Drowing in an ocean of email

Michael Emdy michael at emdy.com
Mon Jan 12 15:16:22 GMT 2004


Help! We run an MS Exchange server internally and we have been plagued
with high load and crashes until we implemented a linux gateway and
chose to run MailScanner with Spam Assassin, RBL's (spamhaus and
spamcop), and use the sa-learn script to train our bayesian DB.  Now we
gained stability internally, but the # of msgs per day are steadily
climbing and at this rate should overwhelm even our linux gateway.  We
are taking upwards of 70-100k msgs a day for a company with just under
1000 employees.  Our biggest problem currently is email destined for
user email addresses that do not exist within our company.  What finally
happens is the "Bad Mail" queues up on our MS Exchange server and cannot
be delivered and just gets shoved into a folder for the admin to deal
with.  Which basically we just have to go in on an hourly basis and
delete the mail.  Before figured out what was happening we had about
25GB of mail to delete from the folder, now we do it hourly until we can
find a solution.

I've configured Sendmail to not accept unresolvable domains, and I
implemented the domain lookup as well.  We still seem to be getting
slammed by something.  We are running out of options.

Does anyone have any suggestions on what else we could do to start
cutting down our daily msg intake?

Any help would be greatly appreciated !

Thanks in advance,

MTE



More information about the MailScanner mailing list