to automate this some fashion to prevent and possible mistakes by engineers as the no of port groups will expand as applications come along

Cheers all,

Lee

0 Kudos
1 Reply
leedudmish
Contributor
Contributor

Hi All

as an update to this if anyone comes across a similar issue, the dvs switch has 32 uplink ports 8 assigned as standard and a further 24 spare, as we have worked through getting more servers online we have increased the uplink could on the LACP Lag and we have reduced the original 8 ports down.

Once we have all our 8 servers online we will have reduced down the standard dvs uplinks to 0 thus it will leave 32 uplinks on the LACP lags and therefore by default when a port group is pushed it will select that uplink as a default as there are none other available so it will automatically assign.

this of course only applies you have used all the uplinks or you reduce the standard uplink count to 0 thus forcing the port group to use the only available uplink LAG.

cheers

Lee

0 Kudos