Messages in this thread | | | From | Bodo Eggert <> | Subject | Re: Linux mdadm superblock question. | Date | Tue, 16 Feb 2010 22:10:07 +0100 |
| |
Bill Davidsen <davidsen@tmr.com> wrote: > Volker Armin Hemmann wrote: >> On Sonntag 14 Februar 2010, you wrote:
>>> In other words, 'auto-detection' for 1.x format devices is using an >>> initrd/initramfs. >>> >> >> which makes 1.x format useless for everybody who does not want to deal with >> initrd/initramfs. >> > > You make this sound like some major big deal. are you running your own > distribution? In most cases mkinitrd does the right thing when you "make > install" the kernel,
I don't know what "make install" will do, so I'll have to expect random results.
I don't expect it to copy bzimage to /boot/linux-version-commentfrommymind, point the "ln" or "lt" entry (depending on if I want to upgrade or to test a new kernel) in lilo.conf to the new kernel and to run lilo -R ln.
I don't expect it to sftp the kernel from my build machine to my server, either.
I expect it to move ~/bin/umount (a wrapper around /bin/umount fusermount -u and smbumount) to initrd:/bin/umount. It might also create an initrd with a passwordless rescue mode. Or it will use a minimal shell, and in case of trouble, I have to fight the shell, too. In short: I expect it to backstab me. (As a bonus, you can't read about rdinit= if you encounter your first initrd-ed system and init=/bin/sh does not work.)
> and if you are doing something in the build so > complex that it needs options, you really should understand the options > and be sure you're doing what you want.
What I do is the most simple thing you can do. No initrd no cry. That's why I have to use the 0.9 format. This - and the fact that my distribution defaults to using the 1.0 format - is what I discovered after upgrading my system last time.
| |