Hi,
I've installed zimbra mta on an old server with just a P3 1.00 Ghz Processor.
Every 60 seconds the cpu load rise to about 80% due to a process called zmmtaconfig which fork a cpu intesive java process:

zimbra 20116 0.0 0.2 7936 2440 ? S 19:54 0:01
/usr/bin/perl /opt/zimbra/libexec/zmmtaconfig
zimbra 2360 0.0 0.1 5376 1060 ? S 21:51 0:00 \_ sh -c
/opt/zimbra/bin/zmprov gs mailhost.cc.univaq.it | grep zimbraServiceEnabled
zimbra 2361 0.0 1.4 470772 15344 ? Sl 21:51 0:00 \_
/opt/zimbra/java/bin/java -client -Xmx256m -Dzimbra.home=/opt/zimbra
-Djava.library.pa
zimbra 2362 0.0 0.0 5000 560 ? S 21:51 0:00 \_
grep zimbraServiceEnabled

Looking into the perl code of zmmtaconfig I've found this main cicle:

readConfig();
compareKeys();
stopStartServers();
print STDERR "Sleeping\n";
sleep 60;
%prevknownkeys=%knownkeys;
%knownkeys = ();
$VARS_SET=0;
$SERVICE_SET=0;
$LOCALS_SET=0;
$firstrun = 0;

and changed as above a "sleep 60" to read "sleep 180". The overall load of the machine decreased of a 10%:

11:30:01 all 6.98 0.00 0.60 0.09 92.33

I'm not in production now, but which consequences could I have with this timing? What does zmmtaconfig provides and why it runs on the mailbox machine too?

Ciao,
Claudio