dnsjava-users — The dnsjava mailing list

DNSJAVA will not load

[Wai P. <sep...@gm...>]

Hi there,

I managed to replaced java's default dns ns with DNSJava in my application.

System.setProperty("sun.net.spi.nameservice.provider.1","dns,dnsjava");

When I run my application as it is from eclipse, my application works.
It loads DNSJAVA. However, when I deployed my application via JNLP, my
application used the native DNS. Is there anything i am missing here?

Thanks in advance!
--
Est Solaris Oth Mithas

getLocalHost throws UnknownHostException

[Wai P. <sep...@gm...>]

Hi

I am running dnsjava to replace the default dns in java. I am using
windows xp and JRE 1.5.0_05. When I try to perform a
InetAddress.getLocalHost, I get an UnKnownHostException. Anyone know's
how to overcome this?

Thanks in advance!
--
Est Solaris Oth Mithas

Address.getByName() returning an invalid InetAddress object?

[Praveen T. <pra...@gm...>]

Hi,

I have a machine setup with two different DNS names. (deathstar =3D zpac23 =
=3D
172.25.181.23 <http://172.25.181.23/> and zpac23 is the canonical hostname)=
.

When I do this:

InetAddress xaddr =3D Address.getByName("deathstar");
InetAddress saddr =3D InetAddress.getByName("deathstar");
System.out.printlin(xaddr.getHostName());
System.out.printlin(saddr.getHostName());

-------------- This is the output I get: ----------------
zpac23
deathstar
--------------

The documentation for
InetAddress.getHostName()<http://java.sun.com/j2se/1.4.2/docs/api/java/net/=
InetAddress.html#getHostName%28%29>,
says this:

"Gets the host name for this IP address.

If this InetAddress was created with a host name, this host name will be
remembered and returned; otherwise, a reverse name lookup will be performed
and the result will be returned based on the system configured name lookup
service........"

So, as per this, Address.getByName() isn't returning a proper InetAddress
object. Right? Is there a solution someone can suggest? I've tried this wit=
h
1.6.4 and 2.0.0 and got the same result. I am using dnsjava in the most
simplistic manner - one call to Lookup.setDefaultResolver() for setup and
regular calls to Address.getByName().

I am using dnsjava with jCIFS and this behavior is causing problems with
jCIFS (triggering an apparent problem with SMB servers, to be precise).

Thanks in advance
-Praveen

Re: properties for DNSJavaNameService

[Brian W. <bwe...@xb...>]

On Mon, 8 Aug 2005, Stefan Groschupf wrote:

> i wish to use javadns to fix some bottleneck my webcrawler has with dns.
> Actually I would love to test the usage of javadns as alternative 
> implementation of sun.net.spi.nameservice.NameService.
>
> I notice that Paul Cowan is one of the authors of 
> org.xbill.DNS.spi.DNSJavaNameService.
> Since I had seen a older implementation of this interface from Paul that 
> supports some properties
> (http://buzzsurf.com/java/dns/doc/com/buzzsurf/dns/DNSJavaNameService.html)
> I asking my self if this properties are still supported but may not 
> documented.

The included name services provider uses the:
 	sun.net.spi.nameservice.nameservers
 	sun.net.spi.nameservice.domain
 	java.net.preferIPv6Addresses
as documented.

> I'm very interested to provide a set of DNS servers and the possibility to 
> load balance in a round robbing style, can this somehow configured just by 
> system properties or do I need to code it myself somehow.

Setting the servers can be done using system properties.  To enable 
round-robin load balancing, you'd need to write code which called 
Lookup.getDefaultResolver() to obtain the default Resolver object and 
ExtendedResolver.setLoadBalance() to enable load balancing.

Brian

properties for DNSJavaNameService

[Stefan G. <sg...@me...>]

Hi there,
i wish to use javadns to fix some bottleneck my webcrawler has with dns.
Actually I would love to test the usage of javadns as alternative  
implementation of sun.net.spi.nameservice.NameService.

