SMTP mail and "unreachable" node

[Bill Pedersen]

I have an issue with mail to some nodes. I get a failure that they are "unreachable", for instance:

%TCPIP-E-SMTP_UNREACHABL, cannot connect to remote host, ingres.com -SYSTEM-F-UNREACHABLE, remote node is not currently reachable

I can ping the node.

But when I traceroute the node it goes for 15 hops and then times out:

traceroute to ingres.com (98.129.177.130): 1-30 hops, 38 byte packets
1 213.173.170.9 (213.173.170.9) 0.0 ms 0.0 ms 0.0 ms
2 atm6-0-1115.ar3.gva.ch.colt.net (212.23.238.121) 10.0 ms 10.0 ms 0.0 ms
3 ge3-0-0-pr2.FRA.router.colt.net (212.74.76.191) 20.0 ms (ttl=248!) 20.0 ms
(ttl=248!) 10.0 ms (ttl=248!)
4 xe-1-2-0.mpr1.fra4.de.above.net (80.81.194.26) 10.0 ms (ttl=246!) 20.0 ms
(ttl=246!) 10.0 ms (ttl=246!)
5 xe-1-1-0.mpr1.cdg12.fr.above.net (64.125.24.6) 50.0 ms (ttl=246!) 30.0 ms
(ttl=246!) 30.0 ms (ttl=246!)
6 * xe-0-0-0.mpr1.cdg11.fr.above.net (64.125.31.225) 20.0 ms (ttl=245!) 30.0
ms (ttl=245!)
7 xe-3-1-0.mpr1.lhr2.uk.above.net (64.125.24.1) 30.0 ms 40.0 ms 30.0 ms
8 so-0-1-0.mpr1.dca2.us.above.net (64.125.27.57) 110 ms (ttl=249!) 120 ms (t
tl=249!) 110 ms (ttl=249!)
9 ge-2-0-0.mpr3.iah1.us.above.net (64.125.25.114) 130 ms (ttl=249!) 160 ms (
ttl=249!) 130 ms (ttl=249!)
10 xe-0-3-0.cr1.iah1.us.above.net (64.125.30.101) 130 ms (ttl=249!) 130 ms (t
tl=249!) 130 ms (ttl=249!)
11 xe-0-1-0.er1.iah1.us.above.net (64.125.28.5) 140 ms (ttl=246!) 140 ms (ttl
=246!) 150 ms (ttl=246!)
12 xe-7-0-0.er1.dfw2.us.above.net (64.125.30.58) 150 ms (ttl=243!) 150 ms (tt
l=243!) 150 ms (ttl=243!)
13 main1.above.net (209.133.126.42) 140 ms 150 ms 150 ms
14 vlan905.core5.dfw1.rackspace.com (67.192.56.229) 150 ms 150 ms 140 ms
15 67.192.56.31 (67.192.56.31) 150 ms 160 ms 150 ms
16 * * *
17 * * *
18 * * *
19 * * *
20 * * *
21 * * *
22 * * *
23 * * *
24 * * *
25 * * *
26 * * *
27 * * *
28 * * *
29 * * *
30 * * *
TCPIP>

SHOW HOST (nslookup) works fine:

TCPIP> show host ingres.com

BIND database

Server: 81.23.73.120

Host address Host name

98.129.177.130 INGRES.COM
TCPIP>

This is on the following configuration:

HP TCP/IP Services for OpenVMS Alpha Version V5.6
on an AlphaServer ES40 running OpenVMS V8.3


Thoughts?

I get this for more than one site. And have gotten it off and on for some time. Is there a logical that needs to be set regarding name translation or something for SMTP?

Thanks,

Bill.

[Jim_McKinney]

> Thoughts?

If HP's SMTP is well behaved it would not be attempting to communicate with the address associated with ingres.com. ingres.com has DNS MX records registered and any SMTP mail should be directed to them. You might want to see if they are reachable. (You may also discover that other SMTP targets that you are having issues with also have MX registration.)

A$ mu nsl/typ=mx ingres.com
Server: LOCALHOST
Address: 127.0.0.1

Non-authoritative answer:
INGRES.COM preference = 10, mail exchanger = proofpoint1.invision.net
INGRES.COM preference = 10, mail exchanger = proofpoint2.invision.net

Authoritative answers can be found from:
INGRES.COM nameserver = pdns4.ultradns.org
INGRES.COM nameserver = pdns1.ultradns.net
INGRES.COM nameserver = pdns5.ultradns.info
INGRES.COM nameserver = pdns3.ultradns.org
INGRES.COM nameserver = pdns6.ultradns.co.uk
INGRES.COM nameserver = pdns2.ultradns.net
pdns4.ultradns.org internet address = 199.7.69.1
pdns4.ultradns.org IPv6 address = 2001:0502:4612::0001
pdns5.ultradns.info internet address = 204.74.114.1
pdns6.ultradns.co.uk internet address = 204.74.115.1
pdns1.ultradns.net internet address = 204.74.108.1
pdns1.ultradns.net IPv6 address = 2001:0502:f3ff::0001
pdns2.ultradns.net internet address = 204.74.109.1
pdns3.ultradns.org internet address = 199.7.68.1
A$ mu nsl proofpoint1.invision.net
Server: LOCALHOST
Address: 127.0.0.1

Non-authoritative answer:
Name: PROOFPOINT1.INVISION.NET
Address: 69.18.136.194

A$ mu nsl proofpoint2.invision.net
Server: LOCALHOST
Address: 127.0.0.1

Non-authoritative answer:
Name: PROOFPOINT2.INVISION.NET
Address: 69.18.136.195

A$

[Robert Gezelter]

Bill,

Please also note that there is no guarantee that ICMP (PING) and SMTP (mail) go to the same actual machine. Many traffic management devices to differential routing based on port number in addition to IP address.

