yeah but there is barely anything there. we don't send out very much
mail at all (the machine is only for internal use). in fact the entire
/var/qmail/queue directory tree is pretty much empty, as is
/var/spool/qmail
this is why it's perplexing... i think it must be a problem with the
install or something; qmail has been acting that way since it was
installed - even when there was barely anyone on the machine.
zugzug# ll
total 16
drwxr-x--- 2 qmailq qmail 4096 Mar 30 2000 ./
drwxr-x--- 11 qmailq qmail 4096 Mar 30 2000 ../
-rw------- 1 qmails qmail 0 Mar 30 2000 sendmutex
-rw-r--r-- 1 qmailr qmail 1024 Jul 2 20:05 tcpto
-rw--w--w- 1 qmails qmail 1 Jul 3 01:20 trigger
zugzug# pwd
/var/qmail/queue/lock
and from roughly the same time...
1:21am up 55 days, 10:22, 2 users, load average: 1.20, 1.24, 1.25
61 processes: 54 sleeping, 2 running, 5 zombie, 0 stopped
CPU states: 0.1% user, 0.6% system, 0.0% nice, 0.5% idle
Mem: 516576K av, 383872K used, 132704K free, 19700K shrd, 257452K buff
Swap: 1020116K av, 3340K used, 1016776K free 93692K
cached
PID USER PRI NI SIZE RSS SHARE STAT LIB %CPU %MEM TIME
COMMAND
32076 qmails 20 0 396 396 308 R 0 90.5 0.0 778:11
qmail-send
19851 root 2 0 1276 1276 680 R 0 2.8 0.2 0:00 top
Jörgen Persson wrote:
>
> On Mon, Jul 02, 2001 at 10:39:43AM -0700, Will Yardley wrote:
> > Jörgen Persson wrote:
> > >
> > > On Mon, Jul 02, 2001 at 01:25:20AM -0700, Will Yardley wrote:
> > > > i've heard from a couple people that the qmail-send process is
> > > > supposed to take up all the avail cpu when nothing else is using it;
> > > > is that right?
> > >
> > > Hardly -- what does the log say?
> > >
> > > Jörgen
> > well the logs aren't really spitting out any errors; they look like this
> > (and a lot of procmail backtraces from various peoples' mail delivery).
> > we do have a lot of people using procmail to filter their mail (probably
> > about 20 or so out of 40 users) and a bunch of people using spambouncer
> > (a bunch of procmail scripts) but the machine itself doesn't have that
> > many users, and the load did the same thing when less people were using
> > procmail and almost noone was using spambouncer.
>
> Procmail may cause it but I doubt it since your problem concerns
> outgoing mail.
>
> Check your queue -- what does ''/var/qmail/queue/lock'' look like?
>
> Jörgen