I notice that Paul Cowan is one of the authors of  
org.xbill.DNS.spi.DNSJavaNameService.
Since I had seen a older implementation of this interface from Paul  
that supports some properties
(http://buzzsurf.com/java/dns/doc/com/buzzsurf/dns/ 
DNSJavaNameService.html)
I asking my self if this properties are still supported but may not  
documented.

I'm very interested to provide a set of DNS servers and the  
possibility to load balance in a round robbing style, can this  
somehow configured just by system properties or do I need to code it  
myself somehow.

Thanks for any hints.
Greetings.
Stefan
  

jnamed.conf

[Matthias J. <ma...@ho...>]

Hello,

I would like to use dnsjava as local DNS Server in my intranet/extranet.

This is the model:

First Computer: (Windows)
Hostname: multiserver   (dhcp, http and ftp server)
IP: 131.107.2.200

Client Computer: (Windows)
Hostname: dhcp1
IP: 131.107.2.201

At the moment I have to enter "http://multiserver" to query my http server, 
but I would like to use
http://www.myserver.sql for doing this.

And I would like to use jana.mail.smtp and jana.mail.pop3 in my email 
clients instead of "multiserver".

I am really thankful if someone would help me to configure my system for 
this.

With kind regards


Matthias

RE: dnsjava non-English windows support - 3

[Brian W. <bwe...@xb...>]

On Wed, 6 Jul 2005, Holger Stenzhorn wrote:

> I have created property files for the languages Spanish, French, Italian 
> and Portuguese. So dnsjava now works on native Windows systems in those 
> languages. I hope that helps a bit... Since I have no CVS commit access 
> I simply attached the contents of the property files below.

Thanks.  You also might want to try the current CVS code - there's new 
code (added yesterday) to use an undocumented API to get the DNS 
properties from the JVM itself (which presumably uses native code to 
retrieve it from the OS), which should be language independent.  It will 
fall back to using the output of ipconfig on Windows if that fails, so 
extra language support is still useful.

Brian

RE: dnsjava non-English windows support - 3

[Holger S. <Hol...@xt...>]

Hi,

I have created property files for the languages Spanish, French, Italian
and Portuguese.
So dnsjava now works on native Windows systems in those languages. I
hope that helps a bit...
Since I have no CVS commit access I simply attached the contents of the
property files below.

Bye,
Holger

DNSServer_es.properties:
------------------------
host_name=3DNombre del host
primary_dns_suffix=3DSufijo DNS principal
dns_suffix=3DLista de b\u00FAsqueda de sufijos DNS
dns_servers=3DServidores DNS


DNSServer_fr.properties:
------------------------
host_name=3DNom de l'h\u00F4te
primary_dns_suffix=3DSuffixe DNS principal
dns_suffix=3DListe de recherche de suffixe DNS
dns_servers=3DServeurs DNS


DNSServer_it.properties:
------------------------
host_name=3DNome host
primary_dns_suffix=3DSuffisso DNS primario
dns_suffix=3DElenco di ricerca suffissi DNS
dns_servers=3DServer DNS


DNSServer_pt.properties:
------------------------
host_name=3DNome do host
primary_dns_suffix=3DSufixo DNS prim\u00E1rio
dns_suffix=3DLista de pesquisa de sufixo DNS
dns_servers=3DServidores DNS

RE: dnsjava non-English windows support

[Brian W. <bwe...@xb...>]

On Tue, 5 Jul 2005, Holger Stenzhorn wrote:

> I just tried out the language support on German Windows and it works
> fine there.
>
> ...except that you have to change either the property files or the class
> ResolverConfig.java:
> In the class it is tried to access the "primary_dns_suffix_key" in the
> properties but in the actual key is only called "primary_dns_suffix".
> (I would opt to change the class to use "primary_dns_suffix" without the
> "_key" at the end because the other property keys ("host_name",
> "dns_suffix" and "sns_servers") are also without the "_key" at the end,
> right?)

