Suddenly lots of "host lookup did not complete" errors

[Dan Muey]

Hello list,

All of a sudden today over half, maybe closer to all( I have to experiemtn more) of the emails sent out of
our Exim server send warnings then fail witht the error:

'host lookup did not complete'

Most if not all of the messages in the queue are there because of that and if I do
A host lookup on any of those domains I get all the records back I should.
I don't control any of the erroring domains and both the local name server and our other two nameservers
Return the mx records and the ip address of the mx record like it should.

Any one have any idea what would suddenly start giving that error all the time?

What is Exim trying to do when it gives that error, reverse dns comparison or ???

Thanks

Dan

--

## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##

[Dan Muey]

Are you running Debian? I chose today to upgrade from Exim3
to Exim4 using

Nope , RedHat 8.0 and I didn't uupgrade or install anything.
It just started happening for a new domain.
And the domain itself has no errors it's the domain's they are sending to.

the latest unstable debian package, but I'm getting the same
error you're
reporting.
Of course, DNS works fine otherwise, and I can lookup MX for
hotmail.com on
the commandline.

Same here.

Haven't found any solution to this problem. Must be a bug in
exim or some
kinde of weird configuration error. Hoping for a fixed package soon.

Me to. Or at least a pointer to what Exim is trying to
do and is not able to do or expecting and not receiving, when that error occurs.

Or what configureation settings might cause that.
Most of the 100+/- messages stuck in queue and eventually returned
to the sender is originally sent by the same domain or trying to send an
error back to the remote sender about and non existant address for that domain.

Any Exim gurus out there? To quote Pink Floyd -Is there anybody in there?-

Thanks!

Dan

Regards,
Eddy

--

## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##

[Dan Muey]

When Exim can't deliver a message because "host lookup did not complete" is
it trying to lookup the host of the sender or reciever?

How is it do that lookup?

What info is it expecting/needing to get?

I ask because we have one domain that can't send anythign
out because they all get stuck in queueu with said error message.

I had them change their smtp to our qmail server and all is creamy and dreamy now.
But I really would like Exim to be their SMTP server.

Any thoughts?

TIA

Dan

--

## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##

[Fernando Sanchez]

Post by Dan Muey
When Exim can't deliver a message because "host lookup did not complete" is
it trying to lookup the host of the sender or reciever?
How is it do that lookup?
What info is it expecting/needing to get?
I ask because we have one domain that can't send anythign
out because they all get stuck in queueu with said error message.

It is looking for the ip of the remote server to where it has to send
the mail

Post by Dan Muey
I had them change their smtp to our qmail server and all is creamy and dreamy now.
But I really would like Exim to be their SMTP server.
Any thoughts?

make sure your dns server is working and your mail gateway can ping any
of the listed servers in the errors, or try to use dig to test for the
MX records

Post by Dan Muey
TIA
Dan
--
## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##

--


Fernando Sanchez
Dpto. Sistemas USFQ



--

## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##

[Philip Hazel]

Post by Dan Muey
When Exim can't deliver a message because "host lookup did not complete" is
it trying to lookup the host of the sender or reciever?

Could be either. This message comes from the dnslookup router, which is
used both to send mail, and to verify senders. It depends *when* the
message is issued - if it is during delivery, it's looking up a
receiver; if it's during reception, it could be verifying a receiver or
a sender, depending on your configuration.

Post by Dan Muey
How is it do that lookup?

DNS.

Post by Dan Muey
What info is it expecting/needing to get?

An IP address.

Post by Dan Muey
I ask because we have one domain that can't send anythign
out because they all get stuck in queueu with said error message.

Not sure what you mean by "one domain that can't send". Are you saying
that other messages *on the same host* but with different sender
domains, can look up that host? If so, I am gobsmacked.

Post by Dan Muey
I had them change their smtp to our qmail server and all is creamy and dreamy now.

Oh, that suggests that you mean one client host can't send rather than
one domain. That makes more sense.

Post by Dan Muey
But I really would like Exim to be their SMTP server.

Run a test on your Exim server using the -bh option, and simulate a mail
from the failing client. See what Exim tells you.

--
Philip Hazel University of Cambridge Computing Service,
***@cus.cam.ac.uk Cambridge, England. Phone: [phone].
Get the Exim 4 book: http://www.uit.co.uk/exim-book


--

## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##

[Dan Muey]

Post by Dan Muey
When Exim can't deliver a message because "host lookup did not
complete" is it trying to lookup the host of the sender or reciever?

