Thanks, your last sentence was the one I need to understand my problem. Bye :)
On Tue, Sep 28, 2010 at 12:31 PM, Victor Duchovni <victor.ducho...@morganstanley.com> wrote: > On Tue, Sep 28, 2010 at 11:24:47AM -0500, Jason Voorhees wrote: > >> I know that Postfix defaults virtual_alias_expansion_limit directive >> to 1000. I have a question: >> >> What happens if I have a virtual_alias_maps that returns more than >> 1000 results? Will postfix will send e-mail to the first 1000 results >> and ignore the rest from 1001? > > What happens is that expansion stops, and any remaining addresses > remain unexpanded, and may generate bounces (if they are in a virtual > alias domain). > > The behaviour when the limit is exceeded is not ideal. It is probably > better to reject the message. Wietse and I discussed this issue off-list > about a year ago, don't recall which, if either, of us was going to look > into it further... > > Because recursive virtual expansion happens in the cleanup(8) server, > it is not possible to reject a single SMTP recipient that expands to a > list over the limit. Rather, the entire message would have to be rejected > after "." with a "queue-file write error" (and a more specific message > in the mail logs). > > It is perhaps time to consider doing virtual expansion in the SMTP server > for a future Postfix 3.0 release. That would potentially allow wild-card > rewrites to co-exist with recipient validation. > >> What's the postfix behaviour in this cases when a virtual_alias_maps >> returns more than 1000 (default) results? >> >> a) It causes a bounce to ALL results of the virtual_alias_maps? > > No. > >> b) It delivers the first 1000 then bounces the rest over the default limit? > > Only the first 1000 undergo expansion, the others are not subjected to > virtual alias rewriting, and this may cause delivery to fail. > > -- > Viktor. >