As you noticed later, it's fixed in 2.0.0.

> Also should the ant-build script be changed to include the windows
> directory with the property files into the packaged JAR file?

Yes, it should.  I'll fix that before the 2.0.1 release.

Thanks,
Brian

RE: dnsjava non-English windows support - 2

[Holger S. <Hol...@xt...>]

Hi again,
=20
I was pointing to version 1.6.6 in my last mail.
As I see the problem with the "_key" at the end of "primary_dns_suffix" =
is fixed in 2.0.0.
=20
Bye,
Holger
=20
-----Urspr=FCngliche Nachricht-----
Von: dns...@li... =
[mailto:dns...@li...] Im Auftrag von Holger =
Stenzhorn
Gesendet: Dienstag, 5. Juli 2005 18:14
An: dns...@li...
Betreff: RE: dnsjava non-English windows support=20


Hi,
=20
I just tried out the language support on German Windows and it works =
fine there.
=20
...except that you have to change either the property files or the class =
ResolverConfig.java:
In the class it is tried to access the "primary_dns_suffix_key" in the =
properties but in the actual key is only called "primary_dns_suffix".
(I would opt to change the class to use "primary_dns_suffix" without the =
"_key" at the end because the other property keys ("host_name", =
"dns_suffix" and "sns_servers") are also without the "_key" at the end, =
right?)
=20
Also should the ant-build script be changed to include the windows =
directory with the property files into the packaged JAR file?
=20
Bye,
Holger
=20
 ----------
I just applied a patch which supports reading the output of ipconfig or
 winipcfg in alternate languages, and added support for German and =
Polish. =20
 If anyone reading this has a way of testing the change (I don"t use
 windows, and only speak English), it would be greatly appreciated. =20
 Additionally, if anyone has a version of Windows in a different =
language,
 and could submit a simple properties file like the existing ones to add
 support for that language, it would also be appreciated.
=20
 Thanks,
 Brian

RE: dnsjava non-English windows support

[Holger S. <Hol...@xt...>]

Hi,
=20
I just tried out the language support on German Windows and it works
fine there.
=20
...except that you have to change either the property files or the class
ResolverConfig.java:
In the class it is tried to access the "primary_dns_suffix_key" in the
properties but in the actual key is only called "primary_dns_suffix".
(I would opt to change the class to use "primary_dns_suffix" without the
"_key" at the end because the other property keys ("host_name",
"dns_suffix" and "sns_servers") are also without the "_key" at the end,
right?)
=20
Also should the ant-build script be changed to include the windows
directory with the property files into the packaged JAR file?
=20
Bye,
Holger
=20
 ----------
I just applied a patch which supports reading the output of ipconfig or
 winipcfg in alternate languages, and added support for German and
Polish. =20
 If anyone reading this has a way of testing the change (I don"t use
 windows, and only speak English), it would be greatly appreciated. =20
 Additionally, if anyone has a version of Windows in a different
language,
 and could submit a simple properties file like the existing ones to add
 support for that language, it would also be appreciated.
=20
 Thanks,
 Brian

dnsjava non-English windows support

[Brian W. <bwe...@xb...>]

I just applied a patch which supports reading the output of ipconfig or
winipcfg in alternate languages, and added support for German and Polish.  
If anyone reading this has a way of testing the change (I don't use
windows, and only speak English), it would be greatly appreciated.  
Additionally, if anyone has a version of Windows in a different language,
and could submit a simple properties file like the existing ones to add
support for that language, it would also be appreciated.

Thanks,
Brian

Re: Usage of NS records as part of query response

[Brian W. <Bri...@no...>]

On Fri, 15 Oct 2004, Alexis Smirnov wrote:

> I've recently investigated a problem involving dnsjava, DNS servers operated
> my ISP and a DNS server I own. 

I just responded to this on the discussion forum.

