TCP port 993 hijacked by mountd, rpc.statd

[Jason M. Leonard]

Greetings,

I have a production mail server that runs imaps.  Sometimes when I reboot,
tcp port 993 (imaps according to /etc/services) is taken by either
rpc.statd or (currently) mountd before inetd starts, which causes imaps to
fail.  I tried rearranging things in rc.d, but the RPC bits seem pretty
determined to start before the inetd bits.  I can manually stop rpc.statd
or mountd, restart inetd, then restart rpc.statd or mountd, but that seems
to screw up RPC for my Sun workstation NFS clients.  In fact they won't
even boot because they can't mount /var/mail.

So right now I have the option of providing email service to my IMAP
clients or providing them with functioning workstations, but not both.

Any suggestions?  Any idea wow can I reserve TCP port 993 so that RPC will
leave it alone?

This is 5.2.1-RELEASEp5 running uw-imap; I can provide more details if
necessary.


:Fuzz