Note: This is a beta release of Red Hat Bugzilla 5.0. The data contained within is a snapshot of the live data so any changes you make will not be reflected in the production Bugzilla. Also email is disabled so feel free to test any aspect of the site that you want. File any problems you find or give feedback here.
Bug 154802 - sendmail timeouts with msgs larger than 100kb+
Summary: sendmail timeouts with msgs larger than 100kb+
Alias: None
Product: Red Hat Enterprise Linux 3
Classification: Red Hat
Component: sendmail
Version: 3.0
Hardware: i386
OS: Linux
Target Milestone: ---
Assignee: Thomas Woerner
QA Contact: David Lawrence
Depends On:
TreeView+ depends on / blocked
Reported: 2005-04-14 10:28 UTC by murray lotnicz
Modified: 2007-11-30 22:07 UTC (History)
1 user (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Last Closed: 2006-08-21 17:07:04 UTC
Target Upstream Version:

Attachments (Terms of Use)

Description murray lotnicz 2005-04-14 10:28:53 UTC
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.6) Gecko/20050223 Firefox/1.0.1

Description of problem:
mails larger than 100kb or so are getting stuck at DATA and are sitting idly in 
/var/spool/clientmqueue but smaller size emails are being sent without probs. /v
ar/log/maillog shows a bunch of timeouts connecting to localhost but the daemon 
is definitely running on localhost on port 25. iptables is completely down. i'm 
not exactly sure what has changed with the introduction of the smmsp aside from 
sendmail running setgid instead of setuid root but i never had any probs with 8.
11 or earlier with the default config. i've compiled 8.13 on this same box and h
ave run into the same prob... ugh. in desperation i upgraded to the 2.4.21-27.0.
2 kernel and the clientmqueue flushed upon rebooting but now they're starting to
 choke again. any help would be sincerely appreciated.

Version-Release number of selected component (if applicable):

How reproducible:

Steps to Reproduce:
1. mail -s test < vacuum.log


Actual Results:  1. message gets stuck in /var/spool/clientmqueue
2. /var/log/maillog shows timeouts with

Expected Results:  mail should be sent successfully

Additional info:

from /var/log/maillog:

Apr 14 00:33:58 testmachine sm-msp-queue[11610]: j3AGmhm2026980: to=postgres, de
lay=3+14:35:15, xdelay=00:00:00, mailer=relay, pri=8149088, relay=[], d
sn=4.0.0, stat=Deferred: Connection timed out with []
Apr 14 00:33:58 testmachine sm-msp-queue[11610]: j3ACR729002195: to=sysadmin@dom,, ctladdr=postgres (503/503), delay=3+19:06:51, xdelay=
0:00:00, mailer=relay, pri=8627790, relay=[], dsn=4.0.0, stat=Deferred:
 Connection timed out with []
Apr 14 00:33:58 testmachine sm-msp-queue[11610]: j39Cmfm2025894: to=postgres, de
lay=4+18:35:17, xdelay=00:00:00, mailer=relay, pri=10558309, relay=[], 
dsn=4.0.0, stat=Deferred: Connection timed out with []
Apr 14 00:33:58 testmachine sm-msp-queue[11610]: j398iVPv002644: to=sysadmin@dom,, ctladdr=postgres (503/503), delay=4+22:49:27, xdelay=
00:00:00, mailer=relay, pri=10947013, relay=[], dsn=4.0.0, stat=Deferre
d: Connection timed out with []

[root@testmachine root]# sendmail -v -Ac -q

Running /var/spool/clientmqueue/j3DMmnm2027088 (sequence 1 of 10)
j3DMmnm2027088: locked

Running /var/spool/clientmqueue/j3DIYtTH032078 (sequence 2 of 10), Connecting to [] via relay...
220 ESMTP Sendmail 8.12.11/8.12.11; Thu, 14 Apr 2005 01:2
7:18 -0700
>>> EHLO Hello localhost.localdomain [], pleased to m
eet you
250 HELP
>>> MAIL From:<> SIZE=186113 AUTH=postgres@testma
250 2.1.0 <>... Sender ok
>>> RCPT To:<>
>>> RCPT To:<>
>>> DATA
250 2.1.5 <>... Recipient ok
250 2.1.5 <>... Recipient ok
354 Enter mail, end with "." on a line by itself

[root@testmachine root]# ps -ef | grep sendmail | grep -v grep
root     11600     1  0 00:23 ?        00:00:00 sendmail: accepting connections
smmsp    11609     1  0 00:23 ?        00:00:00 sendmail: Queue runner@01:00:00 
for /var/spool/clientmqueue
root     11612 11600  0 00:23 ?        00:00:00 sendmail: j3E7Nw29011612 localho
st.localdomain []: DATA
root     15091 11600  0 01:01 ?        00:00:00 sendmail: j3E81oBR015091 localho
st.localdomain []: DATA
root     15095 11600  0 01:02 ?        00:00:00 sendmail: j3E822si015095 localho
st.localdomain []: DATA
smmsp    17360 11609  0 01:23 ?        00:00:00 sendmail: ./j3DMmnm2027088 [127.
0.0.1]: client DATA 354
root     17361 11600  0 01:23 ?        00:00:00 sendmail: j3E8Nw08017361 localho
st.localdomain []: DATA
root     17515 11600  0 01:27 ?        00:00:00 sendmail: j3E8RInh017515 localho
st.localdomain []: DATA

[root@testmachine root]# netstat -an | grep 25
tcp        0      0  *               LISTEN

