At 12:41 19/07/2001 +0100, you wrote: >well, for me. it took 30+ seconds to clear the dead server.
>For instance, if I telneted to port 25 on the afflicted server it took
>30+ seconds to give me the usual prompt.
>This tranlated into timeouts for all of our users and a build up of half
>open connections on our server.
>took out the dead server and boom, back to normal.
That HARDLY constitutes useless. Besides, your users should be exempt from
RBL lookup. If you have specific nets that are YOUR nets, I would strongly
suggest you exempt them from RBL lookup. Am I repeating myself? I guess I do.
As far as Philip's suggestion goes, the logging of those two errors/return
codes would be a BIG step towards identifying the problem without having to
telnet to your port 25. You agree with me on that? I hope so.
