Excerpts from Holger Levsen's message of 2012-12-06 07:12:03 -0800: > Hi, > > On Donnerstag, 6. Dezember 2012, Stefano Zacchiroli wrote: > > On Thu, Dec 06, 2012 at 01:11:11PM +0100, Holger Levsen wrote: > > > so to rewrite cloud-init (as simple-cloud-init or whatever) what > > > functionality is needed? > > Doesn't this go in the "wrong" direction? > > well, yes and no. > > this is sloccount for cloud-init 0.6.3: > > Totals grouped by language (dominant language first): > python: 4956 (90.99%) > sh: 470 (8.63%) > xml: 21 (0.39%) > > > The functionality I need can probably be done in ~100 lines of sh, plus I > really couldnt figure out how to use cloud-init at all, so... here i go. >
Perhaps I missed where the functionality you want isn't handled by cloud-init? The way to use it without EC2 or OpenStack is documented in doc/sources/nocloud/README Basically, you put meta-data into /var/lib/cloud/seed/nocloud/meta-data. This is described in doc/examples/seed/. This file is where you set things like hostname and install authorized_keys. Then to do further system configuration a cloud-config stanza goes in /var/lib/cloud/seed/nocloud/user-data The capabilities of cloud-config are best discovered in doc/examples/cloud-config.txt This idea that "all I need is a 100 line shell script" makes sense if your focus is narrow, but cloud-init has grown out of the idea that a single generic interface for doing first-boot initialization of a clean image is more useful than many many 100 line shell scripts. -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected] Archive: http://lists.debian.org/[email protected]
