Bind 8 special characters

[A.R.]

System: Red Hat 6.0

Version: Bind 8

Problem: Special Characters

Description:

We are implementing dns on this red hat box.
Some computer names on this existing network
have "underscore" characters (_) in their host
name. Is it possible to use underscore characters
in this version of bind.

We have added these computer names as A records
and as CNAME records. An nslookup on these names and
ip addresses do not give a result.

Can this work.

Thank you in advance.

PLEASE email ***@yahoo.com or ***@conres.com.
--
Thanks,

Dr. A. Rahman, Ph.D.
***@yahoo.com


Sent via Deja.com http://www.deja.com/
Before you buy.

[[e-mail]]

Post by A.R.
System: Red Hat 6.0
Version: Bind 8
Problem: Special Characters
We are implementing dns on this red hat box.
Some computer names on this existing network
have "underscore" characters (_) in their host
name. Is it possible to use underscore characters
in this version of bind.

The short answer is no. The long answer depends on who you talk to. I
was told by my ISP back some 5 years now that there was a change to the
RFC that striped underscores as a valid character. Just this week I
discovered that is was not an RFC change but a change in the BIND code
that outlawed underscores. Either way, they don't work and you will
need to change them.

What I have done for my hosts with underscores is replace the
underscore with a dash in DNS. Since the machines talk IP anyway, it
really doesn't matter, unless you have an application that is tied to
the host name....

DAVID


* Sent from RemarQ http://www.remarq.com The Internet's Discussion Network *
The fastest and easiest way to search and participate in Usenet - Free!

[Robert Annandale]

Post by A.R.
System: Red Hat 6.0
Version: Bind 8
Problem: Special Characters
We are implementing dns on this red hat box.
Some computer names on this existing network
have "underscore" characters (_) in their host
name. Is it possible to use underscore characters
in this version of bind.

Later versions of BIND check that host names are compliant with RFC 952.

Underscores are not allowed in host name record unless they are in CNAMEs.

Hyphens are allowed in host names provided the hyphen is not the last
character.

You can still use underscores in e-mail address within an SOA record.

RBA.

[Cricket Liu]

Post by Robert Annandale
Later versions of BIND check that host names are compliant with RFC 952.
Underscores are not allowed in host name record unless they are in CNAMEs.

Actually, it's even a little more subtle than that. An underscore may be
legal in an alias for some query types, and illegal for others. For
example,
in

foo_foo in cname bar
bar in a 10.0.0.1
bar in txt "Comment"

For an address query, foo_foo is an illegal owner name. For a TXT
query, foo_foo is legal.

Post by Robert Annandale
Hyphens are allowed in host names provided the hyphen is not the last
character.

Hyphens are allowed in the labels of domain names as long as they're
not the first of last character of the label.

cricket

Acme Byte & Wire
***@acmebw.com
www.acmebw.com

Attend the next Internet Software Consortium/Acme Byte & Wire
DNS and BIND class! See www.acmebw.com/training.htm for
the schedule and to register for upcoming classes.

[Chris Cell]

I have found a way to do this, but I have been told it is not suggested. You
can go into the named.conf in bind 8.x and put:

options {
check-names master warn

This will allow underscores without failure. The defaults are master - fail,
slave - warn. You can also override the options statement and do it on a
zone basis:

zone "zone.com" {
type master;
file "db.file";
check-names warn;
};

This is not how the config will keep it, but it allows for an easier
migration for the network administrators to change the server names (the
names are important - Microsoft environment about to use dns instead of wins
for resolution).

If you have Dns and Bind by O'Reilly go to page 76, it explains all of this
for earlier versions as well as for the 8.x versions.

Thanks,
Chris

-----Original Message-----
Sent: Thursday, December 30, 1999 10:13 AM
Subject: Re: Bind 8 special characters

Post by A.R.
System: Red Hat 6.0
Version: Bind 8
Problem: Special Characters
We are implementing dns on this red hat box.
Some computer names on this existing network
have "underscore" characters (_) in their host
name. Is it possible to use underscore characters
in this version of bind.

