I have about hundreds of server(all has its own working copy), fabric(python ssh tool for administration) will issue multiple svn update to those machines, many commands (10% machine) return Connection reset by peer. The svnserve's log show not enough updates(for example, svnserve's log show only 4 update for 11 svn update, see docs here).
I want to know how to set backlog for svnserve, and how to set time-out for svnserve(svn protocol), and how many connection can be served concurrently by svnserve command?
Is this http-timeout for http or svn protocol?
