Forum OpenACS Q&A: Connecting to CapitalMediaOnline.com
Netscape says: "a network error occured. Unable to connect to server
(TCP error: no route to host). The server may be down or
unreachable. Try connecting again later.
Access Log: no record of the event.
Server Log: no errors, no work performed.
Postgres: 7.0. capitalmediaonline is a database created for this
service. The database works fine. The data models are loaded, psql
works fine in it.
Postgres Driver: recompiled. when startind nsd, no complaints (I have
the nsd stuff in a different location than the original driver
compile, so I had to recompile it).
ping 18.104.22.168: no problem, 100%
ps -ef | grep nsd: everything's there
nsd.tcl: everything's configured (hardcoded 22.214.171.124 in the
address, hardcoded capitalmediaonline for the hostname, told the three
pools to look for localhost::capitalmediaonline).
capitalmediaonline.tcl: everything's pointing to capitalmediaonline.
21 references to capitalmediaonline in the file. Otherwise,
everything's identical to what works for another service on the same
DNS: my machine can see capitalmediaonline.com without any trouble.
Timespent: all night. I'm really grouchy, frustrated, and desperate.
As near as I can tell, everythings in line. While the DNS was being
replicated over the past few days, I had an identical setup working
locally. I used the same basic tcl files to get things going, with a
few modifications to look locally instead of worldwide. This service
is running on a dedicated server remotely through digital nation. I
spoke with Unix support over there, and they can't see anything wrong
with the network. However, neither he nor I can get through to
capitalmediaonline.com. Any ideas?
The access log does insert a line if I use the IP address instead of the domain name. It reads:
126.96.36.199 - - [26/Jul/2000:14:06:28 -0400] "GET /test.html HTTP/1.1" 302 243 "" "Mozilla/4.0 (compatible; MSIE 5.0; Windows NT; DigExt)"
However, the server.log doesn't log any activity at all.
I think this tells me that the DNS isn't quite right.
Like I said earlier, I've checked into it. However, I want to double check--each hop doesn't need to know the domain name, right? I mean, the client needs to have access to a DNS entry that says what the IP address is, but the 17 other hops (in my case) don't really care about that, right?
If that works, then try the domain name. My guess is that won't work.
Does the server log have a bunch of stuff in it telling you that the ACS is up and running? You should get messages telling you that various script files in the tcl library directory have been loaded, that Postgres is loaded, etc etc.
Your other advice, however is really useful. I was trying to do the telnet thing, but for some reason, I was thinking that I should have a colon between the IP address and port 80, so of course I couldn't get it working. Now I've got some more ideas on how to fix such problems in the future.