- Integrated Systems
- About Us
- Integrated Systems
- About Us
Dec 15, 4:00 am to 10:00 am UTC
Dec 14, 10:00 pm CST to Dec 15, 4:00 am CST
Dec 14, 8:00 pm PST to Dec 15, 2:00 am PST
12-13-2012 12:36 PM
Virtual Connect FlexFabric and PXE configuration
Mahesh had a PXE boot question:
THe customer i am working with has a Virtual Connect Flex fabric and wants to configure the first flex nic 1a for multiple networks. He wants both PXE network (VLAN 194) and vmotion network(VLAN 172) to be available on nic 1a. Is this a possible configuration?
What we have tried already is to untag the PXE network to find the RDP server to complete the deployment and tag it back after the ESX is deployed for the service console to work, but for every ESX server deployment we need to go to the server profile and do the untagging and tagging manually.
So what you trying to do perfectly fine and will absolutely work.
Input from Steve:
However, you don’t need to re-tag the NIC after the deployment.
First, I am assuming that the ESXi management is on VLAN 194, along with the PXE service? Yes
What you need to do is configure NIC 1 (1a) with Multiple Networks 194 and 172, and UNTAG 194 for PXE as the NIC/PXE has no concept of a VLAN tag. The server will PXE boot and RDP can deploy VMware ESXi. Now, once ESX has been deployed, add the host to vCenter and DO NOT set ESXi management for VLAN 194, leave it as untagged on both the NIC within the VC Profile and leave ESX as it is. But the ESX cannot be added to the vCenter because the 194 VLAN IP that it picks up from the RDP does not ping and the host cannot be connected.
As you found, if you change the NIC to tag 194, then you also have to change the ESX management to 194 and vice versa. Simply create a port group on that vSwitch for VMotion and set it to VLAN 194.
So, the NIC will receive untagged frames for VLAN 194, therefore the host should not be configured for VLAN tag 194. However, 172 will be tagged, as this is the VMotion VLAN, the VMotion port group will need be configured for VLAN 194.
Also, I assume that for redundancy, you will do the SAME on NIC 2a and connect it to the same vSwitch?
One other thing, if you go into the NICs on the VMotion function in the vSwitch, you can set one of the NICs to standby, this will ensure that all VMotion traffic will stay on the same NIC, if you do this for all servers in the enclosure, so all servers have VMotion on NIC 2a (for example) and NIC 1a is standby, all VMotion traffic will remain on the same VC module and will only use NIC 1a if there is a failure on NIC 2a.
Once the OS deployment has completed, do not re-tag VLAN 194 in VC. Also, do not configure tagging in the ESXi interface.
I think what is happening is that they are likely using iLO to connect to the blade once the OS is installed, when they do that ESXi will have a DHCP address on VLAN 194 using eth0. They should test the network connect at that time. They may be then setting a fixed address AND configuring VLAN ID 194. If they do this…. They will not be able to connect out, as the server would be tagging on 194 and VC is not expecting tags, , until they re-enable tagging on VLAN 194 in the profile for that NIC.
Make sure VLAN (Optional) is NOT SET in the OS.
Other comments or questions?