Some time ago system administrators bragged about how much uptime their systems had. It was seen by many as a reflection of their skills in keeping their systems up and running without needing a reboot.

There even were some web sites dedicated to displaying such information. People do like to brag πŸ™‚

Then one day someone realized that having uptime that covered multiple years was probably not such a good thing after all… People started worrying what would happen if the system actually needed to be shut down. Would it come up again on it’s own? And if it didn’t, where were the people that had installed it and were familiar with it’s particular quirks? Most had probably moved on to other positions or had even left the company.

And what about security? How many kernel exploits have been fixed since that particular version that is 4 or 5 years old? Long uptimes became increasingly unpopular.

In today’s world a long uptime is usually only found in very special installations, and not frequently in systems that are connected to the internet, the potentially lethal realm where hackers (or even script kiddies) can wreak havoc of an unpatched system in a matter of minutes (or is it seconds?).

Yes, maybe there are firewalls and IDS’ between the system and the internet, but hey, if your systems are unpatched, how up to date are your security systems?

I bet there are two kinds of people reading this. The first is the one that smiles and nods agreeingly, and the other has already stopped reading and is busy checking uptimes πŸ™‚

One Response to uptime, the more the better right?

  • Frank says:

    Agree.

    Times change and now … the number of days that a server is already up and running should not be a source of pride .. but of concern, since it must be highly insecure and with lots of vulnerabilities that can be corrected with patches which sometimes involve breaking the famous uptime and reboot