Community technical support mailing list was retired 2010 and replaced with a professional technical support team. For assistance please contact: Pre-sales Technical support via email to sales@march-hare.com.
Thank you to Peter, Arthur, Tony, Gerard, and Brian for your suggestions and responses! Arthur suggested: > There are so many variables in a CVSNT installation what you > need to do is start isolating things. > > Firstly isolate the network. > > Do the same checkout on the server using :local: protocol then > also on the server using :pserver: protocol. I performed a checkout entirely on the server: set CVSROOT=D:\cvsrepos cvs -Q checkout -p -r BRANCH_NAME module_name > nul ... using both a Windows port of plain CVS 1.11.20 and CVSNT 2.5.03. The results were that plain CVS took 24 seconds, whereas CVSNT took 161 seconds. From this, I conclude that the server's network connection is not my problem, since this all happened locally. Further, the disparity between plain CVS and CVSNT when accessing the same files suggests to me that the disks and RAID drivers aren't my problem either. > Finally if you trace (cvs -ttt co module) then each action > has a timestamp - it may be possible to "see" some specific > points in the trace which are taking too long to execute. I have done so, using a CVSNT client on a different machine. I enabled the registry key to create a server side trace, which seems to prevent the -ttt flag from working. The server side trace can be found at http://external.patientkeeper.com/ServerTraceFile.zip. Brian mused: > Something others haven't mentioned which has come up in > the past. Check your CVSNT Server's network cable. People > have gotten intermittent errors with a dodgy cable. This server has a pair of network interfaces, teamed together via the Broadcom networking drivers. I've checked the statistics on the interfaces, and no errors have been reported since its boot on Monday morning at 3:00. > Are you AD & CVSNT server on the same hub? They're both connected to the same gigabit ethernet switch, yes. Further, pinging the domain controller from the CVSNT server results in 1ms response times, even if the data size is set to 1200 bytes per ping. Does anyone have a good guide on how to read the server trace file? I can look at each line, but I don't have a feel for which actions are taking too long. Again, if there's any information that might shed some light on this situation, please ask. --Mark