Quantcast

smtp-cache problem

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

smtp-cache problem

Thomas Minor
Hello,

I might have a problem with the smtp_connection cache.
Regarding documentation, the cache is enabled on demand by default.

I found a peer site, which is also driven by postfix, with uses greylisting.
I have some 3000 mails to send to this particular server, which starts the session
by greylisting my server. Since postfix uses the smtp_connection_cache by default,
it reuses the connection which receives one greylisting temp fail after the other.
This seems to triggers the $smtpd_hard_error_limit on the other site. My server
continues to use now dead connection until the cache limits take effect.

I tend to assume, that this is a bug and maybe a configurable amount of errors should
trigger the sending postfix to abandon a connection.

Did i miss any configuration options here or do you agree?

Yours,

--Thomas


--

  Thomas Minor, Development

  H & R Netzwerk GmbH
  Am Königsweg 9
  48599 Gronau-Epe

  Sitz:            Gronau / Westf.
  Handelsregister: Amtsgericht Coesfeld, HRB 5886
  Geschäftsführer: Harald Beine

--
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: smtp-cache problem

Wietse Venema
Thomas Minor:

> Hello,
>
> I might have a problem with the smtp_connection cache.
> Regarding documentation, the cache is enabled on demand by default.
>
> I found a peer site, which is also driven by postfix, with uses greylisting.
> I have some 3000 mails to send to this particular server, which starts the session
> by greylisting my server. Since postfix uses the smtp_connection_cache by default,
> it reuses the connection which receives one greylisting temp fail after the other.
> This seems to triggers the $smtpd_hard_error_limit on the other site. My server
> continues to use now dead connection until the cache limits take effect.
>
> I tend to assume, that this is a bug and maybe a configurable amount of errors should
> trigger the sending postfix to abandon a connection.
>
> Did i miss any configuration options here or do you agree?

Why should Postfix abandon a connection that replies with 4xx to a
recipient?  How is this different from sending a multi-recipient
message?

        Wietse
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: smtp-cache problem

Viktor Dukhovni

> On Feb 20, 2017, at 7:19 PM, Wietse Venema <[hidden email]> wrote:
>
> Why should Postfix abandon a connection that replies with 4xx to a
> recipient?  How is this different from sending a multi-recipient
> message?

I guess because hard error limits may induce tarpit controls on the
receiving side?

Mind you, a server doing greylisting should probably not employ tarpit
controls.  Another reason may be that once a remote "421" disconnect is
triggered by too many consecutive errors, we may (begin to) throttle the
destination.

As you point out, with greylisting, a multi-recipient message to O(100)
recipients may run into issues rather similar to the ones just reported
with connection re-use...

--
--
        Viktor.

Loading...