Red Hat

I going to be updating a few of our rhel5.5 servers this weekend to rhel5.10. I'll be making a checklist for pre & post maintenance comparisions.


so far I'll be checking,
services running
filesystems
netstat
ifconfig
uname
oracleasm
multipath

just basic stuff

if it were you what would you include on the checklist?

thanks in advance

Perhaps take a look at Red Hat's support site and read through their release notes and errata and note the changes between 5.5 and 5.10.

Check version info for any non-RHEL software installed. Newer versions of packages could possibly break third party software.

And if you're a tad paranoid, you could enable the rollback feature in the event something does go wrong. I've only done this with some of our more critical systems, but I haven't made a habit of it as it consumes a lot of space storing the packages.

Overall you should be fine. I do RHEL updates all the time and they usually go off without a hitch.

Good luck.

I have done RHEL updates and encountered problems. This is because we install RedHat on hardware not officially supported by RedHat. For example while RedHat supports some Dell Precision workstations, they do not support any Dell Optiplex workstations and we use them. The upgrades typically break because the new kernels need newer drivers. So I get complete info on my hardware configuration before I do an upgrade. The two big items are Video card and Network interface.

Backups and a tested process to recover assuming you damage your server beyond repair is a good thing to have. It's unlikely, but you would look foolish not to have this sorted before you press the button.

You could always re-install, but then do you know all the packages and their configuration updates? You would have to re-create all the users, ssh keys, all sorts of things to communicate with other servers etc.


Not that I'm paranoid, but I'd prefer not to get the sack if I can take sensible steps first.

Robin

I have daily incrementals and weekly full backs that run on these systems, so all pertinent information is saved.

I only asked the question to get a feel for what everybody else does when performing these types of maintenances, as far as easy to fix stuff that may have been lost during a reboot. like mounted filesystems not in fstab or non-static routes. Others have access to these servers and sometimes they make on the fly changes without realizing the impact of a reboot. Only to come to me the next day asking why this isn't working anymore. Then I have to go track down what they did 6 months ago and fix it. PITA made easier with saved pre/post maintenance config files.

I would recommend that you change the root password and make sure that they don't have sudo rights except for specific actions (make sure they can't escape to a shell with sudo too). Make sure all changes are controlled and then you have more of a chance.

If they can do things, then they can do them incorrectly or incompletely.

Another great one is for mounting a new filesystem, but getting it in the wrong sequence in /etc/fstab You either mount it and then over-mount it, or the mount fails.

I hadn't considered static routes. I try to avoid them if at all possible.



Robin