mailscanner 4.2x and postfix 2.0.14 do play well together

[James Davis]

To make a long story short, only 1.1.11+ and less than 2, not 3 nor
3 - 1 and greater than zero... (sorry.;-) and mailscanner can work
in the dual postfix config.

Every combo and fix I've tried of postfix 2+ leaves me with corrupted
queue files on outgoing mail (in the outgoing postfix).
I've not dived into the queue file problem beyond trying to trace a
test message. It looks ok in the inbound postfix, mailscanner gets
it in its queue OK. When it is put into the outgoing postfix I get this:

Aug 27 17:14:52 sunblade1 postfix/postfix-script: [ID 197553 mail.info]
starting the Postfix mail system
Aug 27 17:14:53 sunblade1 postfix/master[27933]: [ID 197553 mail.info]
daemon started -- version 2.0.14-20030812
Aug 27 17:14:55 sunblade1 MailScanner[27946]: MailScanner E-Mail Virus
Scanner version 4.22-5 starting...
Aug 27 17:14:55 sunblade1 MailScanner[27946]: Using locktype = flock
Aug 27 17:15:05 sunblade1 MailScanner[27947]: MailScanner E-Mail Virus
Scanner version 4.22-5 starting...
Aug 27 17:15:06 sunblade1 MailScanner[27947]: Using locktype = flock
Aug 27 17:15:15 sunblade1 MailScanner[27948]: MailScanner E-Mail Virus
Scanner version 4.22-5 starting...
Aug 27 17:15:16 sunblade1 MailScanner[27948]: Using locktype = flock
Aug 27 17:15:25 sunblade1 MailScanner[27949]: MailScanner E-Mail Virus
Scanner version 4.22-5 starting...
Aug 27 17:15:26 sunblade1 MailScanner[27949]: Using locktype = flock
Aug 27 17:15:35 sunblade1 MailScanner[27951]: MailScanner E-Mail Virus
Scanner version 4.22-5 starting...
Aug 27 17:15:36 sunblade1 MailScanner[27951]: Using locktype = flock
Aug 27 17:16:39 sunblade1 postfix/smtpd[27960]: [ID 197553 mail.info]
connect from nermal.douglas.co.us[172.31.125.101]
Aug 27 17:16:39 sunblade1 postfix/smtpd[27960]: [ID 197553 mail.info]
0CCF5F63D: client=nermal.douglas.co.us[172.31.125.101]
Aug 27 17:16:39 sunblade1 postfix/cleanup[27961]: [ID 197553 mail.info]
0CCF5F63D: message-id=<200308272316.h7RNGph2007748 at nermal.douglas.co.us>
Aug 27 17:16:39 sunblade1 postfix/qmgr[27897]: [ID 197553 mail.info]
0CCF5F63D: from=<root at nermal.douglas.co.us>, size=3334, nrcpt=1 (queue active)
Aug 27 17:16:39 sunblade1 postfix/smtpd[27960]: [ID 197553 mail.info]
disconnect from nermal.douglas.co.us[172.31.125.101]
Aug 27 17:16:39 sunblade1 postfix/qmgr[27897]: [ID 197553 mail.info]
0CCF5F63D: to=<jdavis at douglas.co.us>, relay=none, delay=0, status=deferred
(deferred transport)
Aug 27 17:16:40 sunblade1 MailScanner[27946]: Postfix queue structure is depth 1
Aug 27 17:16:40 sunblade1 MailScanner[27946]: New Batch: Scanning 1
messages, 3659 bytes
Aug 27 17:16:41 sunblade1 MailScanner[27947]: Postfix queue structure is depth 1
Aug 27 17:16:41 sunblade1 MailScanner[27948]: Postfix queue structure is depth 1
Aug 27 17:16:41 sunblade1 MailScanner[27951]: Postfix queue structure is depth 1
Aug 27 17:16:41 sunblade1 MailScanner[27949]: Postfix queue structure is depth 1
Aug 27 17:16:41 sunblade1 MailScanner[27946]: Virus and Content Scanning:
Starting
Aug 27 17:16:41 sunblade1 MailScanner[27946]: Uninfected: Delivered 1 messages
Aug 27 17:16:41 sunblade1 postfix/qmgr[27942]: [ID 197553 mail.info]
82C28F667: from=<root at nermal.douglas.co.us>, size=7095, nrcpt=2 (queue active)
Aug 27 17:16:41 sunblade1 postfix/smtp[27969]: [ID 947731 mail.warning]
warning: corrupted queue file: active/8/82C28F667

Without the Mailscanner and second postfix, the inbound postfix can route
mail as a gateway just fine (sans the defer_transports line).

lindsay suggested I try the "hold" queue approach with one postfix. It may
work just fine (haven't tested it myself yet) with postfix 1.1.13 (current
v1 patch release). I however have noted that there is code in the postfix.pm
module that seems to "only" like the deferred and defer directory pathnames.
Still looking and yes my first tests were with v4.23-7.

Has anyone else tried postfix 2.0.14 with the latest mailscanner?

thx

james (i now have mush for brains) davis