gabriel rosenkoetter on Sat, 10 Mar 2001 15:00:14 -0500 |
On Sat, Mar 10, 2001 at 10:51:22AM -0500, Bill Jonas wrote: > The way that load average was explained to me was that it was the number > of processes waiting for time, per second. I could be mistaken about > that, though. Even if it's not exactly what I said, that is what I meant. (And it's also what is implied by saying that the load average is the average number of processes on the run queue for the time frame in question). Oh, I'm not sure if I agree with the "per second" part, but I'm not going to say anything more about this than I've said without going and reading Posix.1{a..g}. > What was the point of the thread again? Oh yeah. I've noticed that a > rogue Netscape thread/process will take my load average to right at 1.00. > (This only occasionally happens, and is usually after Netscape crashes or > becomes lame.) Killing the process solves the problem, of course. I'm > not sure what process might be affecting you. I'd suggest watching top(1) > for a few minutes. Martin's problem seems to be a kernel bug from all I can see. He has two processes in "uniterruptible sleep" state (implying they're doing some kind of I/O, which we're pretty sure they're not), but they're definitely not interested in handling signals (even SIGKILL, which you're the kernel isn't supposed to let you ignore). > [1] # $ while :; do find / -name '*' |xargs md5sum >/dev/null 2>&1 & done > Then, to read the load average when things get *really* sluggish, or you > get a process creation error, use: $ read i </proc/loadavg; echo $i Weird. Anyone know why Linux keeps things (like this) which logically belong in /kern in /proc? As far as reading the loadavg, wouldn't an exec suffice in most shells? ~ g r @ eclipsed.net ______________________________________________________________________ Philadelphia Linux Users Group - http://www.phillylinux.org Announcements-http://lists.phillylinux.org/mail/listinfo/plug-announce General Discussion - http://lists.phillylinux.org/mail/listinfo/plug
|
|