So i have successfully deploy MDT 2012 , and i tried to deploy to one of my machines, no problem whatsoever.
My concern is that when i tried out multiple deployment concurrently ( 6 concurrent deployment to be specific), the installation gets slowed down horribly. Loading the .wim image during PXE boot requires roughly 5 to 6 minutes, and the installation OS is noticeably longer compared to single deployment (didn't manage to get an exact time reading on this).
Understandably that the deployment will get slower during the concurrent deployment, I'd like to know that if MDT was built this way or is there anyway i could enhance the speed for concurrent deployment? (FYI the MDT server is running on a physical server having all the resources to itself, i.e. 64 GB RAM 24 cores etc.)
- Edited by Jack_Tan Friday, August 16, 2013 7:23 AM
If your using Server 2012, there is actually a known issue for this:
And you can try increasing the TFTP block size for PXE:
Please remember to click “Mark as Answer” on the post that helps you, and to click “Unmark as Answer” if a marked post does not actually answer your question. This can be beneficial to other community members reading the thread. ”
Im running Windows Server 2008 R2 and after reading Villalobos's threads i only knew that MDT 2010 supports multicast.
That's why there is this slow down because I'm deploying multiple machines on unicast Nick. I'm running the MDT Server with WDS role on VMWare, so I'm not really sure what config there is on the switch. Any particular details you need from it?
After enabling multicast from MDT, MDT namespace is showing up on the WDS's Multicast Transmission Folder, and when i tried deploying on 1 machine to see if it works, it got stucked at "Attempting multicast Transfer". Its not copying anything as i see the status of the client in the Multicast Transmission Folder "waiting" only.
Any specific config on the switch is needed to allow this multicast to work?
MDT doesn't itself have any RAM or CPU requirements as it is literally just scripts on a network share. The big limiter will be network and disk
Our deployment solution does about 90-110 concurrent deployments and we notice a 2x slowdown in deployment time. We do not use multicast.
- Proposed as answer by Brian Gonzalez Thursday, September 05, 2013 7:40 PM