Frame-Relay over subscribed

Answered Question
Jan 10th, 2008
User Badges:

Good day


I have a quick question for a month or two I have been running into a problem where I have a T1 frame circuit with 10 PVCs on it. Out of the 10 I have only 2 PVC 23 and 38 that constantly go down at the same time, no other pvcs on that T1 frame go down. When I say go down I mean to say that the physical is up..I see the inter face as "up up" but when I look to see OSPF adjcencies they are either in init or exchange state. There are no errors being log on the serial interface at both end or the ethernet interface at both ends. T1 card shows no errors, had the carrier look and they see no issues the only thing they have in common is that they are on the same router, same T1 card and ride the same cable from our datacenter to the smartjack. I'm starting to think that the frame-relay circuit is over subscribed. We have a full T1, 2X512, 2X256, 3X128 and 2X64k connections on here. Am I thinking correctly or have I missed something?

Correct Answer by keegan.holley about 9 years 4 months ago

Have you checked the logs on the remote router for interface failures? It seems strange that the same two pvc's would fail at the same time consistiently. Also, how are they configured? You said that you can see the interface up up so I'm assuming that they are point-to-point sub-interfaces. If not do a show frame pvc when you are having the issue and see if the pvc's are down or deleted or when the last flap occured. I've seen problems with frame circuits where the carrier tech only checks the status of the T1 circuit and not the frame switch. If your PVC went down they should have at least seen the status change, which would tell you what end the problem exists on. You are definitely over subscribed but, if you were experiencing congestion it would affect all of the PVC's. Lastly, if the PVC checks out it could be a higher layer problem. OSPF could be flapping due to misconfigurations or low router resources.

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (1 ratings)
Loading.
Edison Ortiz Thu, 01/10/2008 - 14:18
User Badges:
  • Super Bronze, 10000 points or more
  • Hall of Fame,

    Founding Member

When that happens again:


1) Check the PVC status, not just the physical interface (show frame pvc 23 and/or show frame pvc 38)


2) Try pinging the remote-end and verify if the round-trip is high vs when the circuit is running on normal load.


I also suggest implement Frame-Relay traffic shaping on situations where you have a hub & spoke design. Your hub has a T1 circuit while the hubs have lesser speeds. The hub may be trying to send more data than what the spoke really support.


HTH,


__


Edison.

Correct Answer
keegan.holley Thu, 01/10/2008 - 16:42
User Badges:

Have you checked the logs on the remote router for interface failures? It seems strange that the same two pvc's would fail at the same time consistiently. Also, how are they configured? You said that you can see the interface up up so I'm assuming that they are point-to-point sub-interfaces. If not do a show frame pvc when you are having the issue and see if the pvc's are down or deleted or when the last flap occured. I've seen problems with frame circuits where the carrier tech only checks the status of the T1 circuit and not the frame switch. If your PVC went down they should have at least seen the status change, which would tell you what end the problem exists on. You are definitely over subscribed but, if you were experiencing congestion it would affect all of the PVC's. Lastly, if the PVC checks out it could be a higher layer problem. OSPF could be flapping due to misconfigurations or low router resources.

wgranada1 Fri, 01/11/2008 - 08:23
User Badges:

Thanks guys for your input and keegan I was thinking the same as you that problem on the main T1 would affect everyone else as well. But I finally checked into it and verified that it is indeed the case of the Frame T1 being over utilized/congested. I ran a traffic analysis on the Frame T1 and noticed that it spike to/and a little bit over the 1.5m. I checked those times it spike and compared them to the outages of pvc 23 and 38. Guess what I found everytime the Frame T1 went up to 1.5m there was an outage on 23 and 38 and when traffic went back to normal connectivity restorted. Thinking that this is all a fluke I went back to other outages on 23 and 38 and sure enough I saw the same thing Frame T1 shot up to 1.5/pvc 23 and 38 experienced an outage. So I need to upgrade this Frame T1 so it can handle the traffic.

But I appreciate all your suggestions!!!! Thank you both for the input!!

Actions

This Discussion