SCIENTIFIC-LINUX-USERS Archives

June 2006

SCIENTIFIC-LINUX-USERS@LISTSERV.FNAL.GOV

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Jan Iven <[log in to unmask]>
Reply To:
Date:
Thu, 8 Jun 2006 12:36:34 +0200
Content-Type:
text/plain
Parts/Attachments:
text/plain (64 lines)
On Thu, 2006-06-08 at 10:27 +0100, John Logsdon wrote:
> What is the best way to yum upgrade from RH8 to SL4.3?
> 
> I have 2 machines both running RH8 - 2.4.18 and 2.4.20 kernels which I
> want to upgrade to SL 4.3 when it is available.  I have heard of
> difficulties going straight from RH8 to CentOS4.2 and it is better to go
> RH8-FC2-RHEL4.x or some similar route.
> 
> The complication is that on one machine a number of the partitions are
> reiserfs, which is not in FC2/SL4.*.  I don't want to do a reinstall,
> although that is sometimes the best way.  It would mean doing lots of
> reconfiguration.

I'd seriously reconsider this. An "upgrade" in Red Hat terms is not just
updating packages, but also all kinds of wonderful hacks/workarounds
that need to get applied, but are not part of a RPM (e.g. config file
changes = modprobe.conf vs modules.conf). Which is why an upgrade is not
to be done via the RPM management tools, but using the installer's
(anaconda) "upgrade" option. Each "anaconda" usually knows about the
quirks of the previous release only, which is why you got the
recommendation to do the upgrade in small steps. Each of these steps
requires a reboot, and is significantly slower than a fresh install
(since RPMs are deleted+added one-by-one, lot more disk activity).

In addition, anaconda will only handle the issues it knows about. If you
have installed third-party RPMs or just tweaked your system in a way not
foreseen by Red Hat, the upgrade may leave you with a non-working
machine (even if all the RPMs got updated). Sorting out such a system is
rather time-consuming, and you never know whether a particular problem
is still lurking.

The Reiserfs partitions add a nice twist as well, since the stock
kernels will not support them. So you will not be able to access them
until you have rebuild your own kernel.

Wiping and reconfiguring the machine is at least something you can
estimate the effort of. Just keep the old config files around on a spare
(ext) partition.

Lastly, a RH8-era machine will be rather close to the end of its useful
life cycle by now (dried-out condensators in the power supply, worn out
fan/HDD bearings, dust puppies everywhere, ..). Perhaps leisurely
setting up a replacement service on new HW now is more cost effective
than doing a fancy upgrade now, and providing a new machine in 6 months
(and in a hurry).

Hope this helps
Jan

> Presumably there is no need to run in the new kernels until it has been
> compiled to include reiserfs - or is there one already available?  ie all
> the upgrading can be done under 2.4.x kernels.
> 
> I propose trying this route but does anyone have any comments?
> 
> TIA
> 
> John
> 
> John Logsdon                               "Try to make things as simple
> Quantex Research Ltd, Manchester UK         as possible but not simpler"
> [log in to unmask]              [log in to unmask]
> +44(0)161 445 4951/G:+44(0)7717758675       www.quantex-research.com

ATOM RSS1 RSS2