Poll: When is the last time your production server rebooted?

4 points by ca98am79 ↗ HN
Just curious what the average uptime is for any production servers that the hacker news community works with.

8 comments

[ 2.7 ms ] story [ 32.4 ms ] thread
I can only vote once, I've checked all our servers, they're all (but one) in the 6 months to 1 year range.

Most of the machines were all rebooted about 7 months ago because of a provider power issue, the only outlier was 334 days.

The shortest was 18 days (memory upgrade).

I was involved with an techno-art outfit some years ago. We needed a file and webserver, but didn't have any money (artists never do...) We went to the scrapyard, found a motherboard, CPU, Powerunit and RAM and bought a cheap harddisk. The whole thing was screwed onto a piece of plywood, installed with freebsd and thrown into a corner. It went for around two years without maintenance or rebooting.
[nic@columbia-23 nic]$ uptime 10:00:33 up 1609 days, 21:24, 2 users, load average: 0.08, 0.02, 0.01
> 0.08, 0.02, 0.01

What kind of 'production' is that ?

It's a machine collecting billing records off a telephone exchange. Hasn't stopped in all that time, and definitely mission critical
I actually had to check.. I haven't rebooted my server on Slicehost since I got the account, so it's been up 357 days.
Production servers should never reach 1-2 years without reboot, not if you care about security. Basically every time a vulnerability is found in the kernel you should update => reboot...
If the uptime of any of your Linux servers is more than a month or so, you're playing with fire.