I overheard somewhere (though I can't remember the details) that using the dev install mode makes it easier to upgrade koha. I've never done a koha upgrade; we're not fully migrated yet, but I'd like to facilitate future upgrades as much as possible.

I've had no problem setting up a koha standard install; when I try a dev install, my zebra server won't run. The server reports on boot up (and when I manually start Zebra) that it is starting the Zebra Daemon and Zebra Server, but when I look for the process (top -c), it's not there. No error messages are generated when starting the server.

Is there something special that must be done with Zebra conf on a dev install?

Can someone explain the advantages of a dev install? I understand that it allows you to run the majority of the koha server files directly from a git clone. I see how this could simplify an upgrade, but are there other advantages that will encourage me to get to the bottom of this?

No that's basically it. It keeps everything (configs and code) in one directory, and lets you use git so you have version control over the working branch modifications that you are running on. Then you can pull in new changes from mainline and integrate them as you see fit, switching between branches at will without modifying apache or KOHA_CONF or anything.

For a developer, it is very important to submit patches that are against the current state of the codebase. Therefore we have to keep a reference mainline version current, and generating patches between that and a dev branch is how a submission is made. Git manages all that in a really coherent way, so you don't have to be constantly managing symlinks.

BTW, you might have a problem if the dev install was the second installation on the system to use the same config values (i.e. default port numbers). That would be true for any additional install. Otherwise, check your zebra logs for reasons zebra might be failing.

--joe