Deployment Solution

This problem has been driving me crazy. We have two sites:

Site A: Our local site where I am and do most of the PC Imaging:

Site B: Our remote site were we have another tech who also does imaging

We have our altiris deployment server install at Site A with me. We have a PXE server here at Site A and another at Site B. My imaging job basically goes like this:

1) Boot to Pxe (WinPE)
2) Push hardware independent image (If at remote site it pushes it from a local server there)
3) Script that copies drivers for individual machine (If at remote site it pushes it from a local server there)
4) Run a script that tokenizes sysprep and copies it to machine.
5) Reboot, let windows install, push assorted other software

Step 4 is were its dying at Site B. Its a very small script:

REM ReplaceTokens .\Temp\%ID%.inf .\Temp\%ID%E.inf
.\rdeploy\windows\firm copy .\Temp\%ID%E.inf prod:\sysprep\sysprep.inf


When it gets to this step the remote site gets the error "The system cannot find the file specified" with a Status code of 2.

The funny thing is the ReplaceTokens line works. The ".\Temp\" file was created fine and the tokens were replaced. And this works fine at Site A but not Site B. The only difference is Site B is using a different PXE Server (but the same WinPE boot image). Any ideas what could be causing this?

Lou

Just though I mention it.
the Image on Site B would be different. does the image on site B have the prod:\sysprep\sysprep.inf
directory?

also putting a pause on the steps might help you out troubleshooting.
i.e.
is f:\ mapped correctly? in your case i think it is as you mention it does create the %ID%E in the \temp dir.

Thanks for the reply! The site B hardware independent image (that gets pushed to the PC) is the same as Site A's. And both have the sysprep directory. I have tried putting pauses in and testing out running the command manually and it works. Very confusing. One thing to note that I just found out. If I take the script that does:

REM ReplaceTokens .\Temp\%ID%.inf .\Temp\%ID%E.inf
.\rdeploy\windows\firm copy .\Temp\%ID%E.inf prod:\sysprep\sysprep.inf

and I split it up into two separate scripts so one does:

REM ReplaceTokens .\Temp\%ID%.inf .\Temp\%ID%E.inf

and the next one does

.\rdeploy\windows\firm copy .\Temp\%ID%E.inf prod:\sysprep\sysprep.inf

The first script creates the token file just like it always has EXCEPT I still get the same error on that script. Kind of funny since the command worked. So I set it to ignore the return status code of 2 and it did then bombed out on the second script (which again works fine on Site A). I'll play with this more and see if I can figure anything else out.


Lou


Edit: And F: maps correctly (as does K: to my remote site and S: to my local site)

one more thing I noticed Im sure you are doing it.
on you steps above u do copy the destination sysprep file to the .temp dir right?
i.e. f:\RDeploy\Windows\firm.exe copy prod:sysprep\sysprep.inf .\temp\%ID%.inf
just before step 4
if not then %ID% will not be problem.

I would also recomment using the full path for Firm
.\rdeploy\windows\firm copy .\Temp\%ID%E.inf prod:\sysprep\sysprep.inf
to
f:\rdeploy\windows\firm.exe



here is what mine looks like hope it helps
1. distribute image
2. copy sysprep from image to temp folder.
------------------f:\RDeploy\Windows\firm.exe copy prod:sysprep\sysprep.inf .\temp\%ID%.inf
3. tokenize sysprep and rename extension to .txt
------------------rem ReplaceTokens .\temp\%ID%.inf .\temp\%ID%.txt
------------------f:\RDeploy\Windows\firm.exe copy .\temp\%ID%.txt z:\sysprep\sysprep.inf
4. update sysprep that is on Z: with correct HAL
5. copy sysprep to final location
-----------------f:\RDeploy\Windows\firm.exe copy z:\sysprep\sysprep.inf prod:sysprep\sysprep.inf

I've done a bunch of googling and forum searching and its funny how everyone handles pushing hardware independent images. Its all basically the same but done slightly different. I am copying the sysprep to the temp directory as %id%.inf at the end of my driver pushing script. Since i determine the model in that script for the correct drivers I figure i'd copy the correct inf there too. So basically what your doing but in a different order.

I may have stumbled across something. The fact that just running the replacetokens command would work but still throw the error had me confused. So i did some searching and I found an Altiris KB (can't recall the number. . .doh) that mentioned Altiris is unhappy when there is more then one share named eXpress. I noticed our remote site's PXE server also had its share named eXpress. So I renamed it, renamed everything in the winpe image that pointed to it, renamed anything in my scripts, etc. I then gave it another try. . . .and our WAN link went down. Well not down but our T1 provider is having power problems and it keeps dropping. So I can't test it right now. Once things are back to working order I'll test out to see if this change helped.

Thanks again for the help!

are you also using site B as the source of your DS express mapped to F: on the client

I beleive F:\ needs to be mapped back to the DS server as it relies on the .temp directory for instructions for each one of the task/steps.

if F:\ is mapped to Site B then it will definitly not work. unless youve change it.

I found that each step in the DS is created in the express temp dir every time a task or step is ran on the client side and internaly the client know where to get those instructions.

No matter which site I pxe boot from (Site A or B) it still maps the F: drive to our deployment server as it should. When I first hit this error I suspected it was mapping the remote pxe server to the F: drive but I confirmed its not. I still haven't had a chance to test this out further as AT&T is taking their sweet time fixing the microwave relay station that went down (and took down some of WANs and internet). Its suppose to be fixed by the end of today so barring any other disasters I can test it out again soon.

WAN is back, renaming the share didn't work. This is definitely an interesting problem.

I gave up and called Altiris support. They are stumped as well. Turns out its not the specific replacetoken's script thats failing but ANY script ran while PXE booted AFTER the first script runs (which runs fine). Only on the remote site. So I'm going to update Altiris to the latest version which has some debugging built-in so they can see if that reveals more.