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.
Chris, > I just ran rlog from my cvs.exe cvsnt client, against a test > server and our > production server - the test server (which is not serving any > other clients > and only has one module out of many tens of modules) took 5 > minutes while > our production server took 45 minutes. There were a couple of threads early this year on performance topics. You need to isolate the client or the server, then the disk then the network etc etc. If this is happening equally on all clients then it's likely the server or network... A common fault is to have anti virus on the server scanning the repository, or enabling network compression... Eg: you could run the rlog command "on the prod server" (ie: not from a client) and see if that is fast (rules out the network), or even use the :local: protocol (on server only). Once you narrow the problem down to the server you can move "tmp" to a RAM disk and retry, then move the repo itself to a RAM disk and retry ( by this you are removing the disk drivers from the equation ). The two performance issues posted by people earlier this year turned out to be a faulty disk driver in one case and a faulty network card in another. Regards, Arthur