[Top] [All Lists]

Re: [ietf-smtp] SMTP server "RFC2821 Violation" for EHLO ip-literal.

2019-12-15 14:36:16
On Sun, 15 Dec 2019 18:14:30 +0100, Alessandro Vesely said:
On Sun 15/Dec/2019 17:13:38 +0100 Keith Moore wrote:

There are very many IP networks that do not use DNS but which still
use SMTP to send mail, even if only to a forwarder. Before web
browsers were commonplace, many networks did not use DNS except to
forward email to the Internet.�� The popularity of web browsers has
resulted in much wider support for DNS, but there are still many
networks that aren't intended to support humans using web browsers. �
For some of those networks, DNS is not a feature, it is a liability -
it adds nothing of value for them and is something that can break and
cause (say) assembly lines to fail.

In a walled garden it makes sense to accept mail from anyone who is
able to connect.  To do so globally has proven to be unaffordable.
Anti-spam techniques vary from dropping messages at random to
assessing senders reputation, each has its FPs.

However, consider this:

If we write a standard that says address literals aren't permitted, then the
writers of MTAs will be totally in their rights to refuse to accept mail with
an address literal.

At which point it doesn't matter if it makes sense in a walled garden, because
a device in the garden won't find anybody to talk to.

I once had to deal with the loss of a 17 terabyte LUN in a very large
filesystem - made even more troublesome by the fact that many large files ended
up with holes in the middle because they were striped across multiple LUNs.  We
ended up having to recover over 100T of data.  All caused by a vendor microcode
update borking alert e-mails from a large storage array - so we stopped getting
alerts that drives were failing. We lost 3 drives in an 8+2 RAID6, plus another
2 drives in other raidsets in a sudden flurry of failures before we got
suspicious that no drives had failed - a monthly occurrence when you have many
42U racks full of drive trays.

(For the record, the storage arrays didn't have a concept of a FQDN, and
lived in a 10.X/26 subnet. We could SSH in via a jump host, and relay SMTP
out. Although the devices *had* SNMP traps, our SNMP monitoring system
couldn't deal with a relayed trap where the SNMP packet showed up with
an IP address other than what the trap data said (our network engineering
guys got told "Fixed in the next release".  So we relied on email alerts, which
*had* been totally reliable for the first few years of the arrays.....)

Not that I'm bitter or anything.... :)

Attachment: pgpcRSgCzTT6W.pgp
Description: PGP signature

ietf-smtp mailing list
<Prev in Thread] Current Thread [Next in Thread>