On Fri, Oct 15, 2004 at 09:41:47AM -0400, A.C. wrote:
We've been seeing an issue with a 9.0 server where from time to time, seemingly randomly, the uptime goes from what it should be to 0 as if the machine rebooted, however, the machine stays up through this and services continue to function...all processes suddenly show that they started at the time the uptime goes to 0, but if you check the PID files for each, they still show the timestamp of the day the machine actually booted up. Active logins show idle time based on when the machine booted as well, tho they show login times of when the uptime reset itself.
The system timer is a 32 bit counter which overflows. With a standard kernel it happens in ~497 days. In your case it happens 10 times more often.
Supposedly this isn't a known issue per SUSE, has anybody else seen this, is it hardware..something else..any ideas?
It *is* a known issue per SUSE: http://portal.suse.com/sdb/en/2003/10/90_scheduling.html If you don't like it you can remove it following instructions in http://portal.suse.com/sdb/en/2003/10/pohletz_desktop_90.html Regards, -Kastus