[Mimedefang] Clamd failing after a reboot - cannot connect to clamd.sock

[Rich Stanton]

Has anyone else seen this?  I'm running clamd 0.65 & mimedefang 2.39.  After
a reboot, no messages get through - the error in my maillogs is 'Could not
connect to clamd daemon at /var/spool/MIMEDefang/clamd.sock'.  clamd.sock
exists in the /var/spool/MIMEDefang directory, and is owned by the defang
user.  If I delete the clamd.sock & clamd.pid files, then start the clamd
service & restart mimedefang, everything is fine (until the next reboot).
Clamd is running as the defang user.  My init scripts are the redhat ones
provided with mimedefang.  Is it enough to simply delete the clamd.sock/.pid
files in the 'stop' section of the init script, or does this indicate I've
done something else wrong?  If so, any ideas what?