serverboot

[Roland McGrath]

> I have been actively recommending serverboot over the bootscript
> method.  In almost all cases where people show up on irc with boot
> problems, I tell them to switch to serverboot and it solves the
> problems.

I wish I'd known this.  We've had positive reports about kernel boot
scripts here for the most part.  Are there problems other than poor error
handling?  Someone needs to reproduce and debug the error handling behavior
(when you have bad function names and so forth in your boot script).  But
if you have a proper multiboot setup (i.e. a correct boot script translated
into grub "module" commands), what goes wrong?  I don't think I've seen any
reports about such problems.

> It might be worth noting that I have never gotten the funky bootscript
> method to work.  I also haven't bothered memorizing the commands,
> since serverboot is quite easy and works in every case.

You shouldn't need to memorize anything.  Just write it in a grub boot
menu.  serverboot will stop working one day when the filesystem formats
change and we don't want to bother updating its ancient code.  But grub
will have been updated already, since it is now taking hold among users for
booting Linux too.