Dear all,
I'm hitting a problem on our mail server lately whereby some messages
sent to custom list software are bouncing when they really shouldn't be.
The list is setup so that incoming mails are passed to a list program
using the pipe transport. The timeout on the pipe is left at the default
1 hour.
Here is a partial log entry for an incoming message that
2004-01-15 23:07:22 1Ah89a-0007mv-Pu <= abc123@??? H=mail107.abc.com (abc.com) [x.138.203.7] P=smtp S=526 id=IMTn2W68077a8dca0EoT@???
2004-01-15 23:16:03 1Ah89a-0007mv-Pu Spool file is locked (another process is handling this message)
2004-01-15 23:46:02 1Ah89a-0007mv-Pu Spool file is locked (another process is handling this message)
2004-01-16 00:07:22 1Ah89a-0007mv-Pu ** yura@??? R=my_lists T=my_list_pipe: pipe delivery process timed out
2004-01-16 00:07:22 1Ah95e-0008NN-VP <= <> R=1Ah89a-0007mv-Pu U=exim P=local S=1304 (The message gets bounced to my local debug address)
2004-01-16 00:07:22 1Ah89a-0007mv-Pu Completed
You can see the message comes in and I guess a pipe delivery also gets
started although not logged at 23:07:22 also. This pipe delivery times
out after 1 hour at 00:07:22. The spool file locked entries can be
ignored as they are just a result of other queue runners also trying to
send the message while its on the queue.
On receiving a message, the mail list software usually logs receipt
immediately before processing. It never logged anything for this
message.
My thinking is that its an obscure bug in exim as a quick search on
google for "exim" and "pipe delivery process" also turns up problems
other people are having.
So, two things:
A) Is there any option in exim that would cause such a failure to be
retried. The pipe works the majority of the time so a retry would more
than likely work. This would be a great short term solution until a real
solution can be found.
B) Is it an exim problem or my problem? Any ideas?
Ah, nearly forgot. I'm running RH8 with exim 4.30.
Best regards,
Tom.
