F7: unable to upgrade from FC6 (weird hdaX errors)

[André Costa]

Hi Rahul, thks for replying. Comments below.

On 6/3/07, Rahul Sundaram <sundaram at fedoraproject.org> wrote:
> André Costa wrote:
>
> > It didn't go too far: I was immediately presented with another error:
> >
> > "Error enabling swap partition device hda5: no such file or directory.
> > The /etc/fstab on your upgrade partition does not reference a valid
> > swap partition. Press ok to reboot your system".
> >
> > (notice again the 'hda5' instead of 'sda5'... could it be due to the
> > fact that I don't have LABELs for hda5 and hda6?)
>
> Yes. Due the libata changes from /dev/hd* to /dev/sd* everything
> requires labels to upgrade smoothly. Fedora installer provide labels by
> default but if you did any customization you might have to set it
> manually. More information on how to do this is available in the release
> notes at
>
> http://docs.fedoraproject.org/release-notes/f7/en_US/sn-Installer.html#id3152635

I had read about the libata changes, but I thought Anaconda would
handle it somehow.

The link you sent does not explain how to create labels. But, 'blkid'
was a cool tip, it showed me that FAT partition already had a label
that was just not being used by /etc/fstab. As for the swap partition,
I used mkswap -L to set it (after turning swap off just to be on the
safe side).

After this, Anaconda was able to advance a little further.

... until it stopped due to some bug on:

/usr/lib/anaconda/iw/upgrade_bootloader_gui.py.

It appears line 78 caused an array index out of bounds error =/

Analyzing the error message it suggested it could have had something
to do with my /etc/modprobe.conf (it complained about one of the lines
starting with '#').

I booted back again to FC6, edited /etc/modprobe.conf, removed all
comment and blank line and gave F7 another try. This time it went all
the way through.

(I'll file a bug report about this)

Some additional post-installation quirks:

- FC6 kernel has been removed, but remained as default boot option
- 'smart' hasn't been upgraded, and now raises a Python error:

~ sudo smart --gui
Traceback (most recent call last):
  File "/usr/bin/smart", line 27, in <module>
    from smart import init, initDistro, initPlugins, initPsyco
ImportError: No module named smart

- yumex also fails:

~ yumex
RuntimeError: Bad magic number in .pyc file

- first attempt to login as root failed -- desktop appears momentarily
and then aborts, going back to GDM login screen. I was using beryl
with FC6, this might be the reason.

Well, it's running, but there are still some rough edges to fix...

Regards,

Andre