Hello!
On Wed, Dec 08, 2004 at 11:13:10AM -0600, wayne wrote:
[...]
SRS has an option to do DB-backed forwarding in all cases and this
must be implemented so that there is a fallback in cases where the
MTA being forwarded to rejects the MAIL FROM due to excessive
localpart lengths.
So, what you have described is a subset of SRS.
Ok, our company has a definite decision: No database for return path
rewriting. Period. It's a maintenance nightmare, a single point of
failure in an otherwise redundant system, and doesn't scale.
But there was the intention to *do* some kind of sender rewriting, just
the requirement that it must be stateless.
There's code to implement the hash-based SRS, but you indicate that
that's not enough.
Suggestions?
Kind regards,
Hannah.