BTW, there is a load on this machine, its just not constant.
Every year, I see something like the above.
I've even noticed a lack of latency (knock on wood) in the past year.
That, along with the linode upgrades (more drive space and memory without increase in billing) keeps me a firm advocate for Linode.com.
I've actually indirectly convinced two IRC colleages to join the fold.
21:54:43 up 336 days, 9:06, 3 users, load average: 0.00, 0.00, 0.00
Uptimes like this keep me from wanting to perform upgrades!
I think I've been with Linode at least 5 years now. I'm not going anywhere else!
08:39:00 up 279 days, 12:39, 3 users, load average: 0.00, 0.00, 0.00
I can't remember why I rebooted, it was too long ago…
IMO, uptime == code stability, which is just as important as availability, especially when factoring in virtual instances of Linux via UML.
Reboot once in awhile. Make sure your startup scripts work.
There is some real truth to this. At work, went over to DR for a prod outage to add memory. Machine didn't come back due to some bad params. Took the admin a few hours to fix. Wasn't a big deal as a whole because we have real DR, but was a fun time.
Old and busted: Uptime
New hotness: Availability
I always do the driving. I remember that.