DHCP - ip configuration could not be reserved (no available address timeout etc.)

[hubbardmeister]

Using ESXi6.7 -

I have two guests    A and B  -  Guest A is a transparent bridge, it passes traffic, it also has a management port that i can assign using DHCP (or not)

1/ If i assign a port group to each and use DHCP to an external  upstream DHCP server. All works great for both guests.

2/ Assign static IP's , all day long - not a problem.

3/ If i try and have guest  B obtain a DHCP address via Guest A  with the internalonly port group it will;

   a/ Work 100% of time as long as Guest A  does not have an associated DHCP address it's obtained.

   b/ Will not work and get timeout when Guest A has it's ip address via DHCP

My question is:

Is there a restriction on the number of DHCP IP's a Guest can obtain and where is that  limiter?

[sk591]

Is it possible for you to share an image of the network diagram so that we can better understand how things are configured here? I think the right approach here is to have individual guests configured with either a static IP or have all the guests upstream to an external DHCP server.

[hubbardmeister]

Hi,

Logical attachment along with screenshots. The discovered subnets seems to be significant, initially it shows all nets, then when this is going to fail ESxi has discovered the subnet, but not all of it (it's 70.50 to .100, however it's processing what it sees.)

Additional information -  When  i completely re-cycle/restart  ESxi , all of the initial DHCP request will go through for the VM's. A subsequent down/up on a guest results in a timeout. This seems

like an Esxi problem.

 

[hubbardmeister]

Hi,

Yes, i'd want to use all of them for the DHCP upstream server(Juniper), that is what fails and should not.

Thanks.

[Scott Vessey]

Moderator: Moved to vSphere vNetwork