Todd Denniston wrote:
John Summerfield wrote, On 09/17/2008 10:31 AM:
>
> I'm digging around in the f9 netinst iso atm, trying to solve my
> kernel problem. The initrd for this iso loads a 100 Mbyte or so
> ramdisk that contains basically the rescue disk+installer. It
> shouldn't be hard to copy this to the installed system and boot it
> from grub.
>
> With the recover option, one would expect this might be regenerated as
> new kernels (and tools) are installed.
>
Careful with that thought, if it is automatically regenerated you might
be stuck in the same situation you are now, i.e., bad kernel|mkinitrd no
booty. :)
Also if you do something like that, I think you would want the image to
have an rpm entry that made for it so rpm/yum would not remove THAT
kernel until you have a new known good kernel and image set for your
machine, even if you have more kernels than your installonly_limit setting.
In theory, yum can already be instructed to keep some number of packages
matching a specified pattern, it's supposed to prevent folk like me
finding themselves with one kernel, and that a dud.
In practice, it seems not work that way more often than it does, and at
present it doesn't preserve old packages.
Presumably though, if something is broken (and some's complaining right
now about a segfault partitioning in Anaconda), then it does need to be
rebuilt.
If that means I have five bootable kernels plus five ramdisks of
recovery stuff, that's fine, disks are cheap even if replacing one's a pain.
--
Cheers
John
-- spambait
1aaaaaaa(a)coco.merseine.nu Z1aaaaaaa(a)coco.merseine.nu
-- Advice
http://webfoot.com/advice/email.top.php
http://www.catb.org/~esr/faqs/smart-questions.html
http://support.microsoft.com/kb/555375
You cannot reply off-list:-)