[quote user="Chris Bolton"]
Sorry, gents, I didn't make that very clear, obviously, and I also wrote that MercuryE was blacklisting when I meant MercuryS[/quote]
OK, no problem. I just wanted to be sure that we were all on the same page, so to speak, because it is important to keep straight the order of operations and the "path" of a given message through the delivery process, when attempting to troubleshoot an issue.
[quote]I have Mercury running on a general purpose PC, and the mail client which sent the blocked mail is also on this PC. I will try with a different client PC; good point, thanks.[/quote]
If you're running both Merc and the client on the same system, you will probably have the client set to "send to" the SMTP server (i.e., MercS) at IP 127.0.0.1 (i.e., the loopback interface). So MercS in turn should see that as an "incoming connection" from that same address. But the EHLO/HELO string it sees is still determined solely by the client, NOT Mercury.
[quote]I have a rule in MercuryS as follows H, "*.*", BSN, "554 Invalid HELO"[/quote]
That looks fine.
[quote]and this recently resulted in one of my own mails sent through MercuryE being rejected and MercuryS temporarily blacklisting my local IP. This didn't happen the previous time I used MercuryE.[/quote]
But here you exhibit exactly the same sort of confusion I spoke of above... At THAT point in the proceedings, MercE has not yet even seen the message in question; so it has absolutely nothing to do with this.
[quote]On investigating, the MercuryS logs showed that the EHLO that triggered the blacklist was the XP machine name, while the EHLO that was accepted previously was [192.168.0.1].[/quote]
So did you perchance change anything with the client, be that updating it to a newer version or changing the config in any way? What about the underlying system itself? Did you apply any patches/updates (including automagically via Windows Update) between when Merc worked as desired and when it started rejecting the EHLO/HELO?
[quote]Here are my MercuryS logs:[/quote]
Allow me to re-arrange an annotate these slightly, for clarity...
[quote]
(Log excerpt from March 02, 2011)
T 20110302 192228 4d604c6f Connection from 192.168.0.50
T 20110302 192228 4d604c6f EHLO [192.168.0.50]
T 20110302 192228 4d604c6f MAIL FROM:<address snipped> SIZE=8585697
T 20110302 192228 4d604c6f RCPT TO:<address snipped>
T 20110302 192228 4d604c6f DATA
T 20110302 192233 4d604c6f DATA - 116036 lines, 8585697 bytes.
T 20110302 192233 4d604c6f QUIT
T 20110302 192233 4d604c6f Connection closed with 192.168.0.50, 5 sec. elapsed.
(Log excerpt from March 28, 2011)
T 20110328 203649 4d8f5ebd Connection from 192.168.0.50
T 20110328 203649 4d8f5ebd EHLO user281abc2f54
X 20110328 203649 4d8f5ebd 554 Invalid HELO
T 20110328 203649 4d8f5ebd Connection closed with 192.168.0.50, 0 sec. elapsed.
E 20110328 203658 0 Connection from 192.168.0.50 refused because of short-term restriction.
[/quote]
This confirms what we've both been saying... At some point between March 02 and March 28, *something* caused the mail client (i.e., whatever program you use to read and compose e-mail messages) to change the way it identifies itself when sending mail. While the earlier form of the ID is strictly "legal" per RFC2821, it's generally frowned upon; so maybe that is the underlying "why" some intervening update to either the mail client or the operating system caused that behavior to change?
Also... While not necessarily relevant to your immediate problem, I'm a little concerned that this machine is apparently identifying itself to itself by its "external" IP address. That IP literal really should read: "[127.0.0.1]". Do you have an entry in your local "hosts" file which looks like this:
127.0.0.1 localhost
???
If not, I'd suggest adding such ASAP. Who knows..? It might even fix your HELO/EHLO problem (tho' I rather doubt that).
[quote]I agree neither of these is suitable for my machine to be using as EHLO, and I don't know why it's happening.[/quote]
Well, as I mentioned above, use of an IP Literal IS "legal"; it's just considered "poor form" in some quarters.
[quote]In both the core module "Internet name for this system" and MercuryE "Introduce myself as" I have entered my domain name, cjbolton.plus.com - this has been unchanged since I set Mercury up several years ago. This info is shown in the config screens and in the following extracts from mercury.ini (the MercuryE section was what I included in my first post)[/quote]
That's all well and good, but completely irrelevant. In both cases, those parameters apply ONLY to:
a.) how Mercury (via either MercE or MercC) identifies itself when sending outgoing mail that has already passed through MercS.
b.) the "220- ..." greeting message that MercS sends to the client immediately upon seeing an incoming connection.
(Hmmm... On further thought: It *may* also come into play when pulling incoming mail from an external POP3 server via the MercD Distributing POP3 Client; but since I don't use that module, I can't be sure.)
[quote]The rule is not the problem, as you say; it's just the mechanism by which my odd EHLO has been identified.
Hope this makes sense now; thanks for your help.[/quote]
You're quite welcome. Hopefully the above will be helpful; but the upshot is, you need to be looking at the client program, not Mercury, for the solution. In the meantime, as Rolf pointed out, you can whitelist the localhost to exempt it from Transaction Filtering, and thus side-step the whole issue.
[quote user="Chris Bolton"]
Sorry, gents, I didn't make that very clear, obviously, and I also wrote that MercuryE was blacklisting when I meant MercuryS[/quote]
OK, no problem.&nbsp; I just wanted to be sure that we were all on the same page, so to speak, because it is important to keep straight the order of operations and the "path" of a given message through the delivery process, when attempting to troubleshoot an issue.
