[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Jumpstart install
- To: misc_(_at_)_openbsd_(_dot_)_org
- Subject: Re: Jumpstart install
- From: Chuck Yerkes <chuck+obsd_(_at_)_2003_(_dot_)_snew_(_dot_)_com>
- Date: Sun, 22 Dec 2002 23:35:50 -0800
- Mail-followup-to: Chuck Yerkes <chuck+obsd_(_at_)_2003_(_dot_)_snew_(_dot_)_com>, misc_(_at_)_openbsd_(_dot_)_org
- Reply-to: misc_(_at_)_openbsd_(_dot_)_org
Quoting Avijit Pathania (foulplay_(_at_)_foulplay_(_dot_)_org):
> I was wondering if anyone knew of a way to automate the OpenBSD
> Search on Google mentions of using siteXX.tgz and installing/ selecting
> it along with other OpenBSD archives. I was hoping to automate the tasks
> of disk partitioning as well amongst other things. Basically to have as
> little user input as possible, along the lines of Solaris jumpstart.
More acurately, like the Project Athena builds that predate jumpstart
by a good 5-7 years.
Since you don't say, I'll assume you are using a Unix machine.
Say a Sun
When you get it booting disklessly ('boot net'), the trick is to
have the rc scripts for that rebuild image do the work.
1) Label the disks,
2) format them
3) fill them with data.
4) boot blocks added and reboot
5) the installed stuff has a post install script that runs once.
cfengine is handy for that.
#1 easy enough if you have a limited number of disks you are supporting.
We just kept the labels around for Sun 424's and a couple others.
Being "smart" means writing something to interact with disklabel.
On Suns, DECs, SGIs and such, we don't have the nightmare of fdisk.
#2 newfs them all.
Instead of 1-4, you COULD just dd an image onto the disk and reboot.
We used this when a disk died or on test machines when you managed
to destroy libc.so or something. L1-A; "boot net" and go get a coffee.
When you came back, your workstation was rebuilt.
What's the hard part? #1 perhaps.
Visit your host, monkey.org