6 Replies Latest reply on Nov 1, 2010 3:27 PM by metalhead

    REPO architecture question


      I am trying to fix an issue with my repositories, and I am unsure if I can do what I want it to do.


      The issue, stems from when we rolled out VSE 8.7 in January. We have just under 10K machines and 7 FTP repos. The repos were flooded when we rolled out the update, which killed WAN links across our organization.


      We have a DFS in place already, and there is a local DFS replica at 90% of our sites.


      Our DFS only replicates once a day, late in the evening. Due to the late replication, I don't want to use it for DAT files (in case of outbreak)


      What I was hoping to do, was have the FTP repos used for the engine, DAT, BOF DATS and small files, and the the DFS for installation files.


      However, what I found out was, the agent doesn't check ALL the repos...if it fails on the first one, it stops.


      So, for example, if I want to deploy the anti-spyware module to a box, and the install files live on the DFS, the agent will try the FTP server first (probably because that is where is last got a successful update). The files for the anti-spyware aren't there, so the install fails.


      I was hoping the agent would keep trying other repositories, until it found a good one.


      So, my question is, what is best way for me to adjust my repo architecture to both be fast for small updates, yet non-impacting for large deployments?