Rick Varney wrote on Fri, Feb 05, 2016 at 09:31:40 -0800: > Errors I see in my svnserve.log that *may* be related are: > > 28207 2016-02-03T18:31:50.113582Z xxx.xx.xx.xxx - - ERR - 0 210002 Network > connection closed unexpectedly > > 28540 2016-02-03T18:50:55.871361Z xxx.xx.xx.xxx - - ERR - 0 104 Can't read > from connection: Connection reset by peer > > But these only occurred rarely, and not each time I tried to access the > repo and got the above client error. >
You can do 'svnserve --foreground -d --root=/path/to/repos --listen-port=3691 --log-file=/dev/stdout' and test against that svnserve instance to ensure the log messages are related to a particular client action. (You can run multiple instances in parallel.) > At this point I'm stuck. Am I doing anything obviously wrong? > > Any suggestions on how I can debug this? If *both* sides see a "Network connection closed unexpectedly" error at the same time, I would suspect a firewall or proxy or a network misconfiguration. (That said, I don't see how enabling path-based authz would affect the on-the-wire traffic, given that anon-access=none is set.) Try 'svnauthz accessof' and check that your user is reported as "r" or "rw". Try 'svn info svn://localhost' locally on the server: does it fail in the same way as on the client? Cheers, Daniel P.S. If you don't have /dev/stdout you can use --log-file=/dev/fd/1 or --log-file=$(tty).