[M***@iengines.com]

Post by ***@mantech.com

Post by A.R.
System: Red Hat 6.0
Version: Bind 8
Problem: Special Characters
We are implementing dns on this red hat box.
Some computer names on this existing network
have "underscore" characters (_) in their host
name. Is it possible to use underscore characters
in this version of bind.

The short answer is no. The long answer depends on who you talk to. I
was told by my ISP back some 5 years now that there was a change to the
RFC that striped underscores as a valid character. Just this week I
discovered that is was not an RFC change but a change in the BIND code
that outlawed underscores. Either way, they don't work and you will
need to change them.

Underscore was NEVER a valid character for a hostname.
See RFC 952 and its predecessors. If you connect to the
Internet you play by its rules.

BIND, by default, rejects zones which contain illegal
hostnames or maildomains.

Mark

Post by ***@mantech.com
What I have done for my hosts with underscores is replace the
underscore with a dash in DNS. Since the machines talk IP anyway, it
really doesn't matter, unless you have an application that is tied to
the host name....
DAVID
* Sent from RemarQ http://www.remarq.com The Internet's Discussion Network *
The fastest and easiest way to search and participate in Usenet - Free!

--
Mark Andrews, Internet Engines Inc. / Internet Software Consortium
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: [phone] INTERNET: ***@iengines.com

[Kevin Darcy]

Post by M***@iengines.com
Underscore was NEVER a valid character for a hostname.
See RFC 952 and its predecessors. If you connect to the
Internet you play by its rules.

Just as a historical note, while RFC 952 is not officially obsolete, its purpose
was to specify the format of the HOSTS.TXT file, which is, in fact, obsolete.
Later RFC's, of course, incorporate by reference the restrictions of RFC 952,
disembodied from the HOSTS.TXT context in which they were originally conceived.
Query: was HOSTS.TXT file inclusion ever enforced as a prerequisite for Internet
connectivity?


- Kevin

[Bill Manning]

% Query: was HOSTS.TXT file inclusion ever enforced as a prerequisite for Internet
% connectivity?
%
% - Kevin

Yes.

--bill

[M***@nominum.com]

Post by Kevin Darcy

Post by M***@iengines.com
Underscore was NEVER a valid character for a hostname.
See RFC 952 and its predecessors. If you connect to the
Internet you play by its rules.

Just as a historical note, while RFC 952 is not officially obsolete, its purp
ose
was to specify the format of the HOSTS.TXT file, which is, in fact, obsolete.
Later RFC's, of course, incorporate by reference the restrictions of RFC 952,
disembodied from the HOSTS.TXT context in which they were originally conceive
d.
Query: was HOSTS.TXT file inclusion ever enforced as a prerequisite for Inter
net
connectivity?

RFC 952 was written with the DNS in mind, early versions
existed when RFC 952 was written. RFC 952 is NOT obsolete,
it is very much the current definintion of what is allowed
in a hostname with one modification from RFC 1123.

Yes the restrictions were enforced. 3com had a big battle
even getting the changes in RFC 1123 endorsed.

Mark

Post by Kevin Darcy
- Kevin

--
Mark Andrews, Nominum Inc. / Internet Software Consortium
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: [phone] INTERNET: ***@nominum.com

[Kevin Darcy]

Post by M***@nominum.com

Post by Kevin Darcy

Post by M***@iengines.com
Underscore was NEVER a valid character for a hostname.
See RFC 952 and its predecessors. If you connect to the
Internet you play by its rules.

Just as a historical note, while RFC 952 is not officially obsolete, its purp
ose
was to specify the format of the HOSTS.TXT file, which is, in fact, obsolete.
Later RFC's, of course, incorporate by reference the restrictions of RFC 952,
disembodied from the HOSTS.TXT context in which they were originally conceive
d.
Query: was HOSTS.TXT file inclusion ever enforced as a prerequisite for Inter
net
connectivity?