Brian

Usage of NS records as part of query response

[Alexis S. <al...@pi...>]

I've recently investigated a problem involving dnsjava, DNS servers operated
my ISP and a DNS server I own. 
 
The problem can be summarized as follows: 
 
Record[] r = new Lookup("a.domain.com", Type.A).run(); 
// r is not NULL 
r = new Lookup("b.domain.com", Type.A).run(); 
// r is NULL 
// Lookup returns error code TIME OUT 
 
here, both DNS names (a and b) are actually present in DNS. 
 
 
The problem appear to be related to the way dnsjava interprets the first
(successful) response. My ISP's DNS server returned not only A record, but
also NS record ("ns.domain.com") as part of the response. 
 
The second run of Lookup attempts to resove "ns.domain.com" before trying to
recolve "b.domain.com". 
 
It turns out, I forgot to provide A record for "ns.domain.com" as part of my
zone file for domain.com. Fixing the zone file fixes the problem above, but
begs 2 questions:  
Why do some DNS servers send NS records as part of the response?  
Why does dnsjava attemts to resolve NS names? 
 
Thanks! 
Alexis Smirnov
http://weblog.smirnov.ca <http://weblog.smirnov.ca/> 

 

PS: i'm not monitoring this forum so i'd appreciate receiving a CC for any
response.

 

Re: A Patch for Using Sub-DNS with DNSJava

[testttt <qwo...@ya...>]

Brian Wellington wrote:

>On Wed, 31 Mar 2004, testttt wrote:
>
>  
>
>>Hi, Brian and all others,
>>
>>In order to use DNSJava in my box without extensively configuring it, I 
>>found it's handy if DNSJava can query a sub-DNS before it returns an 
>>empty query result. This will work with the *nix resolv library. So, I 
>>did some very simple change to jnamed.java:
>>
>>1. a new keyword 'sub-dns' for jnamed.conf. It has a ip and an optional 
>>port attribute;
>>2. create a SimpleResolver object for each sub-dns attribute;
>>3. When jnamed is returning response to a query, it checks if the answer 
>>section in the response is empty or not;
>>4. If empty, it will try the configured SimpleResolver object to resolve 
>>the query;
>>5. break until the response has non-empty answer section, or all sub-dns 
>>has been tried;
>>
>>So far, it works ok in my machine. I would like to get some comments 
>>from Brian or whoever is interested. And ultimately, I'd like this 
>>feature in DNSJava or merge my change into DNSJava. Please let me know!
>>    
>>
>
>It looks like you're trying to implement something like BIND's forwarders, 
>which is a reasonable thing to do.  It's unclear from your code what you 
>actually want to do with the queries, though - checking for a non-empty 
>answer section isn't right, because if a server returns a valid negative 
>answer, you should believe it and not ask anyone else.
>
>The bigger question, though, is why are you using the dnsjava server?  
>Its error handling is virtually non-existant, it's not fast, and it 
>doesn't do much.  I'd really recommend just running BIND with forwarders.
>
>Brian
>
>  
>
You are right, my main purpose is not the forwarder feature, instead I 
need another feature which I call "Virtual Zone".  I am not sure if this 
is correct name for this feature, but let me explain below:

I need to access my company's intranet from home using SOCKS5. I use 
TSOCKS for this. It works fine if I know the IP. But I really need DNS 
to work since there are thousands of machines in the intranet and no 
intranet DNS server is on public internet. TSOCKS only handles TCP and 
DNS query is using UDP by default. Yes, TSOCKS has a USE_TCP_4_DNS 
option, but it doesn't seem to work on linux.

So, my workaround is this virtual zone feature: when the DNS server got 
a request for a DNS name which is in the intranet, it would simly return 
a virtual IP address, 0.1.0.3, for example. When TSOCKS receives a 
connect request and the IP is in 0.1.0.x range, it would query the DNS 
server to convert the IP to the DNS name and it then uses the DNS name 
in SOCKS5 connect request. The SOCKS5 server will then query the DNS 
server in the intranet and handle it from there.

