General SMTP/MTA Question

Timothy timothy at jupiter.stcl.edu
Sun Jan 14 02:46:13 UTC 2007


On Saturday 13 January 2007 19:31, Evan Klitzke wrote:
> I am thinking about running an MTA on my home server. At home I am
> connected to the Internet through a regular DSL account, so my IP
> address is not static, and every few days when I get a new DHCP lease
> and the DNS record for my domain has to be updated, which takes about
> half an hour. In the meantime, the DNS record for my domain will point
> to the old IP address. This isn't really a problem with my HTTP server,
> because it is mostly for personal use, so when it isn't reachable it
> isn't a big problem. However, I would be concerned about losing email if
> an MTA tried to pass mail to me and found that they couldn't reach my IP
> address, or that port 25 was blocked on that address.
This is what I did:
1. You have to request port 25 to be open via your provider.  Mine is AT&T, 
and they had it blocked.  I just filled out an online form and in about 24-48 
hours all was well.
2. Register your domain at noip.org.
3. Install their noip2 app which keeps your domain associated with whatever 
your ip is at the moment. Works really good.
>
> How resistant are most MTAs to this kind of failure? In other words, how
> realistic is it to expect that someone else's MTA trying to deliver mail
> to mine will hold it in its mail queue and retry sending long enough for
> it to get through? Will this actually happen, or would a typical MTA
> just drop the message from its mail queue if the domain wasn't reachable
> (or not accepting connections on port 25)? I am hoping some veteran
> sysadmins on the list will have enough experience to know what to
> expect.
I mostly use postfix as my mta and I believe the default is to retry once an 
hour for 3 days.  I believe sendmail is along the same lines.

Cheers,

Timothy

>
> Thanks a lot,
> Evan Klitzke




More information about the redhat-list mailing list