run couchdb as daemon about build-couchdb

[dflourusso]

Run as daemon: couchdb -b
See status: couchdb -s
Kill bg process: couchdb -d

[jhs]

Yes, build-couchdb is just some shortcuts around "make install." After that it is just like normal CouchDB. I believe the wiki has documentation. For most Linux operating systems you simply need to get the init.d script connected. With Ubuntu you can use update-rc.d I believe. Also make sure you have no Ubuntu packages for CouchDB just in case.

[mattezell]

Seems to me that there is more than this and that the documentation is somewhat missing / glossing-over required information to make build-couch work "just like a normal CouchDB". Seems that there are some mixed permissions/ownership issues when using this path that have to be resolved before just getting the "init.d script connected". For example, I have used this project to build - supplying the desired install location ('/opt'). I can kick off /opt/couchdb/bin/couchdb directly without issue (use couch normally). While using the init.d and the update-rc.d does in fact get couch running, it doesn't seem to have correctly binded ip/port and is inaccessible (despite status assuring me to relax)... The inconsistencies in documentation and missed/glossed over steps for setup really do make working with CouchDB somewhat a pain - certainly not relaxing...

Disclaimer: I am not knocking anyone who has worked on any of these project directly - just commenting on the pains associated with setup and resulting from the fragmentation of projects/support/etc. (Additional example: The trunk of CouchDB uses AutoMake - documentation doesn't mention this at all and instead tells you to just enter "./configure" - like that will work on your configure.ac...)

[jhs]

Totally agree. Making CouchDB easier, by various metrics, is my goal. There's still a lot of opportunities for improvement.

[mattezell]

In short, I should do as you are doing and pitch in - that or shut my mouth if I don't have anything constructive to add :)

Great job on your project. Even though there is a bit to do outside of what is explicitly covered in the instructions, this is a great project and you guys are doing a hell of a job.

Keep up the good work!

[jhs]

I definitely admit that I have a perspective problem. I am a sysadmin. To me, if make install finishes with no errors, the hard work is done. Now it's just an edit here and there, and some nice comfortable shell scripts. (The reason this is out of scope so far for Build CouchDB is that every situation is incredibly different, considering disk allocation, and operating system, and requirements the sysadmin has.)

However, I now see your point that most people want CouchDB, not a bunch of files sitting around that in some sense constitute "CouchDB". There are some people doing that. For example, here is a very simple procedure to get Build CouchDB running as a full service on Amazon EC2 Linux: https://gist.github.com/1171217