I'm not an Expert. Take it easy on me! I'm flying by the seat of my pants here, but we do have a Tech person I can pass this along to. He has tried everything he can think of for the past three months with no results.
We have the following conditions.
Problem: Internet connection is lost at random times throughout the day. Internal network connections remain alive on everyone, but access to the Internet is lost for about ten minutes -- but only certain Users -- and this will occur throughout the day, off and on. Every computer can still access all network drives during the outages.
Server is Windows 2003.
Internet connection is DSL provided by XO Communications.
DSL Modem connects to Firewall box.
Firewall is connected to one of three 3Com Baseline Switches - not managed.
There are about 25 Users out of 40 that have Internet access.
There is a mix of computers running XP Pro, NT, Win98 or Win2K.
You always have access to the Network, itself.
Periodically, groups of Internet-access people will not be able to reach the Internet.
Lasts only about 10 minutes during each outage.
We have tried...
XO Communications could not find any problems after monitoring the line for two weeks.
Outages occurred during that time but nothing shows up on their equipment.
We then isolated to one computer connected to the DSL Router, bypassing Firewall and the Network -- connects every time during an outage. Therefore, XO claims it is not them.
Reconnect Router and those who were experiencing an outage still have it until it ends.
We've tried turning off NetBEUI (leftover from an old installation of the original NT Server).
We've replaced all three Switches.
We've replaced the Firewall.
We've replaced the Router.
We've disconnected all Printers during an outage -- no effect.
We've scanned with PestPatrol, SpySweeper, etc. to clear out all Spyware, Malware, etc.
We have Norton Antivirus Corporate Edition feeding down from the Server to all machines.
We've run 3Com and other packet sniffers looking for activity -- all we've found are a large number of Port 1900 activity indicators, but nothing in the way of files being transferred or sent or received.
Only odd behaviour that seems to be happening is all Ports on the back of the Switches are blinking on and off at the exact same time. That seemed to be happening only during the outages when the original Switches were installed. But now it occurs almost all the time on the new Switches.
I saw a similar posting here referring to NAT configuration problems restricting the number of connections.
Anybody think that might be the problem here?
Any other ideas?
Outage seems to move around. It does not appear to be the same people all the time, but the outage seems to last the same period of time. It does not affect only one type of Operating System -- people with XP could be "disconnected" just as easily as those running NT or anything else.
My questions to our Tech guy is...
a) Why can't we monitor all connections and all activity throughout the day and get a report?
b) Can't we setup some kind of constant ping situation so we can see a report of which computer loses the Internet, and when?
Ideas have been that it could be a bad NIC somewhere within the Network -- but why would that cause random blocking and not affect the Network connections, only the Internet?