<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
<body text="#000000" bgcolor="#ffffff">
On 11/19/2009 05:50 PM, Gianluca Cecchi wrote:
<pre wrap="">On Thu, 19 Nov 2009 13:56:18 +0800 He Rui wrote:
<pre wrap="">Hi Gianluca, I experienced the same as you when I preupgraded to rawhide
with anaconda 13.8. But I just retested it, the preupgrade upgraded the old
system to f12 with anaconda 12.46. Then I didn't encounter this again.
Can you confirm your anaconda version when reboot after preupgrade?
How to check?
Now I'm on F12, finally upgraded and running....
My steps were, resuming them, starting yesterday:
- yum update on my F11
- yum install preupgrade
- run preupgrade GUI, select F12 and complete this phase and reboot
- upgrade starts but stops because my redhat-release was "funcky". I
change it and reboot again in upgrade mode
- upgrade starts and proceeds until it finds an incomplete rpm and
offers only to abort
---> here I would like pre-verification during the preupgrade phase,
or to be offered a chance to retry, so that I can put
the right package
- In my case the problem was with libgcj-4.4.2-7.fc12.x86_64.rpm.
it was downloaded as 17.5 MB instead of the correct 18.5Mb.....
While in upgrade with the error I changed to console 2, I downloaded
the package from another machine, put it to an usb key and connect it
to my server, then I chroot and manually ran rpm -Uvh of it and
finally reboot again in upgrade mode
- server boots in upgrade again and flawlessly continues from where it
stopped (as I see fewer packages showed to be upgraded). To have
upgrade run I need to change again redhat-release to F11 version text
(it was already updated and contained F12... so that upgrade complains
of this if I don't reverse it...
----> probably better a sort of lock file instead of testing
contents of a file....
- upgrade completes (about 2.5 hours spended in finishing phase... it
seems it scans all the disk files... perhaps verification) and reboots
- my newly upgrade f12 stops with the message
dracut: /var/log/lvm2.log: fopen failed: No such file or directory
and then inability to scan LVM devices and so in particular my root device.....
- boot again F12 with option "rdinitdebug rdshell"
Here I go with fantasy, also based in previous experience .-)
I'm put into a shell where I can manually create the directory
/var/log and at this point from this rescue shell I can scan and
activate my LVM volumes. So now I can chroot in my F12 root LV and
mount /boot under it and
1) gunzip-cpio initramfs into a temporary directory,
2) create /var/log under this initrd tree and repack into a new
initrd filethat I put inside /boot
3) change grub to point to this new initrd
- the system finally correctly reboots in F12 !!!!
If you tell me where to search for anaconda version you referred (in
/root/upgrade.log or other files eventually...)
I will report
It seems that you preupgraded your system to f12 (constantine), whose
12.46. You can see it by: <br>
yum info anaconda <br>
And your issue is different from mine. So if your problem depended on
package, you can try the latest version <a
. Please file a bug <br>
if it is clarified. <br>
<pre class="moz-signature" cols="72">--
FAS name: Rhe
IRC nick: Hurry(rhe)#fedora-qa
Email: <a class="moz-txt-link-abbreviated" href="mailto:email@example.com">firstname.lastname@example.org</a>