Odd network issues?

[michaellunsford]

my cron scripts that use FTP are failing "unable to locate server"
email is backing up in the mail queue and not going out at all for the last few days.
if I ssh into the server and try most network stuff -- like whois, ping (using a domain name), host, etc, I get a network related error much like the FTP error above.

The ISP seems to think the server needs to be cold rebooted, but it's been more than 48 hours and they still haven't rebooted the server. I've restarted it from ssh, but it doesn't seem to help. The firewall seems to be okay, I even reset the rules and still no go. Any ideas what it could be, or if I can do anything about it without a cold restart?

 

[DaDomainMasta]

the most reliable test if it is your firewall or not - disable it completely, like # /etc/init.d/iptables stop and see if you can ftp and so on. looks like passive ftp range is too narrow

 

[michaellunsford]

still nothing...
[root@www root]# /etc/init.d/iptables stop
Flushing firewall rules: [ OK ]
Setting chains to policy ACCEPT: nat mangle filter [ OK ]
Unloading iptables modules: [ OK ]
[root@www root]# ftp microsoft.com
ftp: microsoft.com: Temporary failure in name resolution

 

[michaellunsford]

looks like I may have found part of the problem. The /etc/resolv.conf file contains a single IP address. That IP address is non-responsive, so I can only guess that whatever it's using for DNS is down.

I added in an address from opendns.org until I can get the ISP to look into it. The good news is ftp and whois are working. The bad news is the mail part of the server is still not responsive. I've checked it in dnsstuff's dns report and it says "unable to connect for 40 seconds." Is there another one of these conf files hidden away somewhere I need to modify?

 

[DaDomainMasta]

make sure NS(es) for your domain name are responding and have all types of records.
try the following way:
# host -t mx [your domainname]
blahblah-you will get MX hostname

# host [MX hostname]
result is MXIP

# ping [MXIP]

for dns stuff the same, but -t ns in host command input

additional test:
# host [yourdomainname] [yourNSip]

 

[michaellunsford]

Okay, this has nothing to do with my domain name DNS configuration in the "ns" name server. This has to do with a file on my server's hard drive called "/etc/resolv.conf" that has a bad IP address in it. I would like to discover the new "nameserver" ip addres for my dedicated server. That address should be on the local network so I don't have to go out on the internet for name resolution.

One change from my previous post, email is now functioning, although DNSStuff still doesn't like how long it takes them to respond (mostly because my nameserver is pointing to an opendns.org IP address somewhere on the internet instead of a local one).