Oh No! Uptime gone!
I logged into the server this morning and hit up the ‘w’ command… I was a little shocked with the results:
[@myserver log]# w
1:10am up 12 days, 3:03, 1 user, load average: 0.10, 0.09, 0.03
USER TTY FROM LOGIN@ IDLE JCPU PCPU WHAT
pts/1 c-68-39-235-113. 12:40am 0.00s 0.38s 0.05s w
I was expecting to see an uptime of 500+ days. As you can see there is an uptime of only 12 days! I started searching through the logs but there was nothing to indicate a system reboot or crash. After looking at some more logs for services and also seeing nothing indicating that the service had restarted, I started googling around.
It has now come to my attention that there is a bug in some older versions of the linux kernel which causes the uptime to reset after 497 days (without system reboot). I had a look at my project uptime logs and this is what I got:
1. 29.03.2005 09:39 GMT 08.08.2006 12:01 GMT 1y 132d 02h 21m 27s
2. 08.08.2006 12:05 GMT active session 12d 02h 55m 23s
Sure enough it had reset after 497 days (365 + 132). There are plenty of pages on the net about it if you fancy some lite reading:
- http://www.jimohalloran.com/2002/12/30/linux-uptime-wraps-around/
- https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=97373
- http://www.adrianbradshaw.co.uk/?p=89
- http://dannyman.toldme.com/2006/08/16/linux-uptime-reset/
- http://www.linux-noob.com/forums/index.php?showtopic=1780
Oh well, since I am now back to zippo uptime, it is probably well and truely time for a system upgrade 
