sorry, I normally lurk on a list for a while before I start posting to
get the feel, but if I don't send this soon, I'll forget all about it an
it'll be lost to the ether forever.
I have a VM and have been working on figuring out why postfix [via
centos 6] wasn't working the way I wanted, I'd get odd results like this:
[root@orca postfix]# postfix stop
postfix/postfix-script: stopping the Postfix mail system
/usr/libexec/postfix/postfix-script: line 153: kill:
ELF>@@8,�: arguments must be process or job IDs
/usr/libexec/postfix/postfix-script: line 153: kill: X�s2��:
arguments must be process or job IDs
/usr/libexec/postfix/postfix-script: line 153: kill: �t2�:
arguments must be process or job IDs
/usr/libexec/postfix/postfix-script: line 153: kill: : arguments
must be process or job IDs
/usr/libexec/postfix/postfix-script: line 153: kill:
��2�0��2�@��2��@�}8�pP@8�:
arguments must be process or job IDs
/usr/libexec/postfix/postfix-script: line 153: kill:
P@�8�`P�8�p`�8�@@���8�p��8�� ��8���8��:
arguments must be process or job IDs
/usr/libexec/postfix/postfix-script: line 153: kill:
�8�����8�: arguments must be process or job IDs
/usr/libexec/postfix/postfix-script: line 153: kill:
���8�@@0�8�@�8�PP�`�8�P��:
arguments must be process or job IDs
postfix/postfix-script: waiting for the Postfix mail system to terminate
postfix/postfix-script: waiting for the Postfix mail system to terminate
postfix/postfix-script: waiting for the Postfix mail system to terminate
postfix/postfix-script: waiting for the Postfix mail system to terminate
^C
I started digging into the script to figure out what is doing what/where
[root@orca postfix]# vi /usr/libexec/postfix/postfix-script
[root@orca postfix]# cd /var/
[root@orca var]# find . -name "master.pid"
./spool/postfix/pid/master.pid
[root@orca var]# file spool/postfix/pid/master.pid
it seems the master.pid file was corrupted by some other process,
overwriting with a core file
spool/postfix/pid/master.pid: ELF 64-bit LSB core file x86-64,
version 1 (SYSV), SVR4-style, from '/usr/sbin/abrtd'
[root@orca var]# rm spool/postfix/pid/master.pid
rm: remove regular file `spool/postfix/pid/master.pid'? y
[root@orca var]# ls spool/postfix/pid/
unix.cleanup unix.defer unix.local unix.retry unix.smtp
[root@orca var]# postfix stop
postfix/postfix-script: fatal: the Postfix mail system is not running
[root@orca var]# postfix start
postfix/postfix-script: starting the Postfix mail system
[root@orca var]# postfix status
postfix/postfix-script: the Postfix mail system is running: PID: 19579
[root@orca var]# ps auxww | grep pos
removed and everything ended up working fine. it seems that there
isn't any validation being done for the contents of the pid file and it
isn't being removed if things go awry (or exit status checked, a few
places where things could be done differently). if I make such
changes/updates, is there a place I can submit my checks to be included
in the next version/patch of postfix? baring that, is there someone
else who'd like to make such changes? I may have hit a very odd edge
case that may well never, ever happen to anyone again, but it still
seems like a reasonable set of changes to make (at least to me).
I believe this VM got into this state due to some upgrade problems I
experienced. specifically I think I had 2 different processes/instances
of the VM attempting to run at the same time due to the way I autostart
VMs, the way the hypervisor software doesn't really deal with locking
and partially complete upgrade to the hypervisor software.
possibly relevant version info
[root@orca var]# cat /etc/redhat-release
CentOS release 6.6 (Final)
[root@orca var]# rpm -q postfix
postfix-2.6.6-6.el6_5.x86_64
thanks
--
public gpg key id: AE60F64C
