I’ve run into a very strange TCP/IP problem that I hope someone
can help me sort out. For a long time now I’ve run a QNX 4
server running “apache” and “snmpd”. This node also serves
as my TCP/IP connection to the world. That is I have another
work station that I connect to the outside world through this
node over the QNX network.
I should note that the registered domain name for this server
is schoenbrun.com.
The first problem I saw some time ago was that whenever I accessed
a web URL that ended in .org, I always got a very strange web page.
After much investigation I found that DNS was always appending .com
to all .org names. The web page was just the default for any
name.org.com URL. This only happened when running the web server
on a QNX node not running Socket.
A few days ago I started noticing an even more severe
problem. Basically DNS did not work, sort of. If I ping’d
any well know name including my own, ping would hang. However
nslookup would find the right IP address, which I could then
ping without any problem.
After much playing around, I found that the root of my problem
was that when starting Socket I used this command:
Socket schoenbrun.com &
That is I gave the hostname of this node the value “schoenbrun.com”.
If I changed the name to anything else it started working fine.
In addition the previous .org problem went away. I thought this was
great, I had finally solved my problem. Then after three days of
not getting any mail, I found that my troubles had not ended. When
a message was sent from another system, there were complaints in the
return mail. Once complaint was that it could not find
NAME.schoenbrun.com, where name was what I called the node. Another
error talked about there being a loop in where the mail was going.
Does anyone have any ideas. I’ve l’ve left the system named
schoenbrun.com so I can get mail, but this leaves my web browser
and other DNS dependent clients useless.
Mitchell Schoenbrun --------- maschoen@pobox.com
A couple of things. traceroute to 198.144.200.2 or schoenbrun.com
reports your host is called “m201-130.dsl.tsoft.com”, although
nslookup reports schoenbrun.com correctly. Second is that whois
reports your record was updated today. Is that a coincidence?
What are the chances that the routing tables at your ISP are
messed up? It’s not impossible.
Richard
Mitchell Schoenbrun wrote:
I’ve run into a very strange TCP/IP problem that I hope someone
can help me sort out. For a long time now I’ve run a QNX 4
server running “apache” and “snmpd”. This node also serves
as my TCP/IP connection to the world. That is I have another
work station that I connect to the outside world through this
node over the QNX network.
I should note that the registered domain name for this server
is schoenbrun.com.
The first problem I saw some time ago was that whenever I accessed
a web URL that ended in .org, I always got a very strange web page.
After much investigation I found that DNS was always appending .com
to all .org names. The web page was just the default for any
name.org.com URL. This only happened when running the web server
on a QNX node not running Socket.
A few days ago I started noticing an even more severe
problem. Basically DNS did not work, sort of. If I ping’d
any well know name including my own, ping would hang. However
nslookup would find the right IP address, which I could then
ping without any problem.
After much playing around, I found that the root of my problem
was that when starting Socket I used this command:
Socket schoenbrun.com &
That is I gave the hostname of this node the value “schoenbrun.com”.
If I changed the name to anything else it started working fine.
In addition the previous .org problem went away. I thought this was
great, I had finally solved my problem. Then after three days of
not getting any mail, I found that my troubles had not ended. When
a message was sent from another system, there were complaints in the
return mail. Once complaint was that it could not find
NAME.schoenbrun.com, where name was what I called the node. Another
error talked about there being a loop in where the mail was going.
Does anyone have any ideas. I’ve l’ve left the system named
schoenbrun.com so I can get mail, but this leaves my web browser
and other DNS dependent clients useless.
Mitchell Schoenbrun --------- > maschoen@pobox.com
Previously, Richard R. Kramer wrote in qdn.public.qnx4.devtools:
A couple of things. traceroute to 198.144.200.2 or schoenbrun.com
reports your host is called “m201-130.dsl.tsoft.com”,
The correct IP is 198.144.201.130.
nslookup of schoenbrun.com gets this IP, but nslookup returns
m201-130.dsl.tsoft.com.
although
nslookup reports schoenbrun.com correctly. Second is that whois
reports your record was updated today. Is that a coincidence?
What are the chances that the routing tables at your ISP are
messed up? It’s not impossible.
Well I’ve come to the conclusion that some of the behavior I’ve
seen is related to my ISP. The DSN problem mysteriously was
fixed today. However the .org.com problem remains. That is
I’m left with these choices.
-
Name my node schoenbrun.com and Voyager appends .com to .org
URL’s, at least when using Voyager.
-
Name my node anything else and the .org problem goes away,
but mail sent to my snmpd mail program gets returned.
Mitchell
Mitchell Schoenbrun --------- maschoen@pobox.com