We figured it out. Changed the batch file, rem-med out the reboot part of it. Reset the hash, tried it again, to help troubleshoot the download. On the client, under the SDMCACHE folder, it has the preferredservers.coreserver.dat file. It listed in there three desktops it was trying to copy from. Looked in the console under Preferred servers at these three desktops. There was no ip range listed. Contacted support, they confirmed, if there is no ip range, it acts like a wild card and any agent will try to download from them. I went into the SQL db for our core and exported out the Preferred servers listing and Preferred servers limit ip, then did a vlookup in excel to cross compare which preferred servers had no ip subnet range. There was a total of about 12, I went in and updated all of these to the ranges needed, and ran the batch file job again. It is working like it should now, downloads and runs in under a few minutes. Going to change the batch file back to add the reboot and rehash it, test it on a test vm.
I think I am also going to go in and set up preferred servers for my server subnets so this does not happen again.
thanks everyone.