DFS-R takes hours to replicate large file changes

i have been evaluating dfs-r in both server 2008 rc2 , select code use replacement our current robocopy process.  move number of virtual host disk files large 10-20gb database files. 

 

i have set dfs-r replication group between 2 2008 servers using full-mesh topology.  initial replication seems work fine, , can make small changes file renames without issue.  file deletions appear replicate in timely fashion.

 

the problem i'm having if change large file - either copying out of replication source group , in, or new big file (big defined 400mb - have tried larger - same problem), takes hours replication happen.  there no network or disk activity on either server.  not bandwidth issue - these servers local gigabit connections. i trying use remote differential compression final deployment on slower 100mb link.

 

i have run reporting tool, , shows fine.  error in dfs event log event id 5002 says "the dfs replication service encountered error communicating partner <partner server> replication group reptest.

 

additional information:  error 1753 there no more endpoints available endpoint mapper.

 

i have plenty of disk space, , have increased staging quota 24gb, larger entire test folder.

 

 

any ideas?  have high hopes in our environment, i'm frustrated lack of real-time management tool of kind can show replication queue length, etc.

 

thanks in advance,

tj cornish

 

thanks response.  ended opening case microsoft support.  have found situation happening because of microsoft forefront antivirus.  unclear exact cause is, have found following observations:

 

• when replication not occur right away, because source server doesn't see there new file replicate.  usn journal not updated correctly.
• renames , deletions work fine.
• we thought related size of file.  have determined not size related, rather file type related.  had been testing ghost image files of various sizes.  expanding size range , types of files working with, found 12mb ghost image fragment not replicate correctly, 350mb executable archive works fine.
• running fsutil usn command on file not replicating causes usn updated, in turn causes replication of file resume.
• rebooting source server causes stuck file replicate.
• the nightly forefront av scan causes file replicate.

my case in hands of forefront team.  have number of debugs , trying replicate problem.  dfs group spoke believes not dfs issue, rather forefront , usn journal issue.

 

i post final outcome thread when hear forefront group.

 

thanks again,

tj

 

Windows Server  >  File Services and Storage