Not quite sure where best to float this idea so I thought I try here
first to reach a wider audience.
Recently we have seen a number of threads on the fora dealing with
broken 'yum update' issues during updates from 5.1 to 5.2. I guess the
mailing list has probably seen it's share of similar issues.
Here's an example of one such thread:
I'm wondering if it would be useful to put together a *generic*
procedure to recover from such issues, one that should work in most cases.
In addition, maybe some advice/tips to minimise the risk of performing a
large update. I'm thinking of things like performing updates in a screen
session when remotely updating over an ssh login in case the ssh
connection gets lost mid-update. Also, I tend to perform updates in
smaller chunks like:
manually install new kernel packages
yum update yum* rpm*
yum update open*
yum update gnome* kde* xorg*
to minimise the damage if it goes horribly wrong at any point. Not sure
if that's good practice or not but it works for me.
Plus any other advice relevant to the update procedure - how many times
did people ask the difference between 'yum upgrade' and 'yum update',
and which they should use
If it's deemed a good idea, I'm happy to help put together a Wiki page
or something although some help with the content (ie, a generic fix,
tips and advice) would be appreciated.
CentOS-devel mailing list