OpenBSD Journal

OpenBSD Kickstart

Contributed by jose on from the rapid-rollouts dept.

Chris writes: " I work in an ISP/Web Hosting company and we have a ton of servers to install. Having a RedHat kickstart like thing would be something nice to have as fast installs push servers out faster.

I would like to know if people out there in the OpenBSD community have built kickstart-like install methods for OpenBSD, and would like to hear their experiences with them. At initial glance, it doesnt seem like something that would be that difficult, yet things always look that way at first and turn out to be more complicated.

Other's experiences would be very valuable."

I know that there are some techniques out there, but does anyone have a working recipe?

(Comments are closed)

Comments

  1. By tom hensel () tom@replic8.net on mailto:tom@replic8.net

    maybe this might be of interest for you:
    http://www.feyrer.de/g4u/

    Comments

    1. By m0nknutz () on

      You run into problems with partition sizes etc when using ghost 4 unix. I would recommend using dd, netcat and nfs. Then you could copy your master image bit for bit to other machines.

      Comments

      1. By Anonymous Coward () on

        If you dd it, wouldn't that mean you'd have to re-install then manually dd after the reinstall?

        Unless you care to give an example with dd and netcat, that would be greatly appreciated on my part!

        Regards.

        Comments

        1. By m0nknutz () on

          Ramdisk with the appropriate files. Bingo.

          I hit google up for an example and found.

          http://www.rajeevnet.com/hacks_hints/os_clone/os_cloning.html

          Hope this helps.

          Comments

          1. By Anonymous Coward () on

            Thx dude!

      2. By Anonymous Coward () on

        ah, g4u *IS* dd but uses ftp instead of nfs.

    2. By Anonymous Coward () on

      Cool! I didn't know about this. Thanks a lot for the URL! This is something I was looking for as well.

  2. By Matt () on

    Check out this port on misc:
    http://marc.theaimsgroup.com/?l=openbsd-misc&m=104313392300688&w=2

    A guy made a jumpstart diskette that can boot via PXE, partition your drives using percentages (elimiating most problems of differing disk sizes) and script all steps in the install process. I've never tried it myself, but it looks very cool.

    It shouldn't be too hard to write a script that can dynamically update his configs to your specifications (I think).

  3. By Matt () on

    Check out this port on misc:
    http://marc.theaimsgroup.com/?l=openbsd-misc&m=104313392300688&w=2

    A guy made a jumpstart diskette that can boot via PXE, partition your drives using percentages (elimiating most problems of differing disk sizes) and script all steps in the install process. I've never tried it myself, but it looks very cool.

    It shouldn't be too hard to write a script that can dynamically update his configs to your specifications (I think).

    Comments

    1. By Matt () on

      I meant to say "check out this post on misc". :)

      Comments

      1. By Anonymous Coward () on

        You're fired!

  4. By Blake () on 2112.net

    not quite exactly what you're looking for, but I've been using flashdist to keep a bunch of network boxes in sync; just update the master and "burn" a new image onto the target drives. It was designed to work with compactflash cards, but it works great with normal hard disks as well.

  5. By Anonymous Coward () on

    http://www.deadly.org/article.php3?sid=20000621021736

    mainly the last paragraph

  6. By Anonymous Coward () on

    The sitexx.tgz files may be of interest too. See http://www.openbsd.org/faq/faq4.html#site

  7. By djm () on

    The installer script is very, very easy to hack. In a previous job, we customised one to automate the entire install. It used ftp to pull disklabel prototypes off an FTP server, preconfigured networking to use DHCP and did the rest of the customisation using a siteXX.tgz package. Imaging machines was a matter of booting off a floppy.

    This took less than a day of one of our developer's time, though it was made somewhat easier by our use of standardised hardware.

  8. By Mike Sullenszino () nospam at sullenszino com on mailto:nospam at sullenszino com

    All I do is keep a master machine, including a few ports I always install. Then I create base33.tgz and etc33.tgz from the master machine (base3.3 is almost everything except /etc and most of dev). I post those to an internal ftp server. I have a script in /root that configures the basic network/smtp/apache stuff. I then can use a standard boot disk, when it asks where my install sets are, I point it to my ftp server with my custom base33.tgz and etc33.tgz. After first install, run my little script and all is basically configured. Not very elegant, but has been effective for three upgrade cycles (from 2.9-3.3) and several small rollouts of new servers.

  9. By David Norman () on http://openbsddiary.org/

    Something I did when I deployed computer labs worth of computers for my university was use Symantec Ghost. Take a master image and you can deploy it to many machines at a time. A switched gigabit network is ideal for imaging over say 30 machines at a time, but even a 10 megabit hub network will work if you have 4-5 hours to burn.

  10. By Chris () on http://unixfu.net/

    thanks for the feedback, it is much appreciated :)

Latest Articles

Credits

Copyright © - Daniel Hartmeier. All rights reserved. Articles and comments are copyright their respective authors, submission implies license to publish on this web site. Contents of the archive prior to as well as images and HTML templates were copied from the fabulous original deadly.org with Jose's and Jim's kind permission. This journal runs as CGI with httpd(8) on OpenBSD, the source code is BSD licensed. undeadly \Un*dead"ly\, a. Not subject to death; immortal. [Obs.]