Subject: | |
From: | |
Reply To: | |
Date: | Wed, 20 Jun 2007 08:36:29 +0800 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
John Logsdon wrote:
> I have a problem.
>
> Scenario:
>
> I have a CentOS4.4 box that I want to upgrade to SL5.0. So I tried some
> procedures on another box.
I have a box that was originally Fedora Core 3. Over time, I built
almost all of RHEL4 source making only those changes necessary - note,
since this is for private purposes I don't have to bother with artwork,
trademark symbols etc.
My ~ on that system has an ancient, troubled history and dates back to
RHL 7.x (at least), and includes a spell on Debian Woody. Yes, there are
still vestiges of that.
I applied all the relevant results and basically got to an unofficial
RHEL. Approximately equivalent to CentOS, SL and WBEL.
Last week I acquired a box to replace this. Rather than fiddle with the
upgrade path you're trying, I chose to install afresh.
On nahant-clone I ran "rpm -qa --qf '%{name}\n'" so as to get a list of
packages (I don't want names polluted with version info).
I then used this as input on SL5 to run "yum install" against every
package not already installed.
It took a while, but the results seem pretty good. It even picked up pine.
Then, I created my user account and copied ~ for a few users from the
old system. I've not copied my crontab yet. Really must do that RSN.
The most serious problem I had was with seamonkey, and that's documented
in the archives.
There is undoubtedly some undone system configuration remaining, but
possibly your approach would have similar concerns.
--
Cheers
John
-- spambait
[log in to unmask] [log in to unmask]
Please do not reply off-list
|
|
|