MailScanner stays at 100% cpu utilization

[Brian Wells]

We are running MailScanner on two mail relay boxes.  They have moderate
traffic, around 48000 messages per day.  These are dual processor
machines.  Quite frequently, we see two MailScanner processes running as if
they were in a loop.  IE:

 12:41pm  up 22:30,  3 users,  load average: 2.67, 2.61, 2.55
85 processes: 82 sleeping, 3 running, 0 zombie, 0 stopped
CPU0 states: 98.1% user,  1.2% system,  0.0% nice,  0.1% idle
CPU1 states: 99.1% user,  0.4% system,  0.0% nice,  0.0% idle
Mem:  1032484K av,  701156K used,  331328K free,       0K shrd,   86704K
buff
Swap: 1020116K av,       0K used, 1020116K free                  322572K
cached

  PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME COMMAND
  550 root       9   0 19960  19M 17148 S    98.6  1.9  1180m MailScanner
  496 root       9   0 20532  20M 17780 S    98.0  1.9  1179m MailScanner

They will usually stay like this until they are manually killed.  A restart
of MailScanner will not always end these processes.  When this happens,
MailScanner usually stops working, but not every time.

We are running MailScanner version 4.23, spamassassin 2.31, F-prot 3.12a,
with Sendmail 8.12.10 on Redhat 8.0 with kernel 2.4.19.

Anyone have any ideas why this happens and how to fix it?

Thanks,

Brian Wells