Re: OT sendmail delay

[David G. Miller (aka DaveAtFraud)]

Bryan Hepworth wrote:
> route
>
> Kernel IP routing table
> Destination Gateway Genmask Flags Metric Ref Use Iface
> 192.168.2.0 * 255.255.255.0 U 0 0 0 eth1
> 169.254.0.0 * 255.255.0.0 U 0 0 0 eth1
> 94.0.0.0 93.0.0.100 255.0.0.0 UG 0 0 0 eth0
> 92.0.0.0 93.0.0.100 255.0.0.0 UG 0 0 0 eth0
> 93.0.0.0 * 255.0.0.0 U 0 0 0 eth0
> default 192.168.2.1 0.0.0.0 UG 0 0 0 eth1

The 169.254.0.0 entry is for compatibility with a Microsoft peer-to-peer 
networking.  It shouldn't hurt anything.  The 94.0.0.0, 92.0.0.0 and 
93.0.0.0 entries are probably not what you want unless the route to 
these subnets should still be out eth0 (the 93.1.1.208 NIC).  My take on 
your original posting was the eth0 was no longer in use.
> dig internal
>
> ; <<>> DiG 9.2.4 <<>> any internal.coxagri.com
> ;; global options: printcmd
> ;; Got answer:
> ;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 8694
> ;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 0

NXDOMAIN is dig's way of saying it can't find an IP address for 
internal.coxagri.com.  See if you can get this boxes name and IP address 
to resolve through dig.  Sendmail likes to have it's hostname resolvable 
through DNS.

Cheers,
Dave

OT sendmail delay

[Bryan Hepworth]

> Bryan Hepworth wrote:
> > route
> >
> > Kernel IP routing table
> > Destination Gateway Genmask Flags Metric Ref Use Iface
> > 192.168.2.0 * 255.255.255.0 U 0 0 0 eth1
> > 169.254.0.0 * 255.255.0.0 U 0 0 0 eth1
> > 94.0.0.0 93.0.0.100 255.0.0.0 UG 0 0 0 eth0
> > 92.0.0.0 93.0.0.100 255.0.0.0 UG 0 0 0 eth0
> > 93.0.0.0 * 255.0.0.0 U 0 0 0 eth0
> > default 192.168.2.1 0.0.0.0 UG 0 0 0 eth1
>
> The 169.254.0.0 entry is for compatibility with a Microsoft peer-to-peer 
> networking.  It shouldn't hurt anything.  The 94.0.0.0, 92.0.0.0 and 
> 93.0.0.0 entries are probably not what you want unless the route to these 
> subnets should still be out eth0 (the 93.1.1.208 NIC).  My take on your 
> original posting was the eth0 was no longer in use.
> > dig internal
> >
> > ; <<>> DiG 9.2.4 <<>> any internal.coxagri.com
> > ;; global options: printcmd
> > ;; Got answer:
> > ;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 8694
> > ;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 0
>
> NXDOMAIN is dig's way of saying it can't find an IP address for 
> internal.coxagri.com.  See if you can get this boxes name and IP address 
> to resolve through dig.  Sendmail likes to have it's hostname resolvable 
> through DNS.

Dave

Thanks for the insight - I'm working on it over the weekend when everyone 
has gone home.

The public ip from this address is an ADSL link name rather than 
internal.coxagri.com

Having had a quick look at other people in the same scenario I have some 
reading to do. If you have any suggestions I'd be happy to hear them.

