I have a Full Mesh Replication between three File Servers, two are 08R2, the other is 03R2. On one file server in particular I had to delete a large number of folders with files in them, some were difficult due to log file names in the paths. Nonetheless all were removed on a Friday afternoon. We’re now four days later and the backlog report is showing the top 100 files being all those that were deleted and they never seem to clear. I ran a dfsrdiag ReplicationState command and it showed there were no currently open connections. Because the backlog report appears to be quite large, it looks like there are no outgoing changes on this folder, it only accepts incoming. WAN link is a 32 MB DS3…bandwidth is not an issue. Staging folders are currently set at 200GB and the only items filtered are ~*, *.bak, *.tmp. Checking downstream servers, neither have the folders and files removed yet.
I ran a general health report and do not see any issues and I don’t find anything in particular listed in the Event logs. It appears that, for all practical purposes, everything deleted is sitting there waiting on something to happen and is clogging everything up.
I’m stumped. I have a few thousand changes in the backlog that wont clear and no feasible way to address it. This is a LARGE replication group, so removing and recreating does not seem to be a viable option other than a last ditch effort. Is there a way to reset the sending member’s backlog?