Bog Roll ::

It's Not Magic, It's Work!

02 Jan 2005

Linux Uptime At Work

On the last day of 2004, I took a production Linux server off-line for 15 minutes while I rebooted after installing a new kernel. Pretty much no body was about, and so nobody noticed. As they say at Red Hat, "rebooting is a sign of weakness", but with a kernel, there isn't much you can do about it, the box had to be rebooted. The box has never had to be rebooted since it went live a year ago, and only went off-line once when the power failed for longer than the UPS could cope.

Once it was done, and all the right things came back during the re-boot I had completed my system upgrade. All the systems are now patched with zero defects, and the HP Insight Manager tool is now working properly on all the machines. During the upgrade cycle I also took the opportunity to strip down a lot of unnecessary services, which should make the boxes safer and faster - less is more.

I still prefer Debian to Red Hat, but at least know I understand Red Hat, it's much easier to work with than before. Either way both are far superior to Microsoft Windows, and thankfully I don't have to work with a Windows server most of the time.


Some friends got us a goat for Christmas. As we don't really need a goat, they asked Oxfam to give it to someone more deserving in the developing world.

I know that charities are notorious for their waste and inefficiency, but I like to think that the money spent on our goat will help someone more than billions spent on more typical Christmas gifts.