Having a site configuration with several MP's, a few in untrusted forest can be a challenge. Especially combined with complex and firewall micro segmented network. We had an issues with clients, that couln't find the correct MP during OSD and application delployment. This resultet in time-out errors, causing the installation to fail. The locationservice.log showed that the agent tried the entire MP list. one by one until it finally came to one it was able to communicate with. We talked about setting a static reg key allowedMP like Annop also explains in his blog post below. https://www.anoopcnair.com/sccm-mp-rotation-issue-sup-rotation-fix/ Setting the AllowedMP will enforce an agent always to communicate with the the MP in the registry, but it will at the same time remove the flexibility build in the agent. So we decided a mixed solution. First step was to solve the timeout issue during OSD. Set allowedMP in the begining of the OSD TS Remove allowedMP at the ...
Some network requires option 82 in the pxeboot package to ensure the client is trusted. Option 82 is removed from the package, if you’re using the traditionel pxeboot configuration. To fix this switch to the pxe responder.
Playing around with TFTPWindowsSize can result in faster pxeboot process over a slow line. One of my colleagues has optimized the pxeboot process by adjusting the TFTPWindowsSize. In our case the load of the pxe wim vent down from 20 min to 2 min. The default value i 1, and the test ended with 16. it migth be required with some test in your environment to find the "right" value for you. The change did not impact the pxe process on VMWare. PS/ Do not change the TFTPBlocksize, this will break the pxe process! Next step will bo to implement BranchCache, to get additional performance improvements
Comments
Post a Comment