2

We resolved an issue caused by a broken access point by simply disconnecting it from the LAN.

The symptoms we experienced were that now and then, our whole network would get "stuck" for a few seconds. That meant that during web calls, the sound and video would shortly freeze, then continue normally, and websites would load with a short delay and very rarely timeout, then on refresh would load normally. Both cable- and WiFi-connected clients, never showed any issues with the connection, but both experienced the issues the same way.

The device itself and all the monitoring tools connected to it were not showing any problems and all health checks were green.

Is there a generic, vendor-agnostic way to detect faulty devices in a network that could cause these problems?

Improve this question
2
  • maybe it is L7 issue, if all devices are facing the same issue at the same time you thats mean you have to check what is in common an all devices, maybe DNS if all clients use the same DNS and the DNS is facing some delays that definitely will impact all you clients, maybe can be a HW issue all clientes are on the same SW or use the same Core? you have to review from L1 to L7 to see what is happening
    – Roid
    Nov 13 at 14:42
  • the DNS is not the issue if the network freezes during a teams call. still, how to even analyze this problem? is there some kind of tool to sniff the network for potential L1 to L7 issues?
    – Theo
    Nov 13 at 16:44

1 Answer 1

Reset to default
4

A broken device can sometime cause a network loop - received packets are just reflected back - leading to broadcast storming. In turn, these storming can deteriorate the entire network performance or even bring down some "weaker" devices due to the sheer packet rate.

I saw something similar with a broken VoIP phone which was stuck booting - and reflecting backup packets. As soon some broadcast packets (ie: ARP discovery/who-is) arrived, it reflected back the very same broadcast packets to the network switch, which againg forwarded them in a loop.

To discover such issues, you can monitor your switch CPU utilization and/or logs. To block them, you should activate your switch loop detection features (STP and/or per-port loop detection/protection).

Improve this answer
1
  • after looking it up, STP was already enabled in my case but the network freezes still occurred. so far the only solution was to simply disconnect each device and see whether this fixes the issue.
    – Theo
    Nov 15 at 13:14

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .