On Sat, 5 Oct 2002 13:29:21 -0400 (EDT) Tim Kellers <[EMAIL PROTECTED]> wrote:
> I have several custom kernel source files that I compile on my fastest > machine and "make installkernel" on my slower desktop and server > machines. > > Granted, the kernel files are bigger than normal --I keep many options > that are not specifically required in the sourcfe files that I don't > need; it just keeps me a bit more sane to only change machine specific > options. > I have all of /usr/src exported from the machine I use to compile and > I > have the exported /usr/src mounted on the machine on which I want to > install. > > I use the same method to "make buildworld," too. It's very helpful > when upgrading the 18 odd workstations we have in our computer labs, > too. > > I know there may be pitfalls in this type of mass install but, so far, > I've been bitten by very few bugs. > > > Tim Kellers > CPE/NJIT What sort of pitfalls could be expected? The reason this issue has come up for me is I've been working on getting a diskless workstation going. The diskless workstation will boot off NICs boot-ROM and will load the kernel compiled on my faster machine, but as soon as the kernel finishes loading the machine just reboots. I can't find documentation anywhere that speaks of problems associated with compiling on one machine and then distributing that kernel to other, possibly heterogenous, machines. It appears that the `make' process somehow factors in machine specific information, exclusive of that in the kernel config file. Is this accurate? Nathan To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-questions" in the body of the message