I've done the prelimiary virtual zone work in DNSJava and it works great 
with my hacked TSOCKS. However, since DNSJava doesn't have this 
forwarder feature, it doesn't know how to handle those normal DNS 
queries. I've added this forwarder feature and it seems to work quite 
fine. I know I could do this virtual zone feature in BIND, but I haven't 
done C/C++ in the past 4 years.

(Hmm, I think I can use BIND as my DNS server and use DNSjava as a 
forward target in BIND. Cool, this should work too. )

This virtual zone feature is vital in my situation. There are some 
commercial product which has this feature, Aventail Connect client, for 
example. Aventail Connect will intercept the call to gethostbyname() and 
it would do the virtual zone stuff there. But Aventail Connect is 
windows only.

I think I should be able to do the same as Aventail: intercept the call 
to gethostbyname(). I worked on this for a while without success. I also 
blamed this becasue I am a Java guy now.

Anyway, please me know if you'd like to include any of my work into your 
project. I am a very experienced Java guy (may not a DNS expert though).

Thanks!

Ben

Re: A Patch for Using Sub-DNS with DNSJava

[Brian W. <bwe...@xb...>]

On Wed, 31 Mar 2004, testttt wrote:

> Hi, Brian and all others,
> 
> In order to use DNSJava in my box without extensively configuring it, I 
> found it's handy if DNSJava can query a sub-DNS before it returns an 
> empty query result. This will work with the *nix resolv library. So, I 
> did some very simple change to jnamed.java:
> 
> 1. a new keyword 'sub-dns' for jnamed.conf. It has a ip and an optional 
> port attribute;
> 2. create a SimpleResolver object for each sub-dns attribute;
> 3. When jnamed is returning response to a query, it checks if the answer 
> section in the response is empty or not;
> 4. If empty, it will try the configured SimpleResolver object to resolve 
> the query;
> 5. break until the response has non-empty answer section, or all sub-dns 
> has been tried;
> 
> So far, it works ok in my machine. I would like to get some comments 
> from Brian or whoever is interested. And ultimately, I'd like this 
> feature in DNSJava or merge my change into DNSJava. Please let me know!

It looks like you're trying to implement something like BIND's forwarders, 
which is a reasonable thing to do.  It's unclear from your code what you 
actually want to do with the queries, though - checking for a non-empty 
answer section isn't right, because if a server returns a valid negative 
answer, you should believe it and not ask anyone else.

The bigger question, though, is why are you using the dnsjava server?  
Its error handling is virtually non-existant, it's not fast, and it 
doesn't do much.  I'd really recommend just running BIND with forwarders.

Brian

A Patch for Using Sub-DNS with DNSJava

[testttt <qwo...@ya...>]

Hi, Brian and all others,

In order to use DNSJava in my box without extensively configuring it, I 
found it's handy if DNSJava can query a sub-DNS before it returns an 
empty query result. This will work with the *nix resolv library. So, I 
did some very simple change to jnamed.java:

1. a new keyword 'sub-dns' for jnamed.conf. It has a ip and an optional 
port attribute;
2. create a SimpleResolver object for each sub-dns attribute;
3. When jnamed is returning response to a query, it checks if the answer 
section in the response is empty or not;
4. If empty, it will try the configured SimpleResolver object to resolve 
the query;
5. break until the response has non-empty answer section, or all sub-dns 
has been tried;

So far, it works ok in my machine. I would like to get some comments 
from Brian or whoever is interested. And ultimately, I'd like this 
feature in DNSJava or merge my change into DNSJava. Please let me know!

Ben.


// below are some codelet I copied out:
...
private ArrayList resolvers = new ArrayList();
.....
             // add any sub-dns configured:
            else if(keyword.equals("resolver")){
                String host = st.nextToken();
                String port = null;
                if(st.hasMoreTokens())
                    port = st.nextToken();
               
                addResolver(host, port);
            }
...

private void addResolver(String host, String port) throws 
UnknownHostException{
    SimpleResolver resolver = new SimpleResolver(host);
    if(port!=null && port.length()>0)
        resolver.setPort(Integer.parseInt(port));
   
    resolvers.add(resolver);
}
....