The reciever.

Post by Dan Muey
What info is it expecting/needing to get?

The internet adress of the server to deliver to.

Post by Dan Muey
Any thoughts?

Not really.

Thanks for the info that will really help track it down.

Do you get "DNS lookup of hotmail.com (MX) gave TRY_AGAIN"
for when running

Nope I get:

[/home/dmuey]# /usr/sbin/exim -d -bt ***@hotmail.com
Exim version 3.36 debug level 1 uid=0 gid=0
Berkeley DB: Sleepycat Software: DB 2.4.14: (6/2/98)
***@hotmail.com
deliver to ***@hotmail.com
router = lookuphost, transport = remote_smtp
host mx4.hotmail.com [65.54.254.151] MX=5
host mx4.hotmail.com [65.54.253.230] MX=5
host mx2.hotmail.com [65.54.252.230] MX=5
host mx2.hotmail.com [65.54.254.145] MX=5
host mx2.hotmail.com [65.54.166.230] MX=5
host mx1.hotmail.com [65.54.254.129] MX=5
host mx1.hotmail.com [65.54.252.99] MX=5
host mx1.hotmail.com [65.54.166.99] MX=5
host mx3.hotmail.com [65.54.253.99] MX=5
host mx3.hotmail.com [65.54.254.140] MX=5
[/home/dmuey]#

That command will be good to know thanks. Here's from one of the offending addresses:

[/home/dmuey]# /usr/sbin/exim -d -bt ***@pilotonline.com
Exim version 3.36 debug level 1 uid=0 gid=0
Berkeley DB: Sleepycat Software: DB 2.4.14: (6/2/98)
***@pilotonline.com cannot be resolved at this time:
host lookup did not complete
[/home/dmuey]# host -t mx pilotonline.com
pilotonline.com mail is handled by 100 mailhost2.infi.net.
pilotonline.com mail is handled by 60 gateway1.pilotonline.com.
pilotonline.com mail is handled by 70 gateway2.pilotonline.com.
[/home/dmuey]#

So now the question is why can I host, dig, or even nslookup
the domain and all is well but exim can't find it?

See isn't that weird?
It seems to only be happening to one domain out of about fifteen.

Thanks!

Dan

That's the weird thing, the error message is "TRY_AGAIN", not
"NONEXIST".

--

## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##

[Dan Muey]

Post by Fernando Sanchez

Post by Dan Muey
When Exim can't deliver a message because "host lookup did not
complete" is it trying to lookup the host of the sender or reciever?
How is it do that lookup?
What info is it expecting/needing to get?
I ask because we have one domain that can't send anythign
out because they all get stuck in queueu with said error message.

It is looking for the ip of the remote server to where it has to send
the mail

That's what I figured but since it was screwy I wans't sure

Post by Fernando Sanchez

Post by Dan Muey
I had them change their smtp to our qmail server and all is

creamy and

Post by Dan Muey
dreamy now. But I really would like Exim to be their SMTP server.
Any thoughts?

make sure your dns server is working and your mail gateway
can ping any
of the listed servers in the errors, or try to use dig to
test for the
MX records

[/home/dmuey]# /usr/sbin/exim -d -bt ***@pilotonline.com
Exim version 3.36 debug level 1 uid=0 gid=0
Berkeley DB: Sleepycat Software: DB 2.4.14: (6/2/98)
***@pilotonline.com cannot be resolved at this time:
host lookup did not complete
[/home/dmuey]# host -t mx pilotonline.com
pilotonline.com mail is handled by 100 mailhost2.infi.net.
pilotonline.com mail is handled by 60 gateway1.pilotonline.com.
pilotonline.com mail is handled by 70 gateway2.pilotonline.com.
[/home/dmuey]#

If I ping, say gateway2.pilotonline.com then they fail not only form the exim server but also
From an internet based ping service.

That's got to be it but why would that happen all of a sudden to all domains that these folks are trying to send to.

Strange.

Thanks

Dan

Post by Fernando Sanchez

Post by Dan Muey
TIA
Dan
--
## List details at http://www.exim.org/mailman/listinfo/exim-users
Exim details at http://www.exim.org/ ##

--
Fernando Sanchez
Dpto. Sistemas USFQ

--

## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##

[Dan Muey]

Post by Dan Muey
Exim version 3.36 debug level 1 uid=0 gid=0

Okay, this just increase the weirdness, because it thought
you were running
exim4 too. So we're on two different codebases? I run 4.20
I'd start looking into a problem with some resolver library,
but everything
_but_ exim seems to be working fine.
Take care when cc'ing this to the mailinglist, I don't know if they
spam-obfuscate adresses automatically!

