On a GW 6.5 SP1 system, I am still getting the '450 Host Down' Error at 
times when an outbound message is going to our own domain.  In the log 
file it looks like: 

"DMN: MSG 5494 Send Failure: 450 Host down (MYDOMAIN.COM)"

Where MYDOMAIN.COM is our own Internet domain.  I have followed the 
suggestions of TID 10060050 and put an entry in the Route.cfg file for 
the GWIA like so:

mydomain.com  [192.168.100.10]

Where 192.168.100.10 is the internal private IP address of our 
BorderManager server that is also running the GWIA.

Why do messages that originate in our domain and are sent to 
'mydomain.com' ever make it to the GWIA in the first place?  What can 
we do to prevent the Host Down errors as per above?

Thanks,

Dan

This is something, that pop's up a lot of time on this newsgroup.

Acording to TID's, the route.cfg file should be placed in the 
wpgate\gwia directory, but......

I've found, that when this fail, simply move the file to sys:system, and 
restart the GWIA.

Please report success/errors if you try this

-- 
Best Regards

Tommy Mikkelsen

IT Quality A/S
Denmark

Thanks, Tommy.  I will give that a try.  I looked at the problem system 
today after sending the email and at least today, it looks like it did 
not get any Host Down errors from its own domain.  I will monitor the 
situation and try that if I see it happen again.

Thanks,

Dan

Hi,

"Daniel C. Hunt" wrote:
> 
> On a GW 6.5 SP1 system, I am still getting the '450 Host Down' Error at
> times when an outbound message is going to our own domain.  In the log
> file it looks like:
> 
> "DMN: MSG 5494 Send Failure: 450 Host down (MYDOMAIN.COM)"
> 
> Where MYDOMAIN.COM is our own Internet domain.  I have followed the
> suggestions of TID 10060050 and put an entry in the Route.cfg file for
> the GWIA like so:

Note that the route.cfg exclusively takes effect for *new* mails. Mails
that are already in your defer directory will still try to use the
official IP when they get retried, thus still giving the same error.

CU,
-- 
Massimo Rosen
Novell Support Connection Sysop
No emails please!
http://www.cfc-it.de

Thanks, Massimo.  In this case, we have had our domain in the Route.cfg 
for a while so I don't think that is it.

I guess I don't understand why the GroupWise MTA even sends a message 
to the GWIA when the address is a domain that is in the GroupWise 
system.  Shouldn't it route it directly to the correct post office in 
this case?  

Shouldn't I be able to send an email to 'user@mydomain.com' when 
'mydomain.com' has been defined as the Default domain name for the 
GroupWise system and have the MTA route it directly to the correct Post 
Office?

Dan

Hi,

"Daniel C. Hunt" wrote:
> 
> Thanks, Massimo.  In this case, we have had our domain in the Route.cfg
> for a while so I don't think that is it.

Actually, rereading your original message, your GWIA runs *on* your
Bordermanager? That would mean it's not natted at all to the outside,
thus the route.cfg approach would be unnecessary.
 
> I guess I don't understand why the GroupWise MTA even sends a message
> to the GWIA when the address is a domain that is in the GroupWise
> system.  Shouldn't it route it directly to the correct post office in
> this case?

Are you sure it comes from your MTA? Usually, mails like this are error
messages *generated* by GWIA itself, aka undeliverable notifications.
These will obviously go out the GWIA the very same way as every other
internet mail, aka your GWIA will lookup the MX record for it's own
domain, and try to connect to whatever it gets back. Again, as long as
your GWIA isn't natted anwhere, this should by all means work without
route.cfg in place.
 
> Shouldn't I be able to send an email to 'user@mydomain.com' when
> 'mydomain.com' has been defined as the Default domain name for the
> GroupWise system and have the MTA route it directly to the correct Post
> Office?

Yes, absolutely. Such an email send internally should *never* hit the
GWIA. If it does anyways, your MTA doesn't know it's own domain
properly, *or* doesn't find that user name. E.g, an email sent to
thisuserdoesntexisthere@yourdomain.com *will* go to GWIA in all cases,
as there's a possibility you have a different email system servicing
your own domain. So if your MTA believes the user you sent such an email
to doesn't exist in your system, it will send it to the GWIA for
processing.
 
CU,
-- 
Massimo Rosen
Novell Support Connection Sysop
No emails please!
http://www.cfc-it.de

Massimo Rosen wrote:
> Actually, rereading your original message, your GWIA runs *on* your
> Bordermanager? That would mean it's not natted at all to the outside,
> thus the route.cfg approach would be unnecessary.


Set NAT Dynamic mode to pass thru = on ?????