       // then when jnamed is going to return a query response, it do this:
       boolean hasAnswer = ... // if the response's answer section is 
not empty.
        if(resolvers.size()>0 && !hasAnswer){
        Iterator it = resolvers.iterator();
        while(it.hasNext()){
            Resolver resolv = (Resolver)it.next();
            response = resolv.send(query);
            if(response.hasAnswer())
                break;
        }
    }

[SPAM sbl-xbl.spamhaus]Re: finding name server / Windows 2003

[Hontvari J. <hon...@so...>]

I checked the source, there are two methods for windows. Dnsjava checks if
the OS is a windows nt descendant, then calls ipconfig, otherwise (win 3.x?,
9x) it calls winipcfg. 2003 is not checked, so it Windows 2003 is handled as
a win 9x. Becase all future windows version will be nt descendant, it would
be better to do the opposite: explicitly check for old windows and the
default should be nt. But I don't know what are the 9.x and 3.x specific
strings.


----- Original Message ----- 
From: "Hontvari Jozsef" <hon...@so...>
To: <dns...@li...>
Sent: Saturday, January 31, 2004 10:29 PM
Subject: finding name server / Windows 2003


> Somehow dnsjava cannot find the local name server on a Windows 2003
machine.
> It is working well on other Windows os-s.
>
> If I use -Ddnsjava.options=verbose and do a lookup, it is shown that it
> sends the requests to 127.0.0.1:53. I know this is the fallback address.
The
> windows ipconfig command correctly displays two DNS servers.
>
> Any idea why it doesn't use the dns servers returned by ipconfig?
>
>
>
> -------------------------------------------------------
> The SF.Net email is sponsored by EclipseCon 2004
> Premiere Conference on Open Tools Development and Integration
> See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
> http://www.eclipsecon.org/osdn
> _______________________________________________
> dnsjava-users mailing list
> dns...@li...
> https://lists.sourceforge.net/lists/listinfo/dnsjava-users

finding name server / Windows 2003

[Hontvari J. <hon...@so...>]

Somehow dnsjava cannot find the local name server on a Windows 2003 machine.
It is working well on other Windows os-s.

If I use -Ddnsjava.options=verbose and do a lookup, it is shown that it
sends the requests to 127.0.0.1:53. I know this is the fallback address. The
windows ipconfig command correctly displays two DNS servers.

Any idea why it doesn't use the dns servers returned by ipconfig?

Re: codepage

[Brian W. <bwe...@xb...>]

On Thu, 6 Nov 2003, Walker, Aaron  L wrote:

> I'm trying to get dnsjava to work on my OS390 system, specifically to get
> the java email server working.  My problem is conversion/codepage.  I'm
> working with ebcdic, and I need to alter anything that reads or writes to
> the outside world to change its codepage.  I've converted over the email
> server, and with regards to dnsjava:  do I only need to worry about working
> with the DataByteInputStream and DataByteOutputStream classes?  I think I'm
> only worried about doing Lookups.

To be honest, I have no idea.  The Name class definitely assumes ASCII,
and the conversions between wire/text format in the Record class, and the
Tokenizer class (which wouldn't be used for lookups).  I'm sure other
things do as well.

The other problem is that any change will have a non-trivial performnace
impact, and since the number of systems not using ASCII is so small, any   
changes to support alternative encodings would not be integrated
(although could be distributed as a patch).

Sorry I can't be more helpful...

Brian

codepage

[Walker, A. L <Aar...@tw...>]

I'm trying to get dnsjava to work on my OS390 system, specifically to get
the java email server working.  My problem is conversion/codepage.  I'm
working with ebcdic, and I need to alter anything that reads or writes to
the outside world to change its codepage.  I've converted over the email
server, and with regards to dnsjava:  do I only need to worry about working
with the DataByteInputStream and DataByteOutputStream classes?  I think I'm
only worried about doing Lookups.

