Deployment Solution

I am using 6.8 as my server. I try to push out a image and all 15 PCs come up as masters. I have multicasting enabled. I have created a NEW PXE file to test on.

for some reason they all show up as masters.

any ideas?

I do know the switches are cisco. And they are on a VLAN.

Are you saying they are all on the same VLAN?
It could be a timing issue. The master will only wait 5 minutes as a default and then it will start imaging by itself which would cause all the clients to image by themselves also.

all of the clients I have tested on came up in under 5 minutes. I also know that all multicasting check boxes are checked so it can use it.... but isnt

Sounds like multicasting might be blocked at the switch level. You probably want to take a look at that.

my network people tell me nothings blocked. Anything I may be able to tell them to look at? im not a network admin

Check your: Control Panel; Altiris Deployment Server Options; Disk Imaging tab. If this is set to "0", then each machine should get its own session, which is not what you want in this case.

OK..I am having the exact same issue. Sent off a job to image 15 stations. I watched them all come up and the each started there own session (each is a master). Looked on the image share and can see each machine pulling a unicast from the share. I have checked and the Disk Imaging Tab is set to 4.

One thing I did to was go into the WinPE config and add the line to make sure the WinPE firewall was off. Before that the jobs would just fail out.

Are you sure you're not using the intel undi driver?

I hadn't even thought of that possibility. Thanks robinw

I am not sure if it is the same thing or not, but I can't multicast either.

* Server setup for multicasting (needing four clients to start)
* Using WinPE 2.1 with all drivers selected (no UNDI issue)
* I kicked off 48 images at once, and my script has a "Run Script" pause in it so that I can be assured all computers are ready before I kick the process off.

The computers all start displaying notices like;

* "Becoming counting master"
* "Becoming counting client"
* Becoming negoiating master"
* etc

After a few minutes (probably the default five minutes actually), all of them start to unicast. Not cool.

I had the problem up till I downgraded back to 6.8 (luckly our DS server is virtual and I had a database backup before the upgrade). Even a clean install gave me grief. I would recommend to anyone, if possible, to bring up another DS server and slowly migrate clients as the problems are resolved. If anyone has any furhter insight, please feel free to continue this thread.

Good luck

I, unfortunately, don't have access to a previous version environment, and cannot even consider a build due to have pressing deadlines. What I did have time for is to build a model-specific DOS boot disk with the required driver loaded. This has allowed me to, albeit extremely slowly (because I didn't have time to optimise the DOS boot disk), deploy the images required.

I have no idea why this problem exists. Does compressing the WIM image (or some other WinPE 2.1) build option cause this? Our network has not changed since our previous deployments. All we have done is change from our WinPE 1.6/Aclient boot disk to WinPE2.1/Dagent disk - which was done as part of a DS 6.1 SP1 instalaltion.