On Mon, Oct 25, 2004 at 10:05:48PM +0200, Maurice van der Pot wrote:
> I noticed that for the 6.0.7 release of coda, venus was changed to daemonize
> after startup was complete. This allows scripts to start venus and do 
> something with the coda file system directly afterwards.
> 
> Is it not true that something similar is needed for the coda server?

Yes, I just hadn't gotten to doing that yet. Basically I was always
afraid of forking once LWP got running and as such was envisioning
things like forking at the beginning and then using signals or a pipe to
pass startup status updates back to the parent. But after thinking
things through it really doesn't matter whether we fork before or after
LWP threads have started since they are not kernel threads.

If the server daemonized after it has started, it will be easy to change
the vice-setup scripts to start the server and create the rootvolume.

> Going through the same steps manually works just fine.
> I'd hate to have to script something to monitor the log file.

Right now that is the only option, the server is up and running whenever
it writes "..:..:.. File Server started .*" to the log.

Jan