Re-replicate to the repository and after it fails grab a copy of the epoapsvr.log and the agent log from the client and post them here. Also post the name of the distributed repository you are failing to replicate to.
The error posted indicates it could be a DNS issue. On the ePO server try running "nslookup <whatever the name is of the machine hosting the repository>" and see if it resolves the correct IP.
Doing an nslookup from the main epo server brings back the correct IP address for the server in question.
Looking in epoapsvr.log it reads -
HTTP Session initialized
20100112171806 I #7176 naInet Connecting to HTTP Server in socket-mode
20100112171806 I #7176 naInet Connecting to Real Server: nav260 on port: 8081
20100112171808 E #7176 naInet failed to find host for HTTP server: nav260:8081, error 11001
20100112171808 I #7176 naInet HTTP Session closed
Where nav260 is for some reason in the place of where the servers actual hostname should be which is what Im finding most confusing!
Is the agent service present and started on the client? If you do a "netstat -anb" on the client is FrameworkService.exe listening on port 8081? From the ePO server can you successfully connect using "telnet <client machine's name> 8081"?
Socket Error 11001 effectively indicates a failure to connect.
Got it! Seems somebody had populated the hosts file with nav260 for an old application and epo agent had somehow picked the name up...
I #'d the entry out and redeployed the agent to the server, after checking the location in the list was correct at servername.domain/Software I replicated successfully finally.
Thanks for your help Jeremy, pointed me in the right direction of name resolution issues after a very long day!!