One good step is to use a telnet session to see what comes back when you connect to the SMTP port.

- Bob Gezelter, http://www.rlgsc.com

[Hoff]

If you are working from an OpenVMS host or cluster at 213.173.170.9, then your local DNS looks to be misconfigured, and that'll definitely get your SMTP flagged either for graylisting or blocked as spam.

[Bill Pedersen]

Ok, yes, A and MX records can be different the issue I see though is that the traceroute works from my laptop, located on an entirely different network and location while it does not from the OpenVMS system which I keep having issues with.

Hoff: You bring up the issue of mis-configuration.

Here is the config information:

TCPIP> sho name

BIND Resolver Parameters

Local domain: stromasys.com

System

State: Started, Enabled

Transport: UDP
Domain: stromasys.com
Retry: 2
Timeout: 5
Servers: 81.23.73.120, 81.23.73.90
Path: No values defined

Process

State: Enabled

Transport:
Domain:
Retry:
Timeout:
Servers:
Path:
TCPIP>

Other than servers am not sure what can be done here...

Enter configuration option: 4

BIND RESOLVER Configuration

A BIND resolver has already been configured.

BIND Resolver Configuration

Transport: UDP
Domain: stromasys.com
Retry: 2
Timeout: 5
Servers: ns1.Cybernetwork.CH, ns2.Cybernetwork.CH
Path: No values defined

* Do you want to reconfigure BIND [NO]:

The SHOW MX_RECORD gives:

TCPIP> sho mx ingres.com


BIND MX database


Server: 81.23.73.120

Gate address Preference Gate name

69.18.136.194 10 proofpoint1.invision.net
69.18.136.195 10 proofpoint2.invision.net


BIND database


Server: 81.23.73.120

Host address Host name

98.129.177.130 INGRES.COM

So, now where?

Bill.

[Bill Pedersen]

On the issue of MX records... I really do not see where this is taking us since my own domain is defined by MX records and I have no issue sending mail to myself from this remote system.

The MX information for my domain is below:

TCPIP> sho mx ccsscorp.com


BIND MX database


Server: 81.23.73.120

Gate address Preference Gate name

207.111.237.9 5 mx2005.mx2.got.net
207.111.237.8 5 mx2004.mx2.got.net
209.66.101.141 5 mx2003.mx2.got.net
209.66.101.140 5 mx2002.mx2.got.net
209.66.101.139 5 mx2001.mx2.got.net
207.111.237.10 5 mx2006.mx2.got.net


BIND database


Server: 81.23.73.120

Host address Host name

207.111.217.162 CCSSCORP.COM

This is in contrast to the MX information for ingres.com...

Now, if there is an issue of graylisting then I need to understand how to get around that...

The number of sites which do not get email from this site is probably 5% or less of the total number I send to as people sign up but I would like to think it should be near zero if they provide valid mail addresses.

Thanks,

Bill.

Thanks,

Bill.

[Jim_McKinney]

If a DNS MX record for a particular host name exists, then any mail for that host name should be delivered to the host specified by the associated MX record(s)(and not the host specified by the A(ddress) resource record). In the absence of MX records SMTP mail will be directed to the host specified by the A record. SMTP mail from your host to ingres.com will be sent to one of the hosts specified by the associated MX records (and not that address you were PINGing) presuming that HP's SMTP client software plays by the rules. In practice many clients ignore the MX records and always target the A records. For this reason most enterprises have assigned an IP address to their domain name rather than risk losing mail from braindead clients. I would expect that HP's SMTP client is doing the correct thing and targeting the MX records (you could verify this by using TCPDUMP).

[Bill Pedersen]

Hoff:

The 213.173.170.9 address is the gateway. I believe it is a router but I do not have the details, since I did not install the net.

The OpenVMS system is at 213.173.170.13.

Bill.

[Hoff]

213.173.170.13 within your DNS environment looks to be misconfigured. No valid rDNS response to dig.

Some of the more permissive SMTP servers around will accept that, but many of the servers will detect this and the message will get grey-listed, or will be flagged as spam and dropped.

The MX entry is the routing path from the client to the server.

The rDNS setting and associated detection is the reverse; the view from the SMTP server back to the originating mail server.

[Bill Pedersen]

Yes, I expect the SMTP is in fact looking at the MX records.

Ok, rDNS. So how does one set this? What settings are necessary or capable withing TCPIP, or do we need a DNS entry someplace which gives this information.

When I do nslookup for 213.173.170.13 and charon.stromasys.com I get:

C:\Documents and Settings\wpederse>nslookup 213.173.170.13
*** Can't find server name for address 192.168.1.1: Query refused
*** Default servers are not available
Server: UnKnown
Address: 192.168.1.1

Name: charon.stromasys.com
Address: 213.173.170.13


C:\Documents and Settings\wpederse>nslookup charon.stromasys.com
*** Can't find server name for address 192.168.1.1: Query refused
*** Default servers are not available
Server: UnKnown
Address: 192.168.1.1

Non-authoritative answer:
Name: charon.stromasys.com
Address: 213.173.170.13

So what more do we need to do?

Thanks,

Bill.