Re: `system clock stops operating' problem from 4-months old system.

[YONETANI Tomokazu <qhwt+dfly@xxxxxxxxxx>]

On Sat, Jul 02, 2005 at 10:31:30AM -0700, Matthew Dillon wrote:
> :Hello.
> :I have three machines which have been running DragonFly-Stable
> :(i.e. pre-RELEASE systems) for about 124 days or so until it stopped
> :updating the system clock.  Kernels are built from sources as of
> :15th of February this year.  Unfortunately, I had to reboot them and
> :update to the latest RELEASE as they are used as production servers,
> :so I can't give much data for investigating possibly remaining bugs.
> :
> :When I logged into these machines, date(1) shows the same time for a long
> :time(or probably not updating at all), and the number from
> :
> :  vmstat -i |grep ^clk
> :
> :didn't change.  I believe that the commit at 2005/03/27 11:25:10 PST
> :should have fixed the bug(which means RELEASE have that fix too), right?
> :
> :Thanks.
> 
>     It should have.  What we need, next time it happens, is a kernel
>     core dump.  To do it get on the system console, kill 1 to drop into
>     single-user mode, sync a couple of times, and then ctl-alt-esc into
>     DDB and panic the box.
> 
>     The release occured after that fix so this sounds like a new issue
>     that needs to be found and addressed.

Maybe my message was very confusing, but I was talking about the system
clock problem with pre-RELEASE kernel, not the RELEASE kernel.  I'm now
crossing my fingers that the bug I hit today has already been fixed :)

>     Did all the machine's system clocks stop after the same amount of
>     uptime ?

Yes, they were rebooted on the same day and had been up for 4 months.