timeout waiting for input from local during Draining Input
Tesla 13
Need help with 'Draining Input' error. This is going on for about 2-3 years
now and it is not getting fixed from KNOWNBUGS. It always comes back to
haunt us.
Amavis-milter scans the mail and passes it on to our custom local mailer.
Custom mailer processes it and wants to deliver using sendmail, then custom
mailer is hang, sendmail is hang, amavisd is hang and mail is not delivered.
Everything needs to be killed. No matter how big timeout values specified,
it will give 'draining input' error -- but only at the end of the timeout. I
will try to see if I can get more errors.
In the logs below, 'draining input' error is given after 2 minutes. It is
obviously issued against Timeout.quit in sendmail.cf which is:
O Timeout.quit=2m
/var/log/maillog
Mar 12 18:46:41 hitler amavis[3477]: (07161752) Passed, <te...@foo.net ->
<te...@foo.net>, Message-ID:
Mar 12 18:46:41 hitler sendmail[3483]: h2CMkQVn003483: Milter add: header:
X-Virus-Scanned: by amavisd-new
Mar 12 18:46:41 hitler clocal[3491]: foo.net: located .custom as
/home/foo/.custom
Mar 12 18:48:41 hitler sendmail[3490]: h2CMkQVn003483: timeout waiting for
input from local during Draining Input
ps auxwwwf
amavis 3678 0.0 0.2 10240 1044 pts/2 S 20:43 0:00 /usr/sbin/amavis-milter -p
local:/var/amavis/amavis-milter.sock
amavis 3683 0.0 2.9 17580 15140 ? S 20:43 0:00 /usr/bin/perl -T
/usr/sbin/amavisd
amavis 3685 0.1 3.0 17876 15732 ? S 20:43 0:00 \_ /usr/bin/perl -T
/usr/sbin/amavisd
root 3688 0.0 0.8 5040 4204 ? S 20:43 0:00 sendmail: accepting connections
root 3697 0.0 1.0 5128 5184 ? S 20:44 0:00 sendmail: h2D0hsNY003690
localhost [127.0.0.1]: data
mail 3698 0.9 1.0 5736 5096 ? S 20:44 0:00 \_ /usr/bin/perl /custom/clocal
foo.net te...@foo.net
root 3700 0.0 0.8 6072 4308 ? S 20:44 0:00 \_ /usr/sbin/sendmail -oi -f
te...@foo.net foo
strace -p 3685 (amavisd)
flock(6, LOCK_EX
strace -p 3698 (clocal)
wait4(3700,
strace -p 3700 (sendmail)
pause() = ? ERESTARTNOHAND (To be restarted)
'Draining Input' error will still be there when amavis-filter is disabled
and no amavis is used (but -DMILTER present). However, if sendmail is
compiled without -DMILTER, the problem is not there.
DETAILS:
kernel 2.4.20-grsec
glibc 2.3.1
gcc 3.2.1
Redhat 8.0 on p4 2.8 GHz
sendmail 8.12.8
Compiled with: DNSMAP LOG MAP_REGEX MATCHGECOS MILTER MIME7TO8 MIME8TO7
NAMED_BIND NETINET NETINET6 NETUNIX NEWDB PIPELINING SASL
SCANF
STARTTLS TCPWRAPPERS USERDB
OS Defines: ADDRCONFIG_IS_BROKEN HASFCHOWN HASFCHMOD HASFLOCK
HASGETDTABLESIZE HASINITGROUPS HASLSTAT HASNICE HASRANDOM
HASRRESVPORT HASSETREGID HASSETREUID HASSETRLIMIT HASSETSID
HASSETVBUF HASURANDOMDEV HASSTRERROR HASUNAME HASUNSETENV
HASWAITPID IDENTPROTO NEEDSGETIPNODE REQUIRES_DIR_FSYNC
USE_DOUBLE_FORK USE_SIGLONGJMP
Kernel symbols: /boot/vmlinux
Conf file: /etc/mail/submit.cf (default for MSP)
Conf file: /etc/mail/sendmail.cf (default for MTA)
Pid file: /var/run/sendmail.pid (default)
libsm Defines: SM_CONF_GETOPT SM_CONF_LONGLONG SM_CONF_MEMCHR
SM_CONF_MSG SM_CONF_SEM SM_CONF_SHM SM_CONF_SSIZE_T
SM_CONF_STDDEF_H SM_CONF_SYS_CDEFS_H SM_CONF_UID_GID
SM_HEAP_CHECK SM_OS=sm_os_linux SM_VA_STD
FFR Defines: _FFR_SMTP_SSL
Canonical name: foo.net
UUCP nodename: foo.net
a.k.a.: foomachine
a.k.a.: [127.0.0.1]
Conf file: /etc/mail/sendmail.cf (selected)
Pid file: /var/run/sendmail.pid (selected)
============ SYSTEM IDENTITY (after readcf) ============
(short domain name) $w = foo
(canonical domain name) $j = foo.net
(subdomain name) $m = net
(node name) $k = foo.net
========================================================
I have been checking all over the net but can't seem to find a solution. We
are unable to modify custom local mailer 'clocal'. Does anyone have a
fix/patch/solution?
Thanks in advance!
Zulya
_________________________________________________________________
Add photos to your e-mail with MSN 8. Get 2 months FREE*.
http://join.msn.com/?page=features/featuredemail