Have you looked at the message at this site?
It says this:
PXE BaseCode and UNDI runtime modules are copied from FLASH or upper memory into the top of free base memory between 480K (78000h) and 640K (A0000h). This memory must be zero filled by the system BIOS. If this memory is not zero filled, the relocation code in the PXE ROMs will assume that this memory is being used by the system BIOS or other boot ROMs.
From the BIOS screen, this machine looks ancient (in computers terms anything over 5 years is ancient to me). You may want to update the BIOS to start out with, if you can find such an old BIOS. A lot of the PXE BIOS changes were made and a lot of machines didn't have them for a few years after 2000. So a new BIOS may help you.
rhyous: Thanks for your information. I upgraded the BIOS of OPTIPLEX 320 from 1.1.5 to 1.1.11 which the one I can get from dell web site. Unfortunately I still got the same error.
I tried on Lenovo T60, still no luck, same error. I am wondering if this is a BIOS version issue. Is there any other possibility?
Do you have any machines that can PXE boot with success? What is the REP installed on? Is their a firewall running on your PXE rep? Make sure that the LANDesk PXE MTFTP Service is running. Is there a firewall between you and your PXE rep?
bnelson, thanks for your reponse. In answer to your questions, even with the same machine if I press F8 during PXE boot it is able to load the WinPE image into memory successfully. But if I schedule a task against it it can not load WinPE.
My PXE rep is on a VWware machine which runs Windows XP, I also tried a physical machine as PXE rep, same result. Firewall is turned off on the PXE rep. LANDesk PXE MTFTP Service is running well on the PXE rep. No other firewalls.
Look in C:\Program Files\LANDesk\PXE\System\images and make sure you have a prov_winpe.img
bnelson: Wow. You saved my life. I should check this before I posted this thread. I guess the PXE rep deployment task did not copy this image file to that location. After I copied peboot.img to prov_winpe.img the Managed Provisioning can boot successfully. Are that two files exactly same?
My previous success is in VMware. Even without Prov_winpe.img on the PXE rep the bare metal machine can still be able to load Winpe image with Managed Provisioning. But with a physical machine (DELL OPTIPLEX320) even HAVING the Prov_winpe.img it still could not load Winpe image. I was confused now.
Does VMware machine do not need the Pro_winpe.img? BTW I'm on LDMS 8.8.
VMWare has nothing to do with it. My guess is you may have multiple PXE representatives. Check in your console (network view, configuration, pxe representatives). When a machine PXE boots, the first PXE representative that responds to it is the one it uses. So the location of machines in your network may affect which PXE rep(s) are used.
In version 8.8, the PXE rep no longer has a prov_winpe.img. It only has the peboot.img, and that is used for both provisioning and OSD. There is only one WinPE image on the PXE rep in version 8.8.
Have you redeployed all your PXE reps since going to 8.8?
IronMike, Thanks for your information. That makes sense. I do have 2 PXE reps, they are both 8.8 new deployed PXE reps.
Now the thing bothered me is that the VMware machine can load WinPE image from the PXE rep, but the physical machine could not do the same thing with the same PXE rep. I tried on another machine (DELL OPTIPLEX GX620), same result. So I do not think this is related to the machine.
But what else can cause this?
1. RAM - I have found that with some machines I would need as much as a gig of ram for some reason.
2. NIC Drivers - Do the PC's actually get an IP address from the DHCP server?
3. Make sure you have only 1 PXE rep per subnet.
Are the errors any different on the PC's vs. the VMWare? Is the PXE rep a vm or a pc? Is the core a VM?
bnelson: thank you for the continuing post.
1. >>> The memory is 2.0 GB, shows 1.96 GB in Windows.
2. >>> The device does get IP address every time when PXE booting, see the screenshot
3. >>> I have two PXE reps in the same subnet, but I tried to disable the PXE services on one PXE rep, only leave anoher running, no difference.
4.>>> for VMware machine, always successful, whatever press F8 or Managed Provisioning.
One PXE rep is on physical machine, another is a VMware machine, both of them did not work for Managed Provisioning if the PXE booting device is a physical machine. Core is a VM.
OK, the coorelation I am seeing is that the Core and 1 PXE rep are VM. When you have another VM, I assume on the same ESX Server, it will provision. When you try to provision an actual PC it fails? There could be an issue with the setup of the the server hosting the VM. Does that server have a firewall? Could be a closed port issue on the VM Host server? It appears as though you have everything setup to work properly, but there is a communication breakdown when you try to TFTP the boot image out of VM to the PC.
bnelson. You caught the root cause. I tried another VM client which is not on the same host which the core runs on. It also could not PXE boot into Managed WinPE. So the problem might be on the VM core or the host which it resides on. I will try more. Thanks.
Further. In the pxemtftp.log on the PXE rep got the following error:
Fri, 07 Mar 2008 17:16:34 PxeMtftp: Unable to get Multicast address for filename X86PC\UNDI\PROVWIN\STARTROM.0