-- 
Best Regards

Tommy Mikkelsen

IT Quality A/S
Denmark

Tommy,

Tommy Mikkelsen wrote:
> 
> Massimo Rosen wrote:
> > Actually, rereading your original message, your GWIA runs *on* your
> > Bordermanager? That would mean it's not natted at all to the outside,
> > thus the route.cfg approach would be unnecessary.
> 
> Set NAT Dynamic mode to pass thru = on ?????

Doubt it. Without that setting (or the more current Nat Implicit
Filtering in inetcfg), he wouldn't be able to receive anything.

However, it *might* be filtering, depending on which source and target
address the gwia uses when trying to connect to itself.

CU,
-- 
Massimo Rosen
Novell Support Connection Sysop
No emails please!
http://www.cfc-it.de

Yes, Massimo, this is a NW 6.5 SP2 small business server so 
BorderManager 3.8 and GWIA are on the same box.  I have also seen it on 
a NetWare 6.0 SP5 server with BorderManager 3.7 and GWIA on the same 
box.  See the following.  I can't tell if it is coming from the GWIA or 
from the MTA.  I assumed it was from the MTA.

For instance, on the 6.0 server, I checked and I haven't seen any today 
but yesterday I saw approx 11 of the following entries in the log file. 
 They were not all the same deferred message:

10-21-04 06:03:17 0  Queuing deferred message: 
BORDER_MGR/VOL1:\DOMAIN\WPGATE\GWIA\send\s17660d4.070

10-21-04 06:03:27 8  DMN: MSG 8837 Sending file: 
BORDER_MGR/VOL1:\DOMAIN\WPGATE\GWIA\THIRD\send\p17660D4.070

10-21-04 06:03:27 8  DMN: MSG 8837 Send Failure: 450 Host down 
(JWHCORP.COM)

10-21-04 06:03:47 0  MSG 8838 Analyzing result file: 
BORDER_MGR/VOL1:\DOMAIN\WPGATE\GWIA\result\r17660d4.070

10-21-04 06:03:47 0  MSG 8838 Detected error on SMTP command

10-21-04 06:03:47 0  MSG 8838  Command:  JWHCORP.COM

10-21-04 06:03:47 0  MSG 8838  Response: 450 Host down (JWHCORP.COM)

10-21-04 06:03:47 0  MSG 8838 Deferring message: 
BORDER_MGR/VOL1:\DOMAIN\WPGATE\GWIA\defer\s17660d4.070

There is no file with that name in the 'defer' directory, nor is there 
a file by the name above in the 'result' directory so it must have 
resolved them but I don't know why it happened in the first place.

BTW NAT Implicit Filtering is enabled.

Any thoughts?  

Dan

Hi,

"Daniel C. Hunt" wrote:
> 
> Yes, Massimo, this is a NW 6.5 SP2 small business server so
> BorderManager 3.8 and GWIA are on the same box.  

Ok. That being clear now, you don't need a route.cfg, unless the primary
MX record for your own domain points elsewhere (i.e your ISP), and even
then it wouldn't be really needed, but it would be nice to have.

> 10-21-04 06:03:47 0  MSG 8838  Command:  JWHCORP.COM
> 
> 10-21-04 06:03:47 0  MSG 8838  Response: 450 Host down (JWHCORP.COM)

This looks a bit strange. In fact, it should say "450 Host down
(mail.jwhcorp.com)", as this is the primary MX for that domain, and it
should by all means try to connect to that *host*, and not to the domain
name. Is that their own domain, or a third party?
 
> There is no file with that name in the 'defer' directory, nor is there
> a file by the name above in the 'result' directory so it must have
> resolved them but I don't know why it happened in the first place.

Either that, or it has given up in the meantime(after 4 days).
 
> Any thoughts?

Yes, we're getting diluted. :-) We should try to concentrate on one
issue, on one server, and solve that.

So to get back to your own box, we apparently have two issues (or not).

1. 450 Host Down for your own domain. We first need to find out if this
is consistent. However, I suggest you remove your route.cfg anyways.

2. Internal Mails apparently going to your GWIA. Again, we need to
verify if this is consistent. However, would both be consistent errors,
you shouldn't be able to send any internal emails, which apparently
isn't the case, right?

CU,
-- 
Massimo Rosen
Novell Support Connection Sysop
No emails please!
http://www.cfc-it.de

Perhaps the public interface is the first one loaded ??????

-- 
Best Regards

Tommy Mikkelsen

IT Quality A/S
Denmark

Tommy,

Tommy Mikkelsen wrote:
> 
> Perhaps the public interface is the first one loaded ??????

Actually, that would be good. Would  it *not* be the first loaded, ho
could face a NAT or filter issue, if GWIA tries to connect to it's
public IP with it's private IP as source. If the public IP is the first
bound one, it should work in all cases, *if* GWIA is otherwise
functional, aka can send and recieve mail from and to others.

CU,
-- 
Massimo Rosen
Novell Support Connection Sysop
No emails please!
http://www.cfc-it.de

Yes and no.

