Colin, I think that you may be right. Now that I think about it, the problem only seems to occur after linux has been running for a long period of time without a logon (i.e. overnight). I think I have the BIOS configured to kick into sleep mode after a certain time of inactivity. This would make sense as when the system slows down the timer interrupt servicing would occur less frequently. I couldn't, however, find a refence to this problem in the Battery Power Mini-HOWTO at www.linux.org. I'll try turning off BIOS power management overnight tonight and let you know what happens. BTW, my kernel is not compiled with power management support, perhaps if I had, this problem wouldn't have occurred? Anyone know for sure?
Scott / Laurent, Thanks for your suggestions. Yesterday I installed the adjtimex package and ran adjtimexconfig. It did it's thing and returned. I then ran linux overnight having set both the CMOS and system clocks to the correct time. Same problem. Checking it at 6 AM today showed a system time of 1:26, but 'hwclock --show' indicated that the CMOS is still correct. If, after tonight, I still have problems I'll post my /etc/adjtime and output from adjtimex -p for scrutiny. Thanks for all your help, Jason >* Laurent PICOULEAU <[EMAIL PROTECTED]> writes: > >> On Thu, 23 Sep, 1999 =E0 10:38:18AM +0000, Jason Christensen wrote: >>> Would anyone be able to speculate as to why my system (kernel) clock is = >slow? >>> I know that my hardware (CMOS) clock is maintaining a time that does not= > drift >>> more than a CMOS clock normally does, but my kernel clock will lose appr= >oximately >>> 4 hours in every 9. >>> = > >> Could you post your /etc/adjtime and the result of adjtimex -p ? And veri= >fy >> if GMT=3D"" in /etc/default/rcS. > >Can it be that your BIOS set the system in sleep mode during that 9 >hours? I have a wrong time whenever that happens (and Laurent is >right in a sense that adjtime then thinks your clock is slow and >tries to correct this). > >I heard the solution to this problem is mentioned in >/usr/doc/HOWTO/mini/Battery-Powered.gz, but I haven=B4t read up on >that topic yet. > >HTH, > Colin > >-- = > >Colin Marquardt <[EMAIL PROTECTED]> > > >-- >Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] < /dev/null > >
