Network latency when DFS replication is active

[rkings]

Hi,

I am running Windows Server 2003 R2 with DFS replication. For over a year,
replication has been working properly and users have not experienced any
latency accessing their files. Recently I've received many complaints about
latency. When I shut down the DFS replication service, everything works
properly. Once I start the service and replication begins, everyone starts
to complain again.

I've cleaned out the staging, conflict and deleted, etc. folders but it
hasn't helped the issue. Has anyone else had this issue? Are there any
other files/folders that I should be cleaning out to bring DFS replication up
to speed again?

Thanks!
Ryan

[HAL07]

Post by rkings
Hi,
I am running Windows Server 2003 R2 with DFS replication. For over a year,
replication has been working properly and users have not experienced any
latency accessing their files. Recently I've received many complaints about
latency. When I shut down the DFS replication service, everything works
properly. Once I start the service and replication begins, everyone starts
to complain again.
I've cleaned out the staging, conflict and deleted, etc. folders but it
hasn't helped the issue. Has anyone else had this issue? Are there any
other files/folders that I should be cleaning out to bring DFS replication up
to speed again?
Thanks!
Ryan

Check event logs regarding DFS and please analyse with reports. You should NOT "clean" (if I read you correctly I believe you
delete?) the staging/conflict/deleted folders!
The staging is holding a buffer/staging that needs to be rebuilt if you delete it. The conflict/deleted contain files that are
removed from the server because they are in conflict. These conflict/deleted folders do not affect performance/latency at all.

My guess is that your staging folders are too small, or you have some other issue in your domain and/or your DFS. If you encounter
DFS errors it's not uncommon to experience latency issues and similar.

If I am incorrect and your DFS is working fine, check out the bandwidth scheduling options in DFS which can be managed from the
DFS snap-in.

[rkings]

Thanks for your reply. You're right that I was deleting the contents of the
staging, etc. folders. I tried it out of desperation...I won't do that again.

I had already tried limiting replication bandwidth to as low a 4Mbps but it
never helped the situation. Microsoft suggested that Symantec real-time
protection was causing the problem. I've turned off real-time protection
(and even disabled the service) and we are still experiencing latency.

I find it hard to believe that my hardware is inadequate. I'm running a
ProLiant DL380 G5 with 4GB of RAM, connected at 1Gpbs. The only services
running are file sharing & DFS replication. HP System Management is telling
me that all hardware is functioning properly and I have no errors on the NIC
card / RAID 5 drives.

The DFS replication Event Viewer log has no errors, and only a few
replication conflict errors.

I'm stumped...especially since RAM utilization averages at 10% and processor
(4 of them) usage is always under 10% as well.