I inherited 2 DFS-R hosts connected via a WAN. We'll call them Michigan and California. We have several replication groups, one of which is massive, in the realm of 1.4million files and 1.6 Terabytes after deduplication. About 2 weeks ago, all of a sudden we had a 1.4 million file back log. For whatever reason, that backlog jumped back up to 1.4 million at least once since then. Fast forward to earlier this week and the backlog of files being sent from California to Detroit is ZERO. However, the backlog of files being sent from Detroit to California is seemingly stuck at right around 80,000 and growing (as users continue to make changes). For the life of me, I can't figure out what's the hold up. The DFSR logs are pretty much greek to me. I've seen suggestions to disable membership of the backlogged node, wait for the changes to replicate in AD and get picked up by the member servers, then re-enable it, to kick off an initial sync. The problem is, I'm under the impression if I do that I'll lose data that hasn't replicated from Detroit to California yet, since California would become the "master". Sure I can run a preemptive backup, but there's still a chance that someone will change something while that 18 hour backup runs.
Are there any ideas of what I can do. I'd love to narrow this down to figure out what exactly is the hold up. I'm at my wit's end with this thing.
Long term, the plan is to replace this current DFS solution with something else or to at least prune it down and/or break it into smaller parts. However, I'm stuck with what I've got for the moment.
Any help would be GREATLY appreciated.