Post by Dan Muey
It seems to only be happening to one domain out of about fifteen.

I haven't found one domain that worked, until I tried your
tiffany up there,
and lo-and-behold, it works!
What the f...

I know. It just gets weirder and weirder, I can't ping the mx
records my host returns for her.
Not even form an internet based utility that isn't even remotely
related to our stuff.

Oh well we'll see what happens...

_________________________________________________________________
Add photos to your e-mail with MSN 8. Get 2 months FREE*.
http://join.msn.com/?page=features/featuredemail

--

## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##

[Dan Muey]

One question, is your exim built for IPv6 and if so, do you have IPv6
support in the kernel?

Not sure, it was all handled by cpanel and Red Hat 8.0 was a pretty vanilla install.

Not sure about the IPv4 or 6 I'll look into that as well.
Thanks

The debian exim package I'm using has IPv6-support, but I
don't have that
built into the kernel. I thought this could be the problem,
but I've changed
exim to only bind on the available interface (IPv4 ethernet)
and also I've
added the option "dns_ipv4_lookup = *" to my exim.conf which
should force it
to use IPv4 lookup for all domains.

Hmm, may have to try that one.

I'm clutching at straws here. :-\
_________________________________________________________________
Protect your PC - get McAfee.com VirusScan Online
http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963

--

## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##

[Dan Muey]

Post by Philip Hazel

Post by Dan Muey
When Exim can't deliver a message because "host lookup did not
complete" is it trying to lookup the host of the sender or reciever?

Could be either. This message comes from the dnslookup
router, which is used both to send mail, and to verify
senders. It depends *when* the message is issued - if it is
during delivery, it's looking up a receiver; if it's during
reception, it could be verifying a receiver or a sender,
depending on your configuration.

Post by Dan Muey
How is it do that lookup?

DNS.

Right, I guess I meant does it issue a 'host' or 'dig' command or ???

Post by Philip Hazel

Post by Dan Muey
What info is it expecting/needing to get?

An IP address.

Post by Dan Muey
I ask because we have one domain that can't send anythign
out because they all get stuck in queueu with said error message.

Not sure what you mean by "one domain that can't send". Are
you saying that other messages *on the same host* but with
different sender domains, can look up that host? If so, I am
gobsmacked.

Post by Dan Muey
I had them change their smtp to our qmail server and all is

creamy and

Post by Dan Muey
dreamy now.

Oh, that suggests that you mean one client host can't send
rather than one domain. That makes more sense.

We have several domains setup in Exim.
All other domains are having no trouble sending (***@domain1.com, ***@domain1.com, etcc)
One domain, no matter what account it is gets a lot of "host lookup did not complete" errors when sending mail.

Here's an example of one address they can't seem to send to:

# /usr/sbin/exim -d9 -bt ***@pilotonline.com
Exim version 3.36 debug level 9 uid=0 gid=0
Berkeley DB: Sleepycat Software: DB 2.4.14: (6/2/98)
Caller is an admin user
Caller is a trusted user
originator: uid=0 gid=0 login=root name=root
sender address = ***@marvin.infiniplex.net
Address testing: uid=0 gid=0 euid=0 egid=0
Testing ***@pilotonline.com
search_open: lsearch "/etc/localdomains"
search_find: file="/etc/localdomains"
key="pilotonline.com" partial=-1
LRU list:
7/etc/localdomains
End
internal_search_find: file="/etc/localdomains"
type=lsearch key="pilotonline.com"
file lookup required for pilotonline.com
in /etc/localdomains
lookup failed
pilotonline.com in local_domains? no (end of list)
address ***@pilotonline.com
local_part=tiffany.nieves domain=pilotonline.com
domain is not local
routing ***@pilotonline.com, domain pilotonline.com
Starting Perl interpreter
lookuphost router called for ***@pilotonline.com
dns lookup: route_domain = pilotonline.com
DNS lookup of pilotonline.com (MX) gave TRY_AGAIN
pilotonline.com in dns_again_means_nonexist? no (end of list)
returning DNS_AGAIN
lookuphost router deferred pilotonline.com
message: host lookup did not complete
***@pilotonline.com cannot be resolved at this time:
host lookup did not complete
search_tidyup called
#

Post by Philip Hazel

Post by Dan Muey
But I really would like Exim to be their SMTP server.

