Post by Pavlos Gerardos
I prestaged 2 servers with the data i wanted to copy from a NAS (which
is to be removed from the network).
It's ~750GB of data.
One copy was made to Srv1 with robocopy and the other (to Srv2) was made
one day earlier with simple windows copy...
There is a difference of ~1GB of data (~1000 files).
The primary server should be Srv1, which holds the most up-to-date and
I created a replication group for the two folder targets some 6 hours
ago and then made 4-5 health reports.
All of them show a warning for Srv2 that "This member is waiting for
initial replication for replicated folder data".
In the following lines i can see now ~400000 Backlogged Sending &
Receiving Transactions and ~53000 Files Received...
And these numbers are increasing hour to hour...
It's definitely not working ok....
Any ideas on what should i do?
Since you copied the data using different methods, not all of the
meta-data will be identical, since the meta-data isn't identical, the
file hashes don't line up exactly, so DFS-R goes through conflict
It will take DFS-R some time to resolve the conflicts. DFS-R is smart
enough to notice that the content itself is identical though, so it will
generally be faster then just letting DFS-R copy out all the files
Under 2003 R2 it could take days for the two servers to sync up with
750GB of data, under 2008 I'd expect DFS-R to finish in a day or so
assuming the file content is identical, longer if the file content needs
to be replicated. User activity will slow the process considerably.
Check your staging areas sizes, I'd probably bump them to at least 16GB
during the initial sync (especially on 2008, which handles larger
staging areas more efficiently then 2003 did), then set them to more
sane limits later.
Honestly though, the initial replication just takes time, little can be
done to speed it up, and whatever you do, don't try and help it along
once replication starts, you'll just generate more work for the system
trying to resolve your changes.