Distributed repositories, trying to disable legacy DAT replication

[JoeBidgood]

As we do deploy new machines/rebuild machines at remote offices from time to time would a good practice for us be to
     - Turn off global updating so we dont get hit again if we ever do a manual DAT pull during the day.

Definitely - the whole point of lazy caching is to remove the need for replication, so by leaving global updating enabled we're defeating the purpose of the exercise

What would happen with a new build machine, would it have to get the full DAT package from the remote main epo server which would then be cached on the superagent server?

Correct - on those days when a client machine needs a full dat, it'll be pulled to the local SA cache.

HTH -

Joe

[csteels]

Almost there Joe, just another thing Im curious about....

If I was to do a weekly replication (out of hours on a Saturday night for example) so in effect the full  DAT would be there on distributed repositories but end up being up to a week old at most, if I then rebuilt a machine in an office that needed a full update rather then incremental would it be able to use the few day old DAT and the locally cached incremental files to get itself up to date?  Or would it always go off to the central epo server for the very latest full DAT file?

Thanks again

Chris

[JoeBidgood]

It'll always ask for the full dat - the updater will try the incremental files first, and if it can't get to where it wants to be it'll request the full dat package.

(If we're being 100% accurate the client doesn't ask the ePO server: the client asks the SA, and the SA requests the file from the master repo.)

HTH -

Joe