From: John Pettitt (no email)
Date: Tue Feb 01 2005 - 12:17:13 EST
Andy Kalin wrote:
>At some point in the past, we may have had a
>non-Postfix program acting somewhere in the postfix
>pipeline (original developers are gone, so hard to
>confirm); today, our intent is to have a fairly
>vanilla postfix install that acts as an outbound MTA
>only.
>
>There have been no changes to the app that relays the
>messages to postfix on this box (it has not changed
>for months and continues to relay messages at the
>expected rate; the messages then immediately begin to
>backlog on the postfix server).
>
>
>
I sounds from the error message you posted that the queue is trashed in
some way. I think I read in the thread that you stopped and started
postfix and that made things better for a while right? I suggest find
the queue entry that used the initial error (the one that have the 0+0
!= 1) and nuke it then restart postfix and see if that helps.
If you still see corrupt queue messages start thinking hardware
(assuming nobody changed anything on the box) - if the box has been up a
long time and it's not server class hardware (no ECC on the memory) then
it may be time for a reboot.
john
|
|
|