[Mimedefang] Solaris 10 errors

[Joseph Brennan]

We put a T2000 into production as an smtp server.  This is the first
Solaris 10 server we have with Mimedefang running.  The max number of
slaves is 50.  This is a high-capacity host.  The load-balancer is
giving it the same weight as some older Solaris 9 hosts and the T2000
should be overkill for that number of messages.

It's getting some sock errors.  It looks like they might be associated
with needing more processes.  Below-- does this look strange?  Look at
all the slaves being started right after the errors.  These are
consecutive lines in syslog.


Nov 13 11:36:53 serviceberry sm-mta[22199]: [ID 801593 mail.error] 
lADGaqD9022199: Milter (mimedefang): error connecting to filter: Connection 
refused by /var/spool/MIMEDefang/mimedefang.sock

Nov 13 11:36:53 serviceberry sm-mta[22199]: [ID 801593 mail.info] 
lADGaqD9022199: Milter (mimedefang): to error state

Nov 13 11:36:55 serviceberry sm-mta[22222]: [ID 801593 mail.error] 
lADGatKp022222: Milter (mimedefang): error connecting to filter: Connection 
refused by /var/spool/MIMEDefang/mimedefang.sock

Nov 13 11:36:55 serviceberry sm-mta[22222]: [ID 801593 mail.info] 
lADGatKp022222: Milter (mimedefang): to error state

Nov 13 11:36:56 serviceberry mimedefang-multiplexor[24689]: [ID 778059 
mail.info] Starting slave 31 (pid 22244) (23 running): About to perform scan

Nov 13 11:36:56 serviceberry mimedefang-multiplexor[24689]: [ID 778059 
mail.info] Starting slave 12 (pid 22245) (24 running): About to perform scan

Nov 13 11:36:57 serviceberry mimedefang[24703]: [ID 516796 mail.debug] 
lADGaK4Z021803: Bouncing because filter instructed us to

Nov 13 11:36:57 serviceberry mimedefang-multiplexor[24689]: [ID 778059 
mail.info] Starting slave 42 (pid 22253) (25 running): About to perform scan

Nov 13 11:36:57 serviceberry mimedefang-multiplexor[24689]: [ID 778059 
mail.info] Starting slave 29 (pid 22254) (26 running): About to perform scan

Nov 13 11:36:57 serviceberry mimedefang-multiplexor[24689]: [ID 778059 
mail.info] Starting slave 19 (pid 22255) (27 running): About to perform scan

Nov 13 11:36:57 serviceberry mimedefang-multiplexor[24689]: [ID 778059 
mail.info] Starting slave 33 (pid 22256) (28 running): About to perform scan



It was running slow, too.  This message took 29 seconds.  Checking RBL
is one of the first things we do, and it's asking local rbldnsd.  Even
once it had the result (by 11:36:51) it took 9 seconds for syslog to
say it's done.  (Recipient changed to xxxx)

Nov 13 11:36:31 serviceberry sm-mta[21803]: [ID 801593 mail.info] 
lADGaK4Z021803: from=<storyorgancoloragony at mail.com>, size=810, class=0, 
nrcpts=1, msgid=<000b01c82613$bae9d5a0$7dda29d3 at whosedrops>, proto=SMTP, 
daemon=MTA-v4, relay=[211.41.218.125]

Nov 13 11:36:51 serviceberry mimedefang.pl[18938]: [ID 702911 mail.info] 
MDLOG,lADGaK4Z021803,bounce,RBL_Spamhaus,211.41.218.125,<storyorgancoloragony at mail.com>,<xxxx at columbia.edu>,Machine 
Shagging   asqwx4aj

Nov 13 11:36:52 serviceberry mimedefang.pl[18938]: [ID 702911 mail.debug] 
filter: lADGaK4Z021803:  bounce=1

Nov 13 11:36:57 serviceberry mimedefang[24703]: [ID 516796 mail.debug] 
lADGaK4Z021803: Bouncing because filter instructed us to

Nov 13 11:37:00 serviceberry sm-mta[21803]: [ID 801593 mail.info] 
lADGaK4Z021803: to=<xxxx at columbia.edu>, delay=00:00:29, pri=30810, 
stat=Your mail server 211.41.218.125 is listed at www.spamhaus.org



Note-- I suspect either a Solaris 10 problem, or maybe something I missed
when configuring the host.  The same Mimedefang setup has been running
fine on Solaris 9.



Joseph Brennan
Columbia University Information Technology