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.
On Sun, 26 Dec 2004 16:43:08 -0700, Glen Starrett <grstarrett at cox.net> wrote: >> As for RAM -- an order for a new server is on the way. Dual-Xeon with 2GB >> RAM. Can anyone comment on what their server configuration is and what size >> of projects and groups it serves? > >I have a small "hobby" server, 2GHZ, 1G RAM, running Linux that updates >within seconds, but the timings are greatly dependent on other factors >such as number of files, if you're updating to the HEAD version or a >branch / tag, the size of the files, etc. etc. > Just an additional remark here: As far as I know the way CVS stores the file information is optimized for working on HEAD. It actually stores the *complete* contents of HEAD as is and then all other revisions are chained backwards diffs from HEAD. So if there are 100 revisions and you have say a branch that starts out from revision 1.65 and has 15 revisions in itself then on updates CVS must first grab HEAD, then go backwards 35 steps to revision 1.65 using the stored deltas, then go forward on the branch 15 revisions again using the deltas to finally arrive at the tip of the branch.... Needless to say that this actually is rather processing intensive and in itself would suggest that the best operational way is to work with development on HEAD and branch off releases to the side for later maintenance work. /Bo (Bo Berglund, developer in Sweden)