[nycbug-talk] clock running backwords

[pete wright]

On 9/23/05, George Georgalis <george at galis.org> wrote:
>
> on an event-less standard install of FreeBSD 5.4, from cdrom, I
> get this in my logs for every pid:
>
> Sep 23 15:37:35 tiva kernel: calcru: negative runtime of -1993158 usec for
> pid 457 (sshd)
> Sep 23 15:37:45 tiva kernel: calcru: negative runtime of -1320025 usec for
> pid 462 (sshd)
> Sep 23 15:38:01 tiva kernel: calcru: negative runtime of -665357 usec for
> pid 466 (sshd)
> Sep 23 15:38:01 tiva kernel: calcru: negative runtime of -2684144 usec for
> pid 465 (sshd)
> Sep 23 15:38:09 tiva kernel: calcru: negative runtime of -671008 usec for
> pid 470 (ls)
> > date
> Fri Sep 23 15:38:46 EDT 2005
> >
>
> The box is old and could have a failed cmos battery, but I doubt
> it cause it's been off for a while and while ntp doesn't seem to
> have been run...
>
> > grep ntp /etc/rc.conf
> > grep ntp /var/log/messages
> >
>
> The clock is to the second, per eye compare with box running ntp.
>
> ...more research and I've found that the problem goes away when I
> boot with no APCI. Maybe a build world will fix it, that will take
> a while, but planned anyway. Anybody have comments or experience
> with this?


Ran into this problem on a system, especially under load. Check this URL:

http://www.freebsd.org/doc/en_US.ISO8859-1/books/faq/troubleshoot.html#CALCRU-NEGATIVE


I had to play with various sysctl's but think i have resolved the issue.
Frustrating because problem was not easy to reproduce...there were also some
threads relating to this issue on the questions@ if I remember correctly...

-p



--
~~o0OO0o~~
Pete Wright
www.nycbug.org <http://www.nycbug.org>
NYC's *BSD User Group
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.nycbug.org/pipermail/talk/attachments/20050923/31380572/attachment.html