If Public is the first one loaded.

NCP can take forever, unless specified in Monitor/Server Params.

But regarding the GWIA, are you sure, that it's not a conflict between 
NAT/Filters and the actually interface used by the GWIA ??

-- 
Best Regards

Tommy Mikkelsen

IT Quality A/S
Denmark

Hi,

Tommy Mikkelsen wrote:
> 
> Yes and no.
> 
> If Public is the first one loaded.
> 
> NCP can take forever, unless specified in Monitor/Server Params.

That would be a client problem, but has nothing to do with the issue
here, as NCP is not involved at all. And I didn't say having the public
IP as first bound is good for all situations, but for his GWIA it would
be.
 
> But regarding the GWIA, are you sure, that it's not a conflict between
> NAT/Filters and the actually interface used by the GWIA ??

As I said, if his first bound IP is *not* the public one, it *could* be
a NAT / Filter issue. Less so NAT, but quite possibly filters.

CU,
-- 
Massimo Rosen
Novell Support Connection Sysop
No emails please!
http://www.cfc-it.de

Sorry for the long delay in replying Massimo.  I have been swamped.  I 
looked at your suggestions.  In my system I have a Server S1 which has 
two NIC's (one private, one public), is the BorderManager server and 
runs a secondary domain MTA and the GWIA.  I have a second server, S3, 
which has the primary domain on it and the single post office in the 
system.

With this being the setup, why would I ever have mail coming from the 
primary domain to the secondary domain and then to the GWIA addressed 
to someone@jwhcorp.com?

I removed the jwhcorp.com entry from the route.cfg file in the 
\wpgate\gwia dir on the S1 server a couple days ago.

I created an MX record today for my intranet to point to the GWIA on 
S1.

I checked and I have not been getting the 'host down' message for my 
own domain in the last week (so before I made any of the above changes) 
but I will continue to monitor it.  I still don't understand why I 
would ever be seeing mail at this GWIA that was being sent out to 
jwhcorp.com.

Your second question:  
>>2. Internal Mails apparently going to your GWIA. Again, we need to
>>verify if this is consistent. However, would both be consistent 
>>errors, you shouldn't be able to send any internal emails, which 
>>apparently isn't the case, right?

It is correct that we can send internal emails just fine in GroupWise 
to other GroupWise recipients in our system (of course they are all in 
the same post office on S3, under the primary domain.  As well as we 
can send email in and out of our system via the GWIA to the Internet.  
I am trying to get rid of the bogus 'host down' errors.

Dynamic NAT is configured properly.  In Netinfo.cfg, it appears that 
the private interface is the first one loaded.

Thanks for your help.

Dan

Hi,

"Daniel C. Hunt" wrote:
> 
> With this being the setup, why would I ever have mail coming from the
> primary domain to the secondary domain and then to the GWIA addressed
> to someone@jwhcorp.com?

Same answer as before: There are two possible reasons I know of:

1. Your primary domain doesn't know that jwhcorp.com is your own domain.
2. Your primary domain doesn't believe the user you sent the mail to
exists in your system, so it goes out to GWIA as last resort.
 
 
> I created an MX record today for my intranet to point to the GWIA on
> S1.

But does your GWIA server actually query that internal DNS?
 
> I checked and I have not been getting the 'host down' message for my
> own domain in the last week (so before I made any of the above changes)
> but I will continue to monitor it.  I still don't understand why I
> would ever be seeing mail at this GWIA that was being sent out to
> jwhcorp.com.

That's not unusual at all for status messages generated by your GWIA,
e.g undeliverable notifications. It *is* unusual for normal user to user
mails coming from your own GW system.
 
CU,
-- 
Massimo Rosen
Novell Support Connection Sysop
No emails please!
http://www.cfc-it.de

Thanks, Massimo for the response.

1. Your primary domain doesn't know that jwhcorp.com is your own 
domain.

I checked and the primary domain, on the Properties | Internet 
Addressing tab shows 'jwhcorp.com' as the domain name.


2. Your primary domain doesn't believe the user you sent the mail to
exists in your system, so it goes out to GWIA as last resort.

 I'll watch for this.
 
But does your GWIA server actually query that internal DNS?

Well, GWIA queries my internal DNS server to resolve IP addresses for 
outside SMTP hosts.  My internal DNS server goes to a forwarder DNS 
sever on the Internet for the names it can't resolve.

I checked again, and since I started investigating it, the problem 
seems to have gone away!

Thanks,

Dan

Hi,

"Daniel C. Hunt" wrote:
> I checked again, and since I started investigating it, the problem
> seems to have gone away!

Whatever it takes. :-)

CU,
-- 
Massimo Rosen
Novell Support Connection Sysop
No emails please!
http://www.cfc-it.de