In DS 6.9 a lot of new environment variables were introduced for all of the pre-boot environments.

Here's a dump of Altiris variable (with sample values) from my Linux environment (but they are also present in DOS/WinPE):

ALTIRIS_BDC=BDC6.9.9023.318
ALTIRIS_BOOTDEVICE=PXE
ALTIRIS_BOOTDISK=
ALTIRIS_BOOTTYPE=AUTOMATION
ALTIRIS_DS_SERVER=JLT-ATRS1
ALTIRIS_DS_SHARE=eXpress
ALTIRIS_PREBOOT=LINUX
ALTIRIS_PREBOOT_ID=LINUX_X86
ALTIRIS_PREBOOT_NAME=Linux
ALTIRIS_PREBOOT_NAME2=Linux
ALTIRIS_PREBOOT_SUBSUBTYPE=LINUX
ALTIRIS_PREBOOT_SUBTYPE=LINUX
ALTIRIS_PREBOOT_UID=LINUX_X86
ALTIRIS_PROCESSOR=x86
ALTIRIS_PXE_BOOTPATH='MenuOption130\X86PC'
ALTIRIS_PXE_CLIENTMAC=000C29E847FC
ALTIRIS_PXE_ID=130
ALTIRIS_PXE_INTERFACE=eth0
ALTIRIS_PXE_IP=169.254.29.5
ALTIRIS_PXE_OID=130
ALTIRIS_PXE_RID=130
ALTIRIS_PXE_SERVER=JLT-ATRS1
ALTIRIS_SHARE=/mnt/ds
ALTIRIS_UNC=//JLT-ATRS1/eXpress
ALTIRIS_VER=DS6.9

I've highlighted the 2 variables that would allow you to map to your local PXE Server (which in 99% of the time is also your local package server / image store).

 

I’m running a pre script the details of which are shown next

In essence I am copying a modified server.lst file as well as the getsrv.exe file from the default express share into a newly created “tools” directory. The getsrv.exe then determines the subnet, compares this against the server.lst file, extracts the appropriate server name and runs this against the bat file. The last net use cmd make the appropriate local share available.

The details concerning the image source. Note the disable path validation. It will complain when you enter this because in essence the G: drive does not exist, but just override it and it will be fine.

The details of the server.lst file. Note it’s placement and that the deployment server folder is shared to the default express donation.

A sample of a modified server.lst file

At my last job we had a few low bandwidth remote locations and mapping all of the remote image stores with asmbmount wasn't a viable option. I wound up utilizing the token for the agent's IP address to figure out which subnet the client was in, then map the appropriate share based on the results of that. This made future expansion easy because I'd only have to edit the imaging script instead of recompile the PXE image.

TheDude is right, part of creating the PXE image is adding mapped locations. There is no need to modify files after creating the image. For users with multiple sites it is actually nice to make a script to map the drives for you. So if a user is at a different site, they can use the same job to image a computer (pulling the image) because you do not want to image a hard drive over the WAN. I don't know how most users an tell the difference between their sites, but just make a job with conditions that will seperate your sites and run a script per site for imaging. The script should like the following:

mkdir -p "/mnt/images"
asmbmount -b -d DOMAIN -f /Deploypw.pwl //ImageSRV/images /mnt/images

DOMAIN is obviously your domain.
DeployPW.pwl is actually the first 8 charactures of the username you are using when creating the PXE image to map the express share. TheDude is most likely using username SysAltiris (just a guess from his mount.local file) instead of sysaltir.
Last ImageSRV is the server housing your image share.

Why edit the "rc.local" file by hand when the wizard will do it for you?
When you perform the steps in Step 2 of this guide, the wizard will add two new lines to the "mounts.local" file for mounting the new share. It will first make the the directory under "/mnt" and then will run "asmbmount" program. It actually does exactly what you're doing by hand. As an example, this is what my "mounts.local" file looks like

#!/bin/sh
# Mount an SMB (Windows) share
# mkdir /mnt/smb
# asmbmount [-d domain] -f pwlfile //server/share /mnt/smb
# Mount an NFS mount
# mkdir /mnt/nfs
# mount -t nfs server:/directory /mnt/nfs -o nolock
# Mount a ramdisk -- / is on /dev/ram0
# mkdir /mnt/tmp
# mke2fs /dev/ram1
# mount /dev/ram1 /mnt/tmp
# Mount a cdrom -- / is on /dev/cdrom
# mkdir /mnt/cdrom
# mount /dev/cdrom /mnt/cdrom
mkdir -p "/mnt/ds"
asmbmount -b -d ADILSTU -f /sysaltir.pwl //Altiris-DS/eXpress /mnt/ds
mkdir -p "/mnt/nas"
asmbmount -b -d ADILSTU -f /sysaltir.pwl //nas1/public /mnt/nas

Also, I think you meant to say "Alt" instead of "Ctrl" when switching between the two different terminals.

We have several main image repositories. We map these in our Linux PXE boot files, then never worry about them again. In the imaging job, you just disable the path box underneath the path to the image, using a typical Linux path like /mnt/images/imagename.img
To do this, add the directory and path in the rc.local file (under /etc) before the last line and after the mounts.local line. You can see this by choosing to edit your boot image.
Actual additions would look something like these 2 lines:
mkdir /mnt/images
asmbmount -b -d workgroup -f ./Administ.pwl //ipaddress/sharename /mnt/images
This always works in our environment. After modifying the rc.local say yes to save and rebuild the boot image.
We do this for many mount points and that way, if we need to go look for something while it's still booted to Linux PXE, we can use Control F2 to look at another screen while it's imaging, then back to Control F1 to see the imaging job progress. It's great!
This extra screen also allows you to do a showdisk to see if the partitions are like you expect them to be, while imaging. That would be /mnt/ds/techsup/Linux/showdisk or /mnt/ds/techsup/Linux/x64/showdisk for 64 bit boxes.