On 10/24/2011 11:40, Chuck Swiger wrote:
On Oct 24, 2011, at 7:52 AM, A C wrote:
The header from top when things work normally (ntpd no longer running at high
priority in this capture):
load averages: 0.10, 0.09, 0.02; up 14+20:55:09 14:44:43
22 processes: 21 sleeping, 1 on CPU
CPU states: 10.5% user, 0.0% nice, 7.4% system, 6.4% interrupt, 75.8% idle
Memory: 5048K Act, 2432K Inact, 300K Wired, 2316K Exec, 1336K File, 200K Free
Swap: 128M Total, 9272K Used, 119M Free
This box has ~8MB of RAM?
Well, that was a reasonable amount for a ~1990-era machine, and NetBSD will still run on
it today, but running X11 clients is sure to push it into swapping out other processes
entirely. That will cause ntpd (and gpsd, etc) to stop responding entirely for many
seconds while the hard drive chugs-- that's the "uninterruptible sleep" you
initially asked about.
Put more RAM into the box, or replace it with something from this or at least
the previous decade. :-)
Regards,
It has 16 MB of RAM right now (one stick). It had 64 at one point while
I was trying to diagnose this issue. I thought I might have a bad stick
of RAM so I was testing them one by one (one stick in the machine at a
time). I didn't reinsert the other three yet. But it doesn't matter,
even with 64 MB ntpd still crashes which is what started this whole
debugging game in the first place.
Besides, I wouldn't call an Xterm and Xclock anything significant in the
way of X11 clients. I'm certainly not intending to run anything beyond
those. The Xterm is for a console and the Xclock I was using to keep an
eye on the machine so I'd know if the system locked up.
_______________________________________________
questions mailing list
questions@lists.ntp.org
http://lists.ntp.org/listinfo/questions
