PXE Boot: Stuck Contacting Server

[67_dbc]

Hi,

 

I've been experiencing an issue where my PXE Boot client is unable to contact the server. The strange thing is that, the Contacting Server step could be 20 minutes or up to an hour trying to attempt. There has been cases where I will sit there and watch it and it will eventually locate the BootImage and allow me to hit F12 to start the TS. Now, PXE is configured in SCCM the same way on my other 30+ DPs and I have 2 that have this same behavior.

 

Please see attached problem.

 

I've removed the PXE option from the DP, I've removed WDS, rebooted the server, re-set everything back up and have no luck. I've tested this on a physical client and on a VM client, same behavior. My WDS role and SCCM are all on the same SCCM box/subnet, therefore i believe IPHelpers is Not required. My SCCM machine is on a ESX farm which is currently on the same subnet pool as the client through testing.

 

I have several SCCM DPs on an ESX farm on different subnets with SCCM/PXE configured and I don't have this issue.

 

Any thoughts?

 

 

[Jorgen Nilsson]

Hi,

IP helpers are required if the client is not on the same network as the server holding the DP(PXE enabled) role, the same as a normal DHCP server. So you need to make sure that IP Helpers are configured, is it the correct server that answers the request? it looks like a WDS server which have admin approval enabled which is responding.

Regards,
Jörgen

[wilbywilson]

Yeah, I would check for the existence of another (non-SCCM) PXE server on these 2 troublesome subnets.

[67_dbc]

Thanks for the reply. The Client is on the same subnet as the ESX host as the SCCM server. There is only one router between the host and the client. The Cisco router itself is doing the DHCP handoff. PXE/WDS reside on the same box. The WDS role was setup then left alone as is. I've compared those settings to be sure from a working PXE to this troubled one and didn't find any differences. The WDS setting related admin approval is not checked, so I have to rule that out. Any other thoughts?

 

Thanks,

 

Eric

[67_dbc]

Update: We decided to open a case with MS and they can't seem to figure out why either as of yet. They been moving my SCCM environment all over the place trying to figure it out. From the community, nobody else has had this issue before??

 

Anyone?

 

Eric

[anyweb]

The Client is on the same subnet as the ESX host as the SCCM server

what network drivers are you using for the Configuration Manager server NIC

[67_dbc]

Niall,

 

I know this is a year later response, just been really busy. Anyways, I've taken the USB Boot Media approach for my guys to get around this issue. The SCCM DP on the ESX host is using vmxnet3 NIC. We are on VMware ESXi 5.1.0. I've tried E1000 NIC but I did not see anything change. I've given MS 3-4 months to work on this and they keep telling me its the network but with several pcap logs, nowhere shows the break in communication during handshakes between Primary, the DP or the test VM Client. I've been given a diagram how PXE flows including pictures the network/sccm team drawn using mspaint how this works, silly, yes. At the time before they called it quit, we could agree on "somewhat" is that the series of dots appear at stage when the client verification check at the database reports unknown or known client. After that it gets lost in space, dots start to appear and goes in a loop. Since then, MS decided that my case needs to close with no resolution and asked me to open it up if anything new comes up.... At this point I've taken a break from it all, partly why I am responding to this post so late.

 

I'd love to get back on this topic if anyone else can chime in.

 

Thanks,

 

EC