Thanks

Bryan

OT sendmail delay

David G. Miller wrote:
> Bryan Hepworth wrote:
> > route
> >
> > Kernel IP routing table
> > Destination Gateway Genmask Flags Metric Ref Use Iface
> > 192.168.2.0 * 255.255.255.0 U 0 0 0 eth1
> > 169.254.0.0 * 255.255.0.0 U 0 0 0 eth1
> > 94.0.0.0 93.0.0.100 255.0.0.0 UG 0 0 0 eth0
> > 92.0.0.0 93.0.0.100 255.0.0.0 UG 0 0 0 eth0
> > 93.0.0.0 * 255.0.0.0 U 0 0 0 eth0
> > default 192.168.2.1 0.0.0.0 UG 0 0 0 eth1
>
> The 169.254.0.0 entry is for compatibility with a Microsoft 
> peer-to-peer networking.  It shouldn't hurt anything.  The 94.0.0.0, 
> 92.0.0.0 and 93.0.0.0 entries are probably not what you want unless 
> the route to these subnets should still be out eth0 (the 93.1.1.208 
> NIC).  My take on your original posting was the eth0 was no longer in 
> use.
> > dig internal
> >
> > ; <<>> DiG 9.2.4 <<>> any internal.coxagri.com
> > ;; global options: printcmd
> > ;; Got answer:
> > ;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 8694
> > ;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 0
>
> NXDOMAIN is dig's way of saying it can't find an IP address for 
> internal.coxagri.com.  See if you can get this boxes name and IP 
> address to resolve through dig.  Sendmail likes to have it's hostname 
> resolvable through DNS.
> Cheers,
> Dave

Dave

Thanks for your reply to this. When I was doing further checks I found 
that it was also failing reverse dns look ups. So I bit the bullet and 
started to learn about dns. Would you have any advice to offer as to 
best practice for this?

I was thinking that we need an internal dns server to keep sendmail 
happy with all the internal people that use it to send out email. 
Sendmail  isn't currently taking mail in yet directly. That's taken from 
the box that's hosted at the ISP and brought in by fetchmail. Long term 
this was going to change and the MX record externally (at the ISP) was 
going to point to our adsl router.

Any advice would be much appreciated.

Bryan

OT sendmail delay

[Oluwagbenga Shobowale]

On 10/6/06, [e-mail]  wrote:
> David G. Miller wrote:
> > Bryan Hepworth wrote:
> > > route
> > >
> > > Kernel IP routing table
> > > Destination Gateway Genmask Flags Metric Ref Use Iface
> > > 192.168.2.0 * 255.255.255.0 U 0 0 0 eth1
> > > 169.254.0.0 * 255.255.0.0 U 0 0 0 eth1
> > > 94.0.0.0 93.0.0.100 255.0.0.0 UG 0 0 0 eth0
> > > 92.0.0.0 93.0.0.100 255.0.0.0 UG 0 0 0 eth0
> > > 93.0.0.0 * 255.0.0.0 U 0 0 0 eth0
> > > default 192.168.2.1 0.0.0.0 UG 0 0 0 eth1
> >
> > The 169.254.0.0 entry is for compatibility with a Microsoft
> > peer-to-peer networking.  It shouldn't hurt anything.  The 94.0.0.0,
> > 92.0.0.0 and 93.0.0.0 entries are probably not what you want unless
> > the route to these subnets should still be out eth0 (the 93.1.1.208
> > NIC).  My take on your original posting was the eth0 was no longer in
> > use.
> > > dig internal
> > >
> > > ; <<>> DiG 9.2.4 <<>> any internal.coxagri.com
> > > ;; global options: printcmd
> > > ;; Got answer:
> > > ;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 8694
> > > ;; flags: qr rd ra; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 0
> >
> > NXDOMAIN is dig's way of saying it can't find an IP address for
> > internal.coxagri.com.  See if you can get this boxes name and IP
> > address to resolve through dig.  Sendmail likes to have it's hostname
> > resolvable through DNS.
> > Cheers,
> > Dave
>
> Dave
>
> Thanks for your reply to this. When I was doing further checks I found
> that it was also failing reverse dns look ups. So I bit the bullet and
> started to learn about dns. Would you have any advice to offer as to
> best practice for this?
>
> I was thinking that we need an internal dns server to keep sendmail
> happy with all the internal people that use it to send out email.

I think you should just configure your DNS for internal use or at
least on the mail server...that way it would resolve mails for your
local domain and send mails which are meant for other domains out of
your network depending on your ISP
> Sendmail  isn't currently taking mail in yet directly. That's taken from
> the box that's hosted at the ISP and brought in by fetchmail. Long term
> this was going to change and the MX record externally (at the ISP) was

This is also possible and would not be much of an issue you would need
to tell your ISP or you could do it on your own. Redirect all mail
traffice MX to the IP address and you really don't have to do anything
on your mail server or on the DNS...

