Date: Sun, 9 Jan 2000 03:06:15 +0100 (CET) From: Oliver Fromme <olli@dorifer.heim3.tu-clausthal.de> To: freebsd-stable@FreeBSD.ORG, freebsd-current@FreeBSD.ORG Subject: Re: load spike strangeness Message-ID: <200001090206.DAA75669@dorifer.heim3.tu-clausthal.de> In-Reply-To: <85820f$qbq$1@atlantis.rz.tu-clausthal.de>
next in thread | previous in thread | raw e-mail | index | archive | help
FreeBSD <freebsd@gtonet.net> wrote in list.freebsd-stable:
> Overclocking is *NEVER* recommended
Neither is posting anonymously (without a realname).
(Sorry -- Back to the topic.)
I have to admit that I've seen the same symptoms, and I have
no idea what's causing it. It happenes very irregularly (and
rarely), but it clearly _does_ happen sometimes. It doesn't
seem to be related to any particular hardware or FreeBSD ver-
sion, I have seen it on both 3.x and 4.0-current boxes. On
some machines it never happened at all (including some busy
servers), at least not while I was logged in and watched it
(it's possible that it happened without me noticing at all).
It _seems_ to happen preferably when a long-time CPU hog has
run (and terminated) recently, such as setiathome. The load
goes up to 1.0 and stays there for some time (could be a few
minutes, or an hour maybe even a few hours), then drops back
to 0.0 for no apparent reason. During that period of load 1.0,
there is no activity. CPU is 100% idle. There is no process
that consumes any significant amounts of CPU time. The box
feels fast and responds quickly to interactive work. vmstat
looks perfectly normal (like an idle machine).
I have come to the conclusion that it must be a subtle bug
somewhere in the kernel's calculation of the load averages.
I tried to track it down in the kernel sources, but without
success. Since it didn't seem to have any ill effects, but
just being a cosmetic problem, I didn't bother to investigate
further. I have to admit that I wasn't even motivated to
submit a PR. Yeah, shame on me.
Oh by the way, I think it happened once even on an OpenBSD/
Alpha box (not sure though, it was a long time ago). Maybe
it's a long-standing BSD bug, or just strange coincidence.
Please excuse me for forwarding this to -current as well, but
I think it's important enough, and -current is affected, too.
Regards
Oliver
--
Oliver Fromme, Leibnizstr. 18/61, 38678 Clausthal, Germany
(Info: finger userinfo:olli@dorifer.heim3.tu-clausthal.de)
"In jedem Stück Kohle wartet ein Diamant auf seine Geburt"
(Terry Pratchett)
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200001090206.DAA75669>