[root@testmachine root]# iptables -vnL
Chain INPUT (policy ACCEPT 1791K packets, 395M bytes)
pkts bytes target     prot opt in     out     source               destination
Chain FORWARD (policy ACCEPT 0 packets, 0 bytes)
pkts bytes target     prot opt in     out     source               destination
Chain OUTPUT (policy ACCEPT 1858K packets, 435M bytes)
pkts bytes target     prot opt in     out     source               destination

[root@testmachine mail]# egrep -v "^#" /etc/mail/access
localhost.localdomain           RELAY
localhost                       RELAY                       RELAY

[root@testmachine mail]# grep -i "max.*message.*size" /etc/mail/
# maximum message size
#O MaxMessageSize=1000000

[root@testmachine mail]# grep -i "max.*message.*size" /etc/mail/
# maximum message size
#O MaxMessageSize=1000000

[root@testmachine etc]# egrep "local|testmachine" /etc/hosts       localhost.localdomain localhost         testmachine

[root@testmachine etc]# egrep hosts /etc/nsswitch.conf
hosts:      files dns

Comment 1 murray lotnicz 2005-04-14 10:30:30 UTC
oops, it's reproducible via:

mail -s test < vacuum.log

Comment 2 Thomas Woerner 2005-04-14 11:25:35 UTC
What are you trying to achieve with your changes?

Comment 3 murray lotnicz 2005-04-14 11:59:54 UTC
Hi Thomas, i don't understand what you mean by "changes." the sendmail config is
untouched, the cf files are unaltered since the installation. do you mean why
don't i use 8.11 instead? 

sorry, i think my description was misleading. the compiled installation of 8.13
is still running on 1 of the servers but there are 8 other rhel_ws3 servers that
are exhibiting this same problem, all with sendmail-8.12.11-4.RHEL3.1 and no
installations of 8.13. are these the changes that you're referring to?

Comment 4 Thomas Woerner 2005-04-14 13:10:36 UTC
Ok, was there a version which was not affected?

Comment 5 murray lotnicz 2005-04-14 20:30:42 UTC
yes, 8.11 had no probs sending larger msgs but my company prefers to use recent
rpms to keep on top of security vulnerabilities/misc bugfixes. i'm quite
surprised that nobody else is reporting similar problems so i wonder if it's a
unique combination of things that these servers are doing. they are dual-cpu
postgresql machines, typically with a load somewhere between 1-3. they're
booting to 2.4.21-27.0.2.ELhugemem kernels and have 16gb ram. there aren't any
other significant daemons running on these machines, just ntpd, sshd, kjournald,

we've discussed rolling back to 8.11 but are reluctant because there is no 8.11
RPM for rhel3 [was there one for rhel3_u1?] and also due to the following:

Comment 6 murray lotnicz 2005-04-20 23:23:18 UTC
Hi Thomas, i'm assuming that your answer is to revert to sendmail 8.11 since you
stopped responding. Thanks for your time, I'll ask sendmail for help.

Comment 7 Thomas Woerner 2005-04-21 09:40:30 UTC
I tried to reproduce your problem, but failed. 

Are you sure that you have no network problems or problems with ipv6?
Please use tcpdump and look at the differences between the 8.11 and 8.12 version. 

Comment 8 murray lotnicz 2005-04-28 00:54:36 UTC
after some more testing it looks like postgres, PGSTAT in particular, is causing
some havoc on the loopback interface. PGSTAT creates UDP links from localhost to
localhost to pass statistical data. stopping postgres and bouncing sendmail has
flushed the clientmqueue on 2 machines. our redhat 7.3 machines that run
postgres and sendmail don't exhibit this behavior tho. lemme know if there's any
other helpful info i can submit.

tcpdump with postgresql running:

[root@testmachine root]# tcpdump -i lo
tcpdump: listening on lo
17:16:40.783274 localhost.localdomain.32772 > localhost.localdomain.32772: udp
46 (DF)
17:16:40.784061 localhost.localdomain.32772 > localhost.localdomain.32772: udp
262 (DF)
17:16:40.794076 localhost.localdomain.32772 > localhost.localdomain.32772: udp
46 (DF)
17:16:40.794586 localhost.localdomain.32772 > localhost.localdomain.32772: udp
157 (DF)
17:16:40.795419 localhost.localdomain.32772 > localhost.localdomain.32772: udp
46 (DF)
17:16:40.795865 localhost.localdomain.32772 > localhost.localdomain.32772: udp
56 (DF)
17:16:40.796064 localhost.localdomain.32772 > localhost.localdomain.32772: udp
46 (DF)
17:16:40.796519 localhost.localdomain.32772 > localhost.localdomain.32772: udp
131 (DF)

Comment 9 Thomas Woerner 2006-08-01 14:19:27 UTC
I think this bug has to get assigned agains postgresql, it seems to be a logging

Comment 10 Tom Lane 2006-08-02 05:05:39 UTC
Um ... I see no evidence here that connects postgres to the mail problem.  I'm not saying that there 
couldn't be a connection, just that being on the same machine at the same time isn't enough connection 
to justify passing the bug my way ...

Comment 11 murray lotnicz 2006-08-05 00:15:25 UTC
i switched to postfix over a year ago which got rid of this problem. feel free
to close this bug. 

Comment 12 Tom Lane 2006-08-21 17:07:04 UTC
I'm switching this back to the sendmail component, since I see no evidence that postgres did anything 
wrong, and closing as CANTFIX because we are unable to reproduce the problem.

Note You need to log in before you can comment on or make changes to this bug.