Josh Cason put forth on 5/13/2010 5:13 PM:

> But this is a static ip number and the mail server it is using is mine.
> These are customers of ours that we monitor our servers. Now If I put
> the extact ip address into mynetworks. It works. But I don't think that
> is proper.

Why would this not be "proper"?  By using IP addresses you're merely
avoiding name lookups, either via dns or a host file.  If you use dns and
the name server goes down or is unreachable for any number of reasons, this
process fails.  A host file is more reliable but it still adds processing
time.  If you use IP addresses, the network itself has to fail or one of the
remote hosts in order to cause a problem.

There are two upsides I can think of WRT using names instead of bare IPs:

1.  If the IP addresses of the remote servers change often (they probably
don't).

2.  Identification purposes within your main.cf.  In this case, you can just
add a remark like (#) next to the bare IP addresses

If bare IPs work, stick with them.  Obviously you have some dns issues.  If
this is a frequent occurrence, you should definitely use bare IPs.  That
will keep this process working when you or your coworkers or provider(s)
screw up your dns.

-- 
Stan






Reply via email to