Thank you (-: I think I can upgrade the kernel without initrd trouble in the futer. And I am trying to compile a kernel without initrd.
On Wednesday 01 August 2001 03:33, der.hans wrote: > Am 29. Jul, 2001 schwäzte Tao Liu so: > > Yes! The problem is resolved! > > I am using 2.4.7 now. > > Thank you! > > > > But what does initrd mean? > > Is it new for 2.4? > > The initial-ramdisk ( initrd ) is a way for Linux to cheat. It needs to > because x86 architecture sucks :). > > There's a very small amount of adressable memory for the kernel to load > from. If the kernel is too big, you can't load all of it. The initrd is a > way of pulling a secondary chunk of kernel off disk after the main chunk of > kernel has been loaded. I don't know the particulars of how it it works. > > > On Sunday 29 July 2001 22:50, you wrote: > > > have you got the line > > > > > > initrd=/boot/initrd-2.4.7-686 > > Instead of this, put initrd=/boot/initrd and if linux.OLD uses and initrd, > then put initrd=/boot/initrd.old in the stanza for it. > > In /boot soft link from the initrd-<version> to initrd for the kernel > you're trying to use and from initrd-<version> to initrd.old for the old > kernel. > > This is better than specifying initrd-<version> in lilo.conf because the > upgrade tools seem to recognize the soft links and update them you you > install new kernel-image packages. In other words, when you install > kernel-image-<newversion> everything will automagically point at the right > thing and continue working. > > It seems that starting with 2.4.x the kernel image packages are depending > on initrds. I've had to fix the soft links and lilo.conf on every box I've > updated to 2.4.x via packages, though. Don't know why that's borken and > haven't researched it enough to file a bug report. Maybe I should anyway as > it keeps coming up on this list... > > ciao, > > der.hans _________________________________________________________ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com