Thanks,
Aaron

Re: DNSJava failing to lookup records?

[Andrew H. <aa...@vo...>]

Great!  That's more than enough to go on, and gives me all the 
information I need to work this out.  I think my best bet is simply to 
log certain key pieces of information; if it happens again I'll 
(hopefully) have enough info to track down the problem.

Thanks for the response,

- a.

Brian Wellington wrote:

>On Mon, 20 Oct 2003, Andrew Houghton wrote:
>
>  
>
>>Here at VolunteerMatch we use dnsjava (1.3.3) to do some basic semantic 
>>verification of user email addresses before they're added to our 
>>system.  Users enter an email address for most important site actions, 
>>and we do logical RFC 2822 checks followed by a simple test along the 
>>lines of 'does this domain have an MX record? if not, does it have an A 
>>record?'.
>>
>>On Sunday, after restarting our servlet engine (Jetty 4.2.12 running on 
>>RH Linux 7.3, Sun JDK 1.4.0), every record lookup was returning null.  
>>This has happened once before on a brand new machine, and we chalked it 
>>up to setup problems (we had some problems in our resolv.conf that 
>>needed fixing).  This time around, though, there weren't any changes in 
>>the system to cause the problem.  For the record, using dig to lookup 
>>records from a shell worked fine for the same domains that dnsjava was 
>>failing on.
>>    
>>
>
>What does Lookup.getErrorString() return for the failed queries?
>
>  
>
>>My first guess was that dnsjava simply had some problems reading or 
>>parsing resolv.conf, and thus the initialization step was booched, but 
>>that doesn't make much sense to me given that on any normal day it 
>>doesn't have this problem.  I'm about to dive into the code to try and 
>>figure out what's going on, but can anyone offhand give any ideas?  Has 
>>anyone seen this problem?
>>    
>>
>
>I've never heard of any problems with the resolv.conf parser, but it's 
>possible.
>
>  
>
>>And, most importantly, is there a way to force dnsjava to initialize and 
>>return some sort if a status code?  I'm not certain what a status code 
>>would mean in this case, but ideally I'm looking for some way of finding 
>>what dnsjava is using as name servers, how it decided on them, and 
>>whether the initialization process had any problems.
>>    
>>
>
>Setting the verbose option (-Ddnsjava.options=verbose) will cause the 
>resolv.conf parser, called from the initialization code, to print a 
>message for each server it parses (in the FindServer.addServer() method).  
>There's also the FindServer.servers() static method, which returns the 
>servers being used (or the FindServer.searchPath() method returning the 
>domain search list).  Setting the verbose option will also cause each 
>lookup to log a message, so while useful for debugging, it's probably not 
>something you'd want on all the time.
>
>I'm not sure how much this helps, but it should tell you if there's a 
>problem with initialization or something later.
>
>Brian
>
>
>  
>

Re: DNSJava failing to lookup records?

[Brian W. <bwe...@xb...>]

On Mon, 20 Oct 2003, Andrew Houghton wrote:

> Here at VolunteerMatch we use dnsjava (1.3.3) to do some basic semantic 
> verification of user email addresses before they're added to our 
> system.  Users enter an email address for most important site actions, 
> and we do logical RFC 2822 checks followed by a simple test along the 
> lines of 'does this domain have an MX record? if not, does it have an A 
> record?'.
> 
> On Sunday, after restarting our servlet engine (Jetty 4.2.12 running on 
> RH Linux 7.3, Sun JDK 1.4.0), every record lookup was returning null.  
> This has happened once before on a brand new machine, and we chalked it 
> up to setup problems (we had some problems in our resolv.conf that 
> needed fixing).  This time around, though, there weren't any changes in 
> the system to cause the problem.  For the record, using dig to lookup 
> records from a shell worked fine for the same domains that dnsjava was 
> failing on.