RFC 952 was written with the DNS in mind, early versions
existed when RFC 952 was written. RFC 952 is NOT obsolete,
it is very much the current definintion of what is allowed
in a hostname with one modification from RFC 1123.

I didn't say RFC 952 was obsolete; I said that the thing (HOSTS.TXT) the format
for which RFC 952 defines, is obsolete. That's two levels of indirection away from
being obsolete. And as for the difference between 1123 "promoting" 952 to the
status of a DNS standard versus simply borrowing a bunch of restrictions from 952
to create a separate DNS standard, that's just a matter of interpretation.

Post by M***@nominum.com
Yes the restrictions were enforced. 3com had a big battle
even getting the changes in RFC 1123 endorsed.

That's an answer, but not really to the question I asked. Did hosts have to
included in HOSTS.TXT in order to connect to the Internet? I'm trying to figure
out what you meant by "If you connect to the Internet you play by its rules". If
machines could connect to the Internet without being in HOSTS.TXT, then it appears
your statement is incorrect, at least insofar as "the rules" consisted of RFC 952,
the only RFC you cited.

Then again, maybe we could just clear all of this confusion by citing 1123 as the
standard and forgetting about 952 altogether, except to the extent it is
incorporated-by-reference into 1123. There's too much historical baggage
associated with 952, in my personal opinion.


- Kevin

[Barry Margolin]

Post by Kevin Darcy
That's an answer, but not really to the question I asked. Did hosts have to
included in HOSTS.TXT in order to connect to the Internet? I'm trying to figure
out what you meant by "If you connect to the Internet you play by its rules". If
machines could connect to the Internet without being in HOSTS.TXT, then it appears
your statement is incorrect, at least insofar as "the rules" consisted of RFC 952,
the only RFC you cited.

Before DNS, if you wanted your host to be accessed by name rather than just
address, you had to get the name->address translation inserted into
HOSTS.TXT.

Since every machine had its own copy of HOSTS.TXT, you could put your name
in your local HOSTS.TXT but not the one at the the NIC. Then your users
could get there, but people at other sites wouldn't be able to look up your
machine.

But RFC 952 doesn't say that a machine has to be in HOSTS.TXT, it simply
specifies the format of the file, which includes the format of hostnames.
This is then referenced in many places, include RFC 1123, so it's
entrenched as the standard for hostnames in TCP/IP applications.

[Kevin Darcy]

Post by Barry Margolin

Post by Kevin Darcy
That's an answer, but not really to the question I asked. Did hosts have to
included in HOSTS.TXT in order to connect to the Internet? I'm trying to figure
out what you meant by "If you connect to the Internet you play by its rules". If
machines could connect to the Internet without being in HOSTS.TXT, then it appears
your statement is incorrect, at least insofar as "the rules" consisted of RFC 952,
the only RFC you cited.

Before DNS, if you wanted your host to be accessed by name rather than just
address, you had to get the name->address translation inserted into
HOSTS.TXT.
Since every machine had its own copy of HOSTS.TXT, you could put your name
in your local HOSTS.TXT but not the one at the the NIC. Then your users
could get there, but people at other sites wouldn't be able to look up your
machine.
But RFC 952 doesn't say that a machine has to be in HOSTS.TXT, it simply
specifies the format of the file, which includes the format of hostnames.
This is then referenced in many places, include RFC 1123, so it's
entrenched as the standard for hostnames in TCP/IP applications.

Exactly. Which is why I suggested that RFC 1123 is really the applicable standard.
RFC 952, while a standard, wasn't a DNS standard nor -- as I've pointed out in
another thread -- a comprehensive "hostname" standard. This is just a minor
difference of opinion as to which RFC applies to a given situation.


- Kevin

[Sam Wilson]

... I suggested that RFC 1123 is really the applicable standard.
RFC 952, while a standard, wasn't a DNS standard ...