Initial DFS replication for a LARGE folder

By p_jones79 ·
Hi all,

I've recently setup a remote office for a client, and thats all working fine.
They use a DB application to wrap jobs into correspondence, drawings, plans, projects etc and it has a data folder over 100 GB.
The users in the remote office need access to this application and the managers want this to function as quick, or as close to the main office.
Now the DB part of this app will be fine over the site 2 site VPN, but the data itself (we're talking 5-10 MB drawings here) needs to be local-enter DFS.

I've been playing with DFS over the last week and I think this will wwork fine, but I've got an issue with initial replication, as it'll be over 100GB over a 2Mbps SDSL IPSec VPN link!
I wanted to move a copy of the data to the remote server and then setup replication, thinking that the server would ignore files which matched, but after testing I see this isnt the case, it created a sub directory called 'NtFrs_PreExisting___See_EventLog' on the remote server and dumps everything that existed before replication started into there, then begins fill replication!.

Is there a way to tell FRS not to replicate items already in place?

This conversation is currently closed to new comments.

Thread display: Collapse - | Expand +

All Answers

Collapse -

Sorry, no way to do this.

by daniel.tisher In reply to Initial DFS replication f ...

To my knowledge, DFSR needs to complete the initial replication itself (indexing for replication / file change counters etc).

One trick I have done on a 600GB+ volume was to set up a PC (with server R2 OS) at the same office location as the authoritative server. Attach an external HD large enough to store the data to this PC. Configure that machine as a server, add it to the DFS replication group and start replicating from the authoritative server.

Once the initial replication has completed, you may take an image of the PC and ship the HD + Image to your branch office. Now have the tech in the branch office stand up the PC image on same hardware and plug the drive in. Now set up replication with the branch office server that you wish to use as your production server and start replication. The replication traffic will go from the local machine that you just put online to the production server. Once that initial replication has completed successfully and your diagnostic reports are all clean you may remove the temporary server from the replication group and dfs namespace.

Sure, it is cludgy but when you have to set up replication of over 600GB between New York and Sydney within 5 days... well, it works. :)

Related Discussions

Related Forums