Hi, first let me thank you for helping me with my adduser hat on last week by giving helpful and quick answers to my question regarding debconf. I really appreciate that.
Now I am coming to you as a User of Debian-Installer and am fully prepared to be sent to some other place with my questions. In a small side project I am currently helping a company to design and implement a deployment process for embedded Debian to a headless multimedia controller based on amd64 (sic!). Those machines do have HDMI and USB, but hidden away inside the box so you need to open the case to access those ports. I would like to avoid that at least for the bulk of installations. Unfortunately, I find the Installation Guide a bit terse on the topic of preseeded, headless installation, but am prepared to help improving the document. I would probably need some guidance in finding the "right" way to do things before writing them down (or, if weirdness turns out to be a bug, filing the appropriate bugs) and am wondering whether this list might be the correct medium to answer installation questions regarding preseeding. I am afraid that technical questions regarding preseeding will get drowned on debian-user, and probably not find in-depth answers on debian-user-german. After a preseeded installation, a common task is to hand over the freshly installed system to some kind of configuration management like puppet. I have seen more or less ugly methods do to that, including a multi-hundred character long shell string as d-i preseed/late_command with advanced multi level quoting hell, having an localinit.deb which is installed along with the base system and which does that initialization and enrollment after first reboot into the fresly installed system, staying around for the entire life of the box, etc. Is there any documentation / opinion about doing this more elegantly? And then: Can I have a single pre-seed file that would allow me to configure the Installer and Apt to choose the first web proxy from a list of proxies defined in some pre-seed data field, choosing the first one that happens to be available and responding, falling back to direct connection if none of the proxies is there? That would allow me to use the same Installer image for installations in a variety of places with various differnt proxy setups, avoiding the work to have an Installer image per site (giving the users the possibility to fail by just choosing the wrong USB stick). Imagine that I have a remote system in a network that I don't manage, for example a hosting network. I have a box running some kind of unnamed linux, a locally provided rescue system, or grml. I am currently installing in such environments by debootstrapping manually, chrooting into that system and using my existing configuration management to "personalize" the installation. Is there a possibility to start d-i proper in such a system to get all of d-i's magic, probably skipping partitioning and filesystem creation (that might already been done)? And last question: How seriously pluggable is the Installer? Can I, for example, build my own udeb for apt configuration or my own, much less flexible partitioner and just throw those udebs into an installer tree and directly use it? Or do I need to delve in-depth into the build process of the entire Installer to do that? Thanks for helping. No need to Cc me any more, I am subscribed now. Greetings Marc -- ----------------------------------------------------------------------------- Marc Haber | "I don't trust Computers. They | Mailadresse im Header Leimen, Germany | lose things." Winona Ryder | Fon: *49 6224 1600402 Nordisch by Nature | How to make an American Quilt | Fax: *49 6224 1600421
