VPN Dropping -- Ping times sometimes as hi as 792ms

[Jillp]

We have (2) sites with (1) connected via VPN to our primary plant. Cisco 1800 series routers on each side. Sometimes -- randomly we loose the VPN connection, sometimes for 10 seconds sometimes for 10 minutes. When I ping from the primary plant I am seeing some really unusual ping times. Sometimes it's between 20ms - 30ms, sometimes as high as 792ms.

Is there any way to determine what is causing this problem? I keep leaning towards my internet provider -- since it is sporadic, but how could I prove it?

Any help would be greatly appreciated, we have NO local consultant that can even offer me advice.

Jill

 

[Grenage]

Try pinging other servers on the Internet, from both ends of the VPN. Are the VPN's on the same ISP? If the delays are swinging wildly in the wind, it's most likely the ISP.

"We can categorically state that we have not released man-eating badgers into the area" - Major Mike Shearer

 

[Jillp]

There lies the problem.

If my consultants, (who are very confused) are correct. Plant 2 users, whose only access is via remote terminal server. Plant 2 router sends via our TW cable to our Plant 1, where router is configured (theoretically) to route all traffic from Plant 2 Subnet through the TW cable installed at Plant 1.

What muddies the water is the 'failover' scenario. If the cable isp goes down we have T1's from competing providers that these routers fail over to. The failover works great since time-warner had a 12 hour outage in our area last week, and, while slow my Plant 2 users did have a connection.

We've had cisco tech's look at this, as well as the consultant who installed it (who did this via remote assistance, since he's 7 state away). Cisco replaced the router on the off chance that it was bad, to no avial.

I'm desparate, there must be some what we can prove what is going on here? As admin I can ping locally at Plant 2, though I'm not sure this local connection is the same that terminal-server is routed through. I will try you ping other server suggestion. Would you surmise that the sporadic behavior must be ISP?

Thanks for your advice, I appreciate the response. Jill

 

[Grenage]

When you were working on the failover connection, was latency varying as much? Is it possible to use the failover link for testing purposes?

I am assuming here that the cable and T1 connections are maintained by different ISPs.

"We can categorically state that we have not released man-eating badgers into the area" - Major Mike Shearer

 

[Jillp]

Don't know the answer to the latency question. I can't do the testing via the failover link. If I could find someone that really understood this stuff, I'm sure they could.

I am currently at Plant 2 pinging Google. The ping is great 20 - 30ms, but it indicates that the T1 vendor is hosting this connection. I think the router looks at the destination of the traffic (primary plant subnet) and routes Plant 1 traffic 'only' over the TW cable connection. So I'm not sure this is a fair test.

Also, at Plant 1 earlier, I did a ping to Google and the ping time was 100ms or more, this also indicated my 'other' T1 vendor was hosting the connection. Again, I believe incoming traffic from Plant 2 subnet is routed to TW cable connection via my plant 1 Cisco router

I had TW come check the connection at Plant 1 a couple of weeks ago. They showed me evidence that the connection had been active for 19 days.

Quite a condundrum, huh?

 

[Grenage]

Yes, I see what you mean. Can you connect to the Cisco router at one of the plants, and traceroute then ping the router's IP address at the other plan?

"We can categorically state that we have not released man-eating badgers into the area" - Major Mike Shearer

 

[Jillp]

Maybe. I'll check it out.

 

[North323]

post scrubbed configs of the devices too

 

[Jillp]

Sorry, scrubbed?

 

[North323]

ya configs with all your IP's and passwords scrubbed out...