Was wondering if anyone could help, i'm trying to setup a Host Profile from a reference host and apply that same profile to 7 other hosts.
I created a vanilla host never used but fully configured, I applied the host profile to another vanilla host that only was added to vcenter and accessible by vcenter.
When I apply the host profile the host stops communicating to vCenter, and all PING packets start getting dropped.
When I connect via console to the host, i've noticed the Management Services show a 0.0.0.0 I am no longer able to configure Management Networks and am forced to reset the management network.
When I finally get the host back accessible in vcenter, I look at the configuration set so far, and all vDS Networks are configured, but the Management VMKernel is still on vSwitch0 and no physical NICs are applied to any of the vDS Uplinks.
Does anyone have any logs I can look at, when I look at hostd.log I'm not seeing anything worthwhile, but I could be missing something, any help would be great, thanks!
You need to have the exact same hardware and a unified networking architecture to get it to work. One issue I have why we can't use them is the port groups our management interfaces are on use different vlans. If you know what the differences that are expected are you can remove them from the host profile, but for me it gets to a point, is it even worth it.
I think I have an idea what might be happening... Most people using host profiles are also using MAC dhcp reservations with the "deny unknown clients" option When you first boot ESXi the vmk that gets created has the same MAC as the physical nic. When host profile "migrates" that vmk0 to the vDS it does so by deleting and recreating the vmk0. This gives the vmk0 a new mac based on combination of the 00:50:56:5 and the last bits are from the physical nic. The information I have on how the virtual MAC is generated are "fuzzy". The key here is that during the host profile remediation vmk0 is deleted and recreated and a virtual mac is generated. In my case the mac is not known in the dhcp scope and then vmk0 doesnt get a dhcp ip address and the end result is host profile fails to apply and the host is left disconnected from the network. The solution that I have not verified yet is to run esxcfg-advcfg -s 1 /Net/FollowHardwareMac. This should make vmk0 retain the MAC address of the phyiscal NIC and DHCP MAC reservations should work.... I have a lab environment where where Im going to test if this actually is the solution. Caution - This also may lead to a duplicate MAC alarm in your environment because vmnic0 and vmk0 will have the same MAC address.