What does Lookup.getErrorString() return for the failed queries?

> My first guess was that dnsjava simply had some problems reading or 
> parsing resolv.conf, and thus the initialization step was booched, but 
> that doesn't make much sense to me given that on any normal day it 
> doesn't have this problem.  I'm about to dive into the code to try and 
> figure out what's going on, but can anyone offhand give any ideas?  Has 
> anyone seen this problem?

I've never heard of any problems with the resolv.conf parser, but it's 
possible.

> And, most importantly, is there a way to force dnsjava to initialize and 
> return some sort if a status code?  I'm not certain what a status code 
> would mean in this case, but ideally I'm looking for some way of finding 
> what dnsjava is using as name servers, how it decided on them, and 
> whether the initialization process had any problems.

Setting the verbose option (-Ddnsjava.options=verbose) will cause the 
resolv.conf parser, called from the initialization code, to print a 
message for each server it parses (in the FindServer.addServer() method).  
There's also the FindServer.servers() static method, which returns the 
servers being used (or the FindServer.searchPath() method returning the 
domain search list).  Setting the verbose option will also cause each 
lookup to log a message, so while useful for debugging, it's probably not 
something you'd want on all the time.

I'm not sure how much this helps, but it should tell you if there's a 
problem with initialization or something later.

Brian

DNSJava failing to lookup records?

[Andrew H. <aa...@vo...>]

Folks --

Here at VolunteerMatch we use dnsjava (1.3.3) to do some basic semantic 
verification of user email addresses before they're added to our 
system.  Users enter an email address for most important site actions, 
and we do logical RFC 2822 checks followed by a simple test along the 
lines of 'does this domain have an MX record? if not, does it have an A 
record?'.

On Sunday, after restarting our servlet engine (Jetty 4.2.12 running on 
RH Linux 7.3, Sun JDK 1.4.0), every record lookup was returning null.  
This has happened once before on a brand new machine, and we chalked it 
up to setup problems (we had some problems in our resolv.conf that 
needed fixing).  This time around, though, there weren't any changes in 
the system to cause the problem.  For the record, using dig to lookup 
records from a shell worked fine for the same domains that dnsjava was 
failing on.

My first guess was that dnsjava simply had some problems reading or 
parsing resolv.conf, and thus the initialization step was booched, but 
that doesn't make much sense to me given that on any normal day it 
doesn't have this problem.  I'm about to dive into the code to try and 
figure out what's going on, but can anyone offhand give any ideas?  Has 
anyone seen this problem?

And, most importantly, is there a way to force dnsjava to initialize and 
return some sort if a status code?  I'm not certain what a status code 
would mean in this case, but ideally I'm looking for some way of finding 
what dnsjava is using as name servers, how it decided on them, and 
whether the initialization process had any problems.

Thanks for any help,

- a.

p.s. Yes, the versions of dnsjava and the JDK are old.  They're both 
being upgraded soon, but the question still holds.

dnsjava 1.4.1 released

[Brian W. <bwe...@xb...>]

A few bugs have been found in 1.4.0, and 1.4.1 fixes them.  No new 
features were added, but there was one optimization to the Name.equals() 
method.

Changelog below.

Brian

------------

9/2/2003
	- 1.4.1 released.

9/1/2003
	- The Update.delete(Record) method needs to reset the TTL of the cloned
	  record to 0. (reported by Edwin R. Rivera <er...@mo...>)

8/23/2003
	- Change Record.fromString() to expect an EOL/EOF token at the end
	  of the string, so that extraneous tokens can be detected.  The master
	  file parser was updated to this change.
	  (noticed by Bob Halley <bob...@no...>)

8/22/2003
	- Changing TXTRecord.getStrings() from returning a List of Strings
	  to a List of byte []'s was a bad idea.  Change it back, and
	  add TXTRecord.getStringsAsByteArrays().
	  (reported by Blake Ramsdell <bl...@br...>)