Greetings Fedora Core mavens,

Both Fedora Core 2 and 3 bail out during installation from cd after package selection with errors to the effect that they can't read my freebsd partition(s) so they have to stop and then tell me that "it is safe to reboot".

I have instructed both installers to use/format an existing ext3 partition (eg: no fdisk), to use/format an existing (shared) linux swap partition, and to not install a boot loader, so I'm thinking the bail out is when it goes to create an fstab (?).

I have successfully installed Knoppix, Mepis, Debian, Mandrake, Slackware and SuSE from cd on this same drive with this same freebsd installation existing using their normal installers, but for some reason fedora core 2 and 3 trip over the existing freebsd partition(s).

Anyone have any ideas or suggestions? Have any of you other FreeBSD users tried this?

thanks for listening

"I have instructed both installers to use/format an existing ext3 partition (eg: no fdisk)"

I read this to mean that you tell Fedora to format the partition before installing the packages to the partition. You might try using Knoppix to format the partition and then tell the Fedora installer to use the partition but do not format it.

----------------------------
Steve Stites

Hi, and thanks for your very quick response. I will give it another try with a pre-formatted ext3 partition and will tell the installer to not bother with a format.

If I may ask: what has got you thinking that the formatting of the target partition would be an issue with fedora's installers such that they would have to bail? (when they bail, the target partition is left untouched, eg: it still boots the "old" distro [slack9.0] fine)

To verify things I just booted the freebsd part and I'm posting this from freebsd 5.2 (booted from my Slack9.1 lilo) right now, so obviously the freebsd partition and it's "slices" are not corrupted, this is weird ay?

Hello again.

I'm back from the abyss that is the fedora core installer on a drive that has an existing freebsd partition...and...I'm now short one perfectly good installation of slack9.0, but it's ok, I've got 9.1 and 10 to comfort me ;^) and folks, please join us in praying for Patrick's full and speedy recovery.

anyway:

I pre-formatted /dev/hda1 from slack9.1 with a mkfs.ext3 /dev/hda1 and told the fedora installer to not format it or even the swap partition either, anaconda happily complied.

Results: (/dev/hda3 is my freebsd partition)

"Error informing the kernel about modifications to partition /dev/hda3 - Device or resource busy . This means linux won't know about any changes you made to /dev/hda3 until you reboot so you shouldn't mount it or use it in any way before rebooting".

This had ignore/cancel as options, so I chose "ignore" since my freebsd part was obviously not mounted and thus changing it somehow via the fedcore 3 installer running off cd would have been quite a feat, ay?

Next I got this nifty dialog box, entitled:
"Exception occured" (the wrong spelling is fedora 3's not mine, they get it right in the message below)

"An unhandled exception occurred. This is most likely a bug. Please copy the full text of this exception and file a detailed bug report against anaconda at http://someurl.com. Error: error informing the kernel about modifications to partition /dev/hda3"

I still end up feeling like fedora 3 is trying to add my freebsd partition(slices) to it's fstab and finding it's freebsd partition reading/file reading capabilities to be lacking (though I'm not sure how/why this would be), but yet refusing to simply drop that part from fstab when it encounters an error.

any ideas?


thanks for listening