When running Replication Client, the client will fail to update after about a two minute delay of waiting on the Replication Web service. Examining the Miner event log on the Replication Web host shows errors similar to the following:
Message: LIBRSYNC: Error in xecuting Command <C:\inetpub\wwwroot\ArcFMGDBReplicationService\bin\rdiff.exe> with args < delta "\\CARDINAL\ServerBase\replica\Extract_2014-09-19-07-15-47\Gas.gdb.sig\a0000002d.gdbtable.sig" "\\CARDINAL\ServerBase\replica\Extract_2014-09-22-07-15-47\Gas.gdb\a0000002d.gdbtable" "\\CARDINAL\ServerBase\replica\Extract_2014-09-19-07-15-47\2014-09-22-07-15-47.delta\a0000002d.gdbtable.delta">
The rdiff command is taking longer than the default time allowed for it to perform its calculations that generate the update file for the client. After 110 seconds, Replication Web will kill the process, and the client will not receive an update.
Open the web.config file found in C:\inetpub\wwwroot\ArcFMGDBReplicationService. Add the following text to it to increase the timeout to 10 minutes (600 seconds), rather than the default of 110 seconds:
<httpRuntime executionTimeout="600" />
This text can be added anywhere within the file as long as it lies only within the <configuration> and </configuration> tags. If your file already has a system.web section, then add the executionTimeout line within that existing section. Attached is the sample web.config that ships with Replication Web, which has been modified to include this change (three lines from the bottom).