This approach is better is better than fetchmail...
> going to point to our adsl router.
>
> Any advice would be much appreciated.
>
> Bryan
>
> --
> fedora-list mailing list
> [e-mail]
> To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list

OT sendmail delay

[Tim]

On Fri, 2006-10-06 at 10:57 +0100, [e-mail] wrote:
> Thanks for your reply to this. When I was doing further checks I found 
> that it was also failing reverse dns look ups. So I bit the bullet and 
> started to learn about dns. Would you have any advice to offer as to 
> best practice for this?

Get it working internally, first, and be certain you're familiar with it
(your server, and DNS records in general) before you move beyond
internal DNS serving.
> I was thinking that we need an internal dns server to keep sendmail 
> happy with all the internal people that use it to send out email. 
> Sendmail  isn't currently taking mail in yet directly. That's taken from 
> the box that's hosted at the ISP and brought in by fetchmail. Long term 
> this was going to change and the MX record externally (at the ISP) was 
> going to point to our adsl router.

First advice:  Before setting yourself up with a SMTP server accepting
input from the public, learn about spam control.  Once you start
handling your own mail, you've also got to deal with all the spam that
someone else would have been managing for you.  You have to learn how to
kill it properly, not get exploited, and not get blacklisted.

For internal networking, it probably is easier to have a local DNS
server that takes care of address resolution (easier than maintaining
hosts files, etc.).  But be careful how you organise your internal mail
if you want users to be able to post to the outside world using the same
e-mail addresses.  You won't be able to post from a domain name that's
not recognised outside your LAN.  There's nothing stopping you from
having different responses to domain names inside and outside of your
LAN (i.e. using a public domain name, inside and out, but inside your
machines all have internal LAN IP addresses, for internal work, outside
your domain has a real internet IP address for mail checks, etc.).

OT sendmail delay

[Bryan Hepworth]

Subject: Re: OT sendmail delay

Tim

Thanks for the reply...
> On Fri, 2006-10-06 at 10:57 +0100, [e-mail] wrote:
> > Thanks for your reply to this. When I was doing further checks I found
> > that it was also failing reverse dns look ups. So I bit the bullet and
> > started to learn about dns. Would you have any advice to offer as to
> > best practice for this?
>
> Get it working internally, first, and be certain you're familiar with it
> (your server, and DNS records in general) before you move beyond
> internal DNS serving.
> > I was thinking that we need an internal dns server to keep sendmail
> > happy with all the internal people that use it to send out email.
> > Sendmail  isn't currently taking mail in yet directly. That's taken from
> > the box that's hosted at the ISP and brought in by fetchmail. Long term
> > this was going to change and the MX record externally (at the ISP) was
> > going to point to our adsl router.
>
> First advice:  Before setting yourself up with a SMTP server accepting
> input from the public, learn about spam control.  Once you start
> handling your own mail, you've also got to deal with all the spam that
> someone else would have been managing for you.  You have to learn how to
> kill it properly, not get exploited, and not get blacklisted.

The mailserver currently runs spamassassin, and testing with smtp auth and 
starttls also works. This was a requirement for getting port 25 opened by 
the isp. I've read about grey listing. Is there anything else you would 
recommend?
> For internal networking, it probably is easier to have a local DNS
> server that takes care of address resolution (easier than maintaining
> hosts files, etc.).  But be careful how you organise your internal mail
> if you want users to be able to post to the outside world using the same
> e-mail addresses.  You won't be able to post from a domain name that's
> not recognised outside your LAN.  There's nothing stopping you from
> having different responses to domain names inside and outside of your
> LAN (i.e. using a public domain name, inside and out, but inside your
> machines all have internal LAN IP addresses, for internal work, outside
> your domain has a real internet IP address for mail checks, etc.).

Everything that goes out is masqueraded as coxagri.com. It goes out to the 
smarthost at the ISP currently. My test ones made it out of the system OK to 
an external address. Is there anything else to be aware of?

Thanks

Bryan