I have an svn server running on a Windows Server 2016 with about 10 main repositories each with a number of projects. I moved from CVS to svn 3 months ago. On CVS I had a simple backup scheme where I used robocopy to nightly copy out all changed files within the last 2 days and zip them up and then send them to an offsite server. This worked OK since CVS uses a 1:1 map of project files against the reposirory repo files. So expanding the zips towards an initial copy of the repo created a mirror.
Now I would like to do something similar for subversion and I have found that there is a built-in svnsync command for this purpose. Is there some current HOWTO I can use to set this up? I mostly find old advice discussing improvements in upcoming versions like 1.5 etc... The plan is to use an Ubuntu server with the same version of svn installed as the sync target. It would be connected via the Internet through an OpenVPN tunnel or an SSL port on the target. Some questions that have popped up: - Should the nightly sync job run on the source or the target? - Can the servers be using different OS (Win/Linux)? - Should I use Apache or plain svn on the destination Linux server? - I can port forward the svn server to the gateway on the target but not on the source, does this affect where the job should run? (If the target server is on a forwarded port then the OpenVPN tunnel need not be set up, but the sync job must in that case run on the Windows server source). Advice welcome! -- Bo Berglund Developer in Sweden
