|
Some users observe RX Discard packet count increases over time.
This depends on port configurations. Port51 and port52 are both members of ICL MCLAG trunk, meaning the Rx Discard count will have large numbers.
edit "_FlInK1_ICL0_"
set mode lacp-active
set auto-isl 1
set mclag-icl enable
set members "port51" "