Really Odd Intermittent Fault

Simple office setup,

Code:

   Router         Server       Server 2
192.168.0.220   192.168.0.1   192.168.0.2
     |              |            | 
        Switch ZyXEL Gigabit 24 port
              |              |
        Printers x3    Desktops x12

Intermittently I can't ping the router, so no internet access. Every 1 to 5 minutes for between 30 seconds to 5 minutes.

When I can't ping the router I can ping every other device on the LAN.

I've tried 3 different routers now and they all have the problem. The latest router, a draytek 2700 worked without fault for 2 weeks and now is having the same issue.

I've also tried removed clients in groups of five from the switch when I cannot ping the router, and no combination brings it back instantly.

Any advice on what tools to use or what to test now?

 

As stupid as it sounds- can you test another switch?

 

I had sort of ruled out it being the switch as it doesn't matter which port I plug the router into or what ip I give the router it doesn't work.

Also I can ping any other device on the LAN when I cant ping the router.

I take it from your reply that it could be the switch still?

 

It's unlikely that it's the switch- there's not much to wrong with one- it either works or it doesn't but it's the only thing apart from the router that can go wrong.
Can you test something else- when you have no Internet connectivity from computers connected to the switch, can you just plug a laptop straight into the router and see if it's going to work?
If it's gonna work it's not the router if it doesn't work it's the router.

 

Yep if I plug a laptop into the router and pull the cat6 to the switch it works. EDIT only after changing port

 

I tell a lie, it only works if I change the port it's plugged into in the router

 

It kind of point towards the switch doesn't it?
Do you have another one lying around to test it?

Also change the cable between the switch and the router- just for the sakes of it.

 

Going to purchase D-Link Web Smart DGS-1210-48 Switch | Ebuyer.com and see if that does it.

If not it must be the router... in which case that router 4 that has the problem. Could a dodgy ADSL line kill routers?

 

I don't think it can be an ADSL line. If it was in fact it- it would be far more spectacular than what you're experiencing.

 

I got a brand new 48 port switch and the problem persisted, intermittently cant ping the router so no internet but can ping every other device on the lan.

So I put the router directly into a second NIC on my server 2008 r2 and setup Routing and Remote access to provide NAT for my internal LAN. This took the problem of the internet going down out, but it caused the server to hard reboot twice, checking event log showed a memory dump and i'm not clever enough to work out exactly what happened.

As my server 2008 machine is so vital to my company operating I put together a P4 machine and installed untangle, linux gateway. 2 NICS, router on NIC 1 and LAN on NIC 2.

Appeared to run fine for a bit but suddenly couldn't ping the gateway... I had a console screen running on the local terminal and the router was still responding to requests.

So it appears what ever machine is the default gateway is being intermittently knocked off the net / the lan.

The Smart Switch shows no packet loss or erroneous packets...

There are no ip conflicts on the LAN either.

Any ideas where I can go from here?

 

I have seen a few where one single port will fail and eventually a second, and third... Though not applicable here.

I would almost be willing to bet that you have a bad network card on your network, possibly in the server you used for routing, which would explain the crashing.

Disconnect one machine at a time for a bit and see if the problem persists. It stinks, but it's likely the only way you will track it down.