Josh Remus writes:
« HTML content follows »
I've got a problem that I encountered this morning. We have a 384 pipe in
and out of here, and one of our marketing people sent out a mass-emailing
(1.31 MB a piece) that completely clobbered our bandwidth. But of course,
courier handled itself just as I had hoped it would, it had no problem
with the number.
When I got in, people were screaming about the bandwidth. Turning down
the maximum number of ESMTP connections (which doesn't seem to like less
than 40 for that setting) doesn't cut it. So - here's what I had to do: I
You are probably tuning the maximum number of incoming smtp connections,
which of course doesn't affect outgoing mail.
For outgoing mail, you need to twiddle etc/module.esmtp; turning down
MAXDELS (but never below MAXHOST), and then restart Courier.
moved the folders under /var/spool/courier/msgs to someplace temporary,
since there were 800 in the queue. This stopped the flood for the moment
& let the normal email go in and out again.
I don't know if this is kosher at all, but courier seemed to handle those
folders disappearing okay. But I need two things:
Well, it's not exactly 100% kosher; you should have a bunch of complaints in
syslog somewhere, but that's about it.
1. I need to reintroduce those messages a little at a time, because
while I could cancel the marketing email, some legit email probably got
moved along with all the marketing junk.
2. I need to throttle the outgoing connections or something like that
so that the mail server can not normally consume all that bandwidth, or
all that bandwidth for all that time.
Ideally (which never usually happens) - I could make some changes to the
configs on the mail server and somehow move all the messages back in -
then they would slowly deliver themselves while keeping some of the
bandwidth available for other things.
Stop Courier, turn down max # of connections to 8. Make sure that there are
no links to the pulled messages in the msgq directory. There are hard links
from msgq to msgs, which you didn't do anything about, which is the main
problem. Go through all the folders you took out, and make sure that every
file has exactly one link. If there are more than one link to a file, the
second link is somewhere in msgq. Find it and remove it.
Then move all the files in all the removed folders (NOT copy them) to
/var/spool/courier/tmp (or /usr/lib/courier/var/tmp, as the case may be).
There's going to be one or two numeric subdirectories there. All files
should be moved into the last numeric subdirectory. Then restart Courier.
Hopefully things will work now. If not, you'll just have to make sure
everything's stopped, then blow away everything in msgs and msgq
subdirectories (but not the subdirectories themselves) and restart; and
write off the lost mail as a learning experience.
-------------------------------------------------------
This SF.net email is sponsored by: ObjectStore.
If flattening out C++ or Java code to make your application fit in a
relational database is painful, don't do it! Check out ObjectStore.
Now part of Progress Software. http://www.objectstore.net/sourceforge
_______________________________________________
courier-users mailing list
[EMAIL PROTECTED]
Unsubscribe: https://lists.sourceforge.net/lists/listinfo/courier-users