[quote]I have Mercury running on a general purpose PC, and the mail client which sent the blocked mail is also on this PC. I will try with a different client PC; good point, thanks.[/quote]
If you're running both Merc and the client on the same system, you will probably have the client set to "send to" the SMTP server (i.e., MercS) at IP 127.0.0.1 (i.e., the loopback interface).&nbsp; So MercS in turn should see that as an "incoming connection" from that same address.&nbsp; But the EHLO/HELO string it sees is still determined [b]solely[/b] by the client, [b]NOT[/b] Mercury.
[quote]I have a rule in MercuryS as follows&nbsp;&nbsp;&nbsp;&nbsp; H, "*.*", BSN, "554 Invalid HELO"[/quote]
That looks fine.
[quote]and this recently resulted in one of my own mails sent through MercuryE being rejected and MercuryS temporarily blacklisting my local IP. This didn't happen the previous time I used MercuryE.[/quote]
But here you exhibit exactly the same sort of confusion I spoke of above...&nbsp; At [b]THAT[/b] point in the proceedings, MercE has not yet even seen the message in question; so it has [b]absolutely nothing[/b] to do with this.
[quote]On investigating, the MercuryS logs showed that the EHLO that triggered the blacklist was the XP machine name, while the EHLO that was accepted previously was [192.168.0.1].[/quote]
So did you perchance change anything with the client, be that updating it to a newer version or changing the config in any way?&nbsp; What about the underlying system itself?&nbsp; Did you apply any patches/updates (including automagically via Windows Update) between when Merc worked as desired and when it started rejecting the EHLO/HELO?
[quote]Here are my MercuryS logs:[/quote]
Allow me to re-arrange an annotate these slightly, for clarity...
[quote]
(Log excerpt from March 02, 2011)
T 20110302 192228 4d604c6f Connection from 192.168.0.50
T 20110302 192228 4d604c6f EHLO [192.168.0.50]
T 20110302 192228 4d604c6f MAIL FROM:&lt;address snipped&gt; SIZE=8585697
T 20110302 192228 4d604c6f RCPT TO:&lt;address snipped&gt;
T 20110302 192228 4d604c6f DATA
T 20110302 192233 4d604c6f DATA - 116036 lines, 8585697 bytes.
T 20110302 192233 4d604c6f QUIT
T 20110302 192233 4d604c6f Connection closed with 192.168.0.50, 5 sec. elapsed.
(Log excerpt from March 28, 2011)
T 20110328 203649 4d8f5ebd Connection from 192.168.0.50
T 20110328 203649 4d8f5ebd EHLO user281abc2f54
X 20110328 203649 4d8f5ebd 554 Invalid HELO
T 20110328 203649 4d8f5ebd Connection closed with 192.168.0.50, 0 sec. elapsed.
E 20110328 203658 0 Connection from 192.168.0.50 refused because of short-term restriction.
[/quote]
This confirms what we've both been saying...&nbsp; At some point between March 02 and March 28, *something* caused the mail [b]client[/b] (i.e., whatever program you use to read and compose e-mail messages) to change the way it identifies itself when sending mail.&nbsp; While the earlier form of the ID is strictly "legal" per RFC2821, it's generally frowned upon; so maybe that is the underlying "why" some intervening update to either the mail client or the operating system caused that behavior to change?
Also...&nbsp; While not necessarily relevant to your immediate problem, I'm a little concerned that this machine is apparently identifying itself [b]to itself[/b] by its "external" IP address.&nbsp; That IP literal really [b]should[/b] read: "[127.0.0.1]".&nbsp; Do you have an entry in your local "hosts" file which looks like this:
&nbsp;&nbsp; &nbsp;127.0.0.1&nbsp;&nbsp; &nbsp;&nbsp;&nbsp; &nbsp;localhost
???
If not, I'd suggest adding such ASAP.&nbsp; Who knows..?&nbsp; It might even fix your HELO/EHLO problem (tho' I rather doubt that).
[quote]I agree neither of these is suitable for my machine to be using as EHLO, and I don't know why it's happening.[/quote]
Well, as I mentioned above, use of an IP Literal [b]IS[/b] "legal"; it's just considered "poor form" in some quarters.
[quote]In both the core module "Internet name for this system" and MercuryE "Introduce myself as" I have entered my domain name, cjbolton.plus.com&nbsp; - this has been unchanged since I set Mercury up several years ago. This info is shown in the config screens and in the following extracts from mercury.ini (the MercuryE section was what I included in my first post)[/quote]
That's all well and good, but completely irrelevant.&nbsp; In both cases, those parameters apply ONLY to:
a.)&nbsp; how Mercury (via either MercE or MercC) identifies itself when sending outgoing mail that has [b] already[/b] passed through MercS. &nbsp;
b.)&nbsp; the "220- ..." greeting message that MercS sends to the client immediately upon seeing an incoming connection.
(Hmmm...&nbsp; On further thought:&nbsp; It *may* also come into play when pulling incoming mail from an external POP3 server via the MercD Distributing POP3 Client; but since I don't use that module, I can't be sure.)
[quote]The rule is not the problem, as you say; it's just the mechanism by which my odd EHLO has been identified.
Hope this makes sense now; thanks for your help.[/quote]
You're quite welcome.&nbsp; Hopefully the above will be helpful; but the upshot is, you need to be looking at the [b]client[/b] program, not Mercury, for the solution.&nbsp; In the meantime, as Rolf pointed out, you can whitelist the localhost to exempt it from Transaction Filtering, and thus side-step the whole issue.