Run a test on your Exim server using the -bh option, and
simulate a mail from the failing client. See what Exim tells you.

I'll take look at that also.

Thanks for yout input!

Dan

Post by Philip Hazel
--
Philip Hazel University of Cambridge Computing Service,
Get the Exim 4 book: http://www.uit.co.uk/exim-book

--

## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##

[Philip Hazel]

Post by Dan Muey

DNS.

Right, I guess I meant does it issue a 'host' or 'dig' command or ???

Neither, it calls the DNS resolver library.

Post by Dan Muey
We have several domains setup in Exim.
All other domains are having no trouble sending

I'm sorry, but I genuinely don't understand what you mean by "a domain
sending". Do you mean that all messages (from whatever host) that have a
particular domain in the sender address are having trouble, or do you
mean that all messages from a specific host (that happens to use one
particular domain for all its sending addresses) are having trouble?

Post by Dan Muey
One domain, no matter what account it is gets a lot of "host lookup did not complete" errors when sending mail.

By "account" do you mean "login account on a host" or just "customer
account".
Oh good, debug output. That I can probably understand ...

Post by Dan Muey
Exim version 3.36 debug level 9 uid=0 gid=0

Ah. Exim 3. Sorry, I can't remember too much about Exim 3 these days. It
is now very obsolete. However, "pilotonline.com" there is a recipient
domain, not a sending domain.

<snip>

Post by Dan Muey
DNS lookup of pilotonline.com (MX) gave TRY_AGAIN

But that error is clear. The DNS lookup failed in some way. The problem
is with the DNS resolver or the DNS nameserver. What output do you get
from

dig pilotonline.com mx

? On my host, it works fine.

--
Philip Hazel University of Cambridge Computing Service,
***@cus.cam.ac.uk Cambridge, England. Phone: [phone].
Get the Exim 4 book: http://www.uit.co.uk/exim-book


--

## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##

[eloj .]

Post by Philip Hazel

Post by Dan Muey
DNS lookup of pilotonline.com (MX) gave TRY_AGAIN

But that error is clear. The DNS lookup failed in some way. The problem
is with the DNS resolver or the DNS nameserver. What output do you get
from
dig pilotonline.com mx
? On my host, it works fine.

See, that's the problem exactly. See here:

nynaeve:/home/eddy# exim4 -d -bt ***@hotmail.com
Exim version 4.20 uid=0 gid=0 pid=10985 D=fbb95cfd
Berkeley DB: Sleepycat Software: Berkeley DB 3.2.9: (August 12, 2002)
Support for: IPv6
Authenticators: cram_md5 plaintext
Routers: accept dnslookup ipliteral manualroute queryprogram redirect
Transports: appendfile/maildir/mailstore/mbx autoreply lmtp pipe smtp
changed uid/gid: forcing real = effective
uid=0 gid=0 pid=10985
auxiliary group list: <none>
nynaeve in dns_ipv4_lookup? yes (matched "*")
configuration file is /var/lib/exim4/config.autogenerated
log selector = 020d99d8
trusted user
admin user
user name "root" extracted from gecos field "root"
originator: uid=0 gid=0 login=root name=root
sender address = ***@MYDOMAIN.net
Address testing: uid=0 gid=8 euid=0 egid=8
Testing ***@hotmail.com
Considering ***@hotmail.com
routing ***@hotmail.com
--------> dnslookup router <--------
local_part=srm_dfr domain=hotmail.com
checking domains
hotmail.com in "@:MYDOMAIN.net:localhost:nynaeve"? no (end of list)
hotmail.com in "! +local_domains"? yes (end of list)
calling dnslookup router
dnslookup router called for ***@hotmail.com
domain = hotmail.com
DNS lookup of hotmail.com (MX) gave TRY_AGAIN
hotmail.com in dns_again_means_nonexist? no (option unset)
returning DNS_AGAIN
dnslookup router: defer for ***@hotmail.com
message: host lookup did not complete
***@hotmail.com cannot be resolved at this time:
host lookup did not complete
search_tidyup called

Post by Philip Hazel

Post by Dan Muey

Exim pid=10985 terminating with rc=1 >>>>>>>>>>>>>>>>

nynaeve:/home/eddy# dig hotmail.com mx

; <<>> DiG 9.2.2 <<>> hotmail.com mx
;; global options: printcmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOTIMP, id: 29356
;; flags: qr ra; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 0

;; QUESTION SECTION:
;hotmail.com. IN MX

;; Query time: 6 msec
;; SERVER: 192.168.0.254#53(192.168.0.254)
;; WHEN: Fri Jul 11 11:47:30 2003
;; MSG SIZE rcvd: 29

nynaeve:/home/eddy# host -d -t MX hotmail.com
;; res_send()
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 11038
;; flags: rd; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 0
;; hotmail.com, type = MX, class = IN
;; Querying server (# 1) udp address = 192.168.0.254
;; got answer, 29 bytes:
;; ->>HEADER<<- opcode: QUERY, status: NOTIMP, id: 11038
;; flags: qr ra; QUERY: 1, ANSWER: 0, AUTHORITY: 0, ADDITIONAL: 0
;; hotmail.com, type = MX, class = IN
;; Query failed, 0 answers, status: not implemented
hotmail.com MX record not found, no recovery

Understand our confusion? With this I'm beginning to again eye the
resolver libs:

nynaeve:/lib# locate libresolv
/lib/libresolv-2.3.1.so
/lib/libresolv.so.2
/usr/lib/libresolv.a
/usr/lib/libresolv.so
nynaeve:/lib# ls -l /lib/libresolv*
-rw-r--r-- 1 root root 56652 Apr 19 20:57
/lib/libresolv-2.3.1.so
lrwxrwxrwx 1 root root 18 Jun 9 15:34 /lib/libresolv.so.2
-> libresolv-2.3.1.so
nynaeve:/lib# ls -l /usr/lib/libresolv*
-rw-r--r-- 1 root root 73990 Apr 19 20:58 /usr/lib/libresolv.a
lrwxrwxrwx 1 root root 19 Jun 9 15:34
/usr/lib/libresolv.so -> /lib/libresolv.so.2

What is the easiest way to find out which one host and exim is using? I
forget...

_________________________________________________________________
Help STOP SPAM with the new MSN 8 and get 2 months FREE*
http://join.msn.com/?page=features/junkmail


--

## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##

[eloj .]

Post by Philip Hazel
? On my host, it works fine.

You'll have to excuse my precious message (which is currently awaiting
approval by the moderator), but I was under the impression that "dig" worked
where exim and host did not.

Anyway, I went ahead and replaced the DNS I had been using, which is the one
supplied in the "SpeedTouch Alcatel 410" cable modem. I went in and checked
the configuration, and found it was forwarding to two different IPs supplied
by my ISP. So I reactivated bind on my linux host instead, entered said IPs
as forwarders and...

nynaeve:/etc/bind# dig hotmail.com mx

; <<>> DiG 9.2.2 <<>> hotmail.com mx
;; global options: printcmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 55390
;; flags: qr rd ra; QUERY: 1, ANSWER: 4, AUTHORITY: 4, ADDITIONAL: 14

;; QUESTION SECTION:
;hotmail.com. IN MX

;; ANSWER SECTION:
hotmail.com. 1127 IN MX 5 mx3.hotmail.com.
hotmail.com. 1127 IN MX 5 mx4.hotmail.com.
hotmail.com. 1127 IN MX 5 mx1.hotmail.com.
hotmail.com. 1127 IN MX 5 mx2.hotmail.com.

;; AUTHORITY SECTION:
hotmail.com. 3537 IN NS ns2.hotmail.com.
hotmail.com. 3537 IN NS ns3.hotmail.com.
hotmail.com. 3537 IN NS ns4.hotmail.com.
hotmail.com. 3537 IN NS ns1.hotmail.com.

;; ADDITIONAL SECTION:
mx3.hotmail.com. 184 IN A 65.54.253.99
mx3.hotmail.com. 184 IN A 65.54.254.140
[...]

;; Query time: 1864 msec
;; SERVER: 127.0.0.1#53(127.0.0.1)
;; WHEN: Fri Jul 11 12:10:53 2003
;; MSG SIZE rcvd: 416

Seems that my immediate problem is now solved.

_________________________________________________________________
Help STOP SPAM with the new MSN 8 and get 2 months FREE*
http://join.msn.com/?page=features/junkmail


--

## List details at http://www.exim.org/mailman/listinfo/exim-users Exim details at http://www.exim.org/ ##