Very Computer

I have a PII 450 MHz w/512 MB ram and a clean RedHat 6.0 w/updates. The
/home partition is NFS mounted from and SGI. When I take a look at cpu
usage i kpm/qps most of the cpu is idle (~85%), but load average reports
around 2. Isn't load average a function of cpu usage (only, mostly)? Can
it be that some process is running, but not showing in ps/top/kpm? I
feel that the system response has been reduced quite a bit after
upgrading to RedHat 6.0.

Please email me as well as post to this NG!

Ole Jacob
--
GexCon AS, Bergen, Norway  <http://www.gexcon.com>
Tel : +47 55574334 (office) +47 55558650 (home)
Mob.tel. : +47 95080525  Fax : +47 55574331
PGP key : <http://home.c2i.net/ojt/pgp.txt>

[posted and e-mailed]

The proc(5) man page says

        loadavg

        The load average numbers give the number of jobs in
        the run queue averaged over 1, 5 and 15 minutes.

Processes waiting for (slow NFS) disk operations would be in the run
queue, but not necessarily consuming much CPU.

Have you tried tweaking the NFS mount options to try to get better
performance?

--

<snip>

Really?  I thought processes waiting on IO were not in the run queue;
only processes that were "ready to run".  Am I completely off base?

--
Kelly R. Burkhart

[The litigation] industry was, of course, up and running before the
tobacco litigation, but that taught lawyers just how lucrative it
could be to blame individuals' foolishness on, say, Joe Camel.
   -- George F. Will

The manpage is wrong.

Quote:>Really?  I thought processes waiting on IO were not in the run queue;
>only processes that were "ready to run".

That should be correct.

The loadavg shows the number of "active" tasks. Active does not only
mean "running", but also "doing critical I/O".

kernel/sched.c:

static unsigned long count_active_tasks(void)
{
        struct task_struct *p;
        unsigned long nr = 0;

        read_lock(&tasklist_lock);
        for_each_task(p) {
                if ((p->state == TASK_RUNNING ||
                     p->state == TASK_UNINTERRUPTIBLE ||
                     p->state == TASK_SWAPPING))
                        nr += FIXED_1;
        }
        read_unlock(&tasklist_lock);
        return nr;

Quote:}

All tasks are counted that are either TASK_RUNNING,
TASK_UNINTERRUPTIBLE or TASK_SWAPPING.

Peter
--
   _   x    ___

  \/>'~~~~//
    \_____/   signature V0.2 alpha

[code from kernel/sched.c snipped]

Quote:> All tasks are counted that are either TASK_RUNNING,
> TASK_UNINTERRUPTIBLE or TASK_SWAPPING.

Okay, but try the following experiment on an NFS client:

#!/bin/sh
while /bin/true; do
   cat > /dev/null LIST_OF_LONG_NFS_MOUNTED_FILES
done

In my list of long files I have stuff like the emacs-20.4 source tar
file (i.e., several MB long).  Each "cat" is taking ~1 minute, and
"top" reports its %CPU at ~10%, but my load average is slightly above
1.

--

Your tasks are doing I/O. They are most likely in the state
TASK_UNINTERRUPTIBLE and thus increasing the load. That's expected
behaviour. Load does not mean "CPU load" but more generally "system
load".

Peter
--
   _   x    ___

  \/>'~~~~//
    \_____/   signature V0.2 alpha