[Mimedefang] SA upgrade difficulty
NFN Smith
worldoff9908 at sacbeemail.com
Wed Oct 12 14:47:03 EDT 2005
I have a box on which I have RedHat 7.3, sendmail 8.12.8, MIMEDefang
2.25 and SpamAssassin 2.64.
I'm working on upgrading SpamAssassin to 3.0.3, and as a result, I'm
finally getting MIMEDefang updated to 2.53. However, the upgrade
process is going rougher than I expected, and I need a little help on
cleanup.
Things I've done:
1) The upgrade I'm installing is a .RPM from dag.wiers.com,
mimedefang-2.53-1.0.rh7.rf.i386.rpm.
When I installed with rpm, rpm reported conflicts with most of my
existing MIMEDefang files, including mimedefang-filter, the init script
in /etc/rc.d/init.d, /etc/sysconfig, the mimedefang binaries and perl
scripts installed in /user/bin, and man pages. I reran the rpm install
(perhaps unwisely) with the 'force' option. From there I got .rpmnew
files for mimedefang-filter and /etc/sysconfig/mimedefang.
2) Since I didn't see anything obvious for keeping the existing
sysconfig, renamed that one to a backup name, and copied the .rpmnew
file to be the primary file.
On mimedefang-filter, our existing script has some custom coding in it,
so I did the minimum necessary for upgrading the script, renaming the
md_log and md_log_enable settings to md_graphdefang_log and
md_graphdefang_log_enable, respectively.
3) Following starting the mimedefang daemon, I got log errors noting
errors noting that Unix::Syslog was not found, specifically not finding
./Unix/Syslog.pm. I resolved this one by copying
/usr/lib/perl5/5.6.1/i386-linux/Sys/Syslog.pm to
/usr/lib/perl5/5.6.1/i386-linux/Unix/Syslog.pm.
4) From there, following restart of both sendmail and MIMEDefang, things
seem to be working mostly correctly, but I am seeing some log errors on
an inbound message coming from yahoogroups, and slaves dying unexpectedly.
I've quoted relevant log entries below.
I think I'm missing something small here, but I'm not sure of what to do
next to fix this one. For now, the only place I'm seeing this is with a
single yahoogroups message, and I'm not especially concerned if that
particular message gets lost. However, to me, it seems that something
isn't quite right, and I want to get that fixed, before I move on to
upgrading SpamAssassin, and replicating my updates to several parallel
servers.
Relevant log entries:
> Oct 12 11:25:08 alpha sendmail[31466]: j9CIP51w031466: from=<address redacted>, size=4079, class=0, nrcpts=1, msgid=<msgid redacted>, proto=SMTP, daemon=MTA, relay=n3.bulk.dcn.yahoo.com [216.155.201.68]
> Oct 12 11:25:08 alpha sendmail[31466]: j9CIP51w031466: Milter add: header: X-Virus-Scanned: ClamAV 0.83/1130/Wed Oct 12 05:34:00 2005 on aerogram.net
> Oct 12 11:25:08 alpha sendmail[31466]: j9CIP51w031466: Milter add: header: X-Virus-Status: Clean
> Oct 12 11:25:08 alpha mimedefang-multiplexor[30124]: Slave 1 stderr: Argument "LOG_MAIL" isn't numeric in bitwise or (|) at /usr/bin/mimedefang.pl line 522. Undefined subroutine &main::openlog called at /usr/bin/mimedefang.pl line 560.
> Oct 12 11:25:08 alpha mimedefang-multiplexor[30124]: Slave 1 died prematurely -- check your filter rules
> Oct 12 11:25:08 alpha mimedefang-multiplexor[30124]: Reap: Idle slave 1 (pid 31032) exited normally with status 255 (SLAVE DIED UNEXPECTEDLY)
> Oct 12 11:25:08 alpha mimedefang-multiplexor[30124]: Slave 1 resource usage: req=7, scans=7, user=6.110, sys=0.300, nswap=0, majflt=592, minflt=17774, maxrss=0, bi=0, bo=0
> Oct 12 11:25:08 alpha mimedefang[31468]: Error from multiplexor: ERR No response from slave
> Oct 12 11:25:08 alpha sendmail[31466]: j9CIP51w031466: Milter: data, reject=451 4.7.1 Please try again later
> Oct 12 11:25:08 alpha sendmail[31466]: j9CIP51w031466: to=<address redacted>, delay=00:00:01, pri=30961, stat=Please try again later
> Oct 12 11:25:11 alpha mimedefang-multiplexor[30124]: Starting slave 1 (pid 31470) (2 running): Bringing slaves up to minSlaves (2)
Thanks in advance.
Smith
More information about the MIMEDefang
mailing list