Packet InError Detection

Packet InError Detection identifies links that receive more number of bad frames than configured threshold and disables them to avoid instability in the network. For instance, if a network has redundant uplinks, usually only one link is in forwarding state and the rest are redundant and blocked. If one of the redundant links becomes faulty, it may drop the PDUs and become a forwarding link. This can cause loops in the network. Packet InError Detection detects the faults in the link and disables the link to prevent loops in the network.

Packet InError Detection counts an ingress frame that has one or more of the following errors as an inError packet:
  • Alignment error
  • CRC error
  • Oversized frame error
  • Internal received MAC address error (Errors that do not fall in the above 3 types)
  • Symbol error (includes the fragmented, short, or undersized frames)

You can configure the number of inError packets allowed per port in a specified sampling interval. If the port receives more than the configured number of inError packets in two consecutive sampling intervals, then the port becomes error-disabled. The output of the show interface ethernet command for the affected port will show the status of the port as “ERR-DISABLED (packet-inerror)”.

NOTE
It is recommended to use Packet InError Detection only on required ports. If you enable this on a large number of ports in a device and use a very short sampling interval, it may lead to heavy CPU usage.
NOTE
The inError count configured on the primary port of a LAG is inherited by other member ports of the LAG. However, the LAG ports are individually sampled for inError packets. Therefore, inError packets on a port disable only that port and not the entire LAG.
NOTE
Executing commands that clear the packet counters, such as the clear statistics command may interfere with the proper functioning of Packet InError Detection because these commands reset the inError packet count.
Parent topic: Spanning Tree Protocol