A couple weeks ago I opted to use the MIMEDefang embedded perl interpreter (by turning on MX_EMBED_PERL=yes in the MD init script), however, I have noticed some stability problems when using the embedded interpreter. Every morning at 5am the MD multiplexor essentially craps out, requiring MIMEDefang, clamd, and sendmail to be restarted. I didn't see much in the log files indicating what was causing the multiplexor to die off, but I do know that the daily CRON jobs run at around 4-5am every night.
The embedded perl interpreter definitely minimizes the memory usage by the MD slaves, but I don't really notice an overall improvement in total memory usage when using the embedded perl option. All 1.5GB of RAM is still used up entirely, and I have the same number of min/max slaves that I used when I was not using the embedded perl interpreter. I also use a RAM disk for the /var/spool/MIMEDefang directory, and have that set at 128MB. Maybe I'm doing something wrong here, but I haven't found much documentation on the embedded perl interpreter, other than it improves throughput and reduces memory usage. Any insight would be appreciated. - Chris _______________________________________________ Visit http://www.mimedefang.org and http://www.canit.ca MIMEDefang mailing list [EMAIL PROTECTED] http://lists.roaringpenguin.com/mailman/listinfo/mimedefang
