[Mimedefang] My MD install went wacko

[Bill Randle]

On Sun, 2003-06-08 at 16:15, Justin Shore wrote:
> Jun  8 17:20:03 bubba sm-mta[20150]: h58MII3E020150: 
> milter_read(mimedefang): cmd read returned 0, expecting 5
> Jun  8 17:20:03 bubba sm-mta[20150]: h58MII3E020150: Milter (mimedefang): 
> to error state
> Jun  8 17:20:03 bubba sm-mta[20150]: h58MII3E020150: Milter: data, 
> reject=451 4.7.1 Please try again later
> 
> and started temp failing the message (which it did 5 more times).  This is 
> weird.
> 
> I'm not literally restarting Sendmail and MIMEDefang from cron every 5 
> minutes.  Any ideas?

Justin,

This sounds suspiciously like the problems I was having (see the thread
"[Mimedefang] mimedefang lockup - help needed" and related followups
[one of which was a reply from you :-)]. In my case the mimedefang
processes were dieing for unknown reasons.

When the system is hosed and you do a "ps -Alw -y -H | grep mime", what
does it look like (compared to when the system is normal)?

In my case, when it was wacko, the ps tree looked like this:
(some columns deleted for brevity)

S     PID  PPID   RSS   WCHAN  TTY          TIME   CMD
S   23796     1   288   do_sel ?        00:00:00   mimedefang-mult
S   23797 23796 116888  pipe_w ?        00:03:04     mimedefang.pl
S   23833 23796  14304  pipe_w ?        00:00:35     mimedefang.pl
S   25103 23796  21944  pipe_w ?        00:00:05     mimedefang.pl
S   25110 23796  21860  pipe_w ?        00:00:05     mimedefang.pl
S   23821     1   696   rt_sig ?        00:00:00   mimedefang

Whereas during normal operation, it was more like this:

S     PID  PPID   RSS   WCHAN  TTY          TIME   CMD
S   30756     1   288   do_sel ?        00:00:11   mimedefang-mult
S   31641 30756  15644  pipe_w ?        00:04:56     mimedefang.pl
S    2855 30756  13076  pipe_w ?        00:00:32     mimedefang.pl
S    6208 30756  22792  pipe_w ?        00:00:05     mimedefang.pl
S   30779     1   444   do_sel ?        00:00:02   mimedefang
S   30783 30779   444   do_pol ?        00:00:04     mimedefang
S   30784 30783   444   rt_sig ?        00:00:00       mimedefang
S   32414 30783   444   do_sel ?        00:00:00       mimedefang
S    1178 30783   444   do_sel ?        00:00:00       mimedefang
S    1353 30783   444   do_sel ?        00:00:00       mimedefang

I found restarting mimedefang would usually clear it up for a while.

In my case, I never did figure out what the underlying cause was.
After upgrading to MIMEDefang-2.34-BETA-5 (and previously to kernel
2.4.19) fixed it.

I know you said you're already using 2.34-BETA-5, but it would be
interesting to know if your mimedefang processes are dieing off
and not restarting like it was for me. It might help David or
someone else track down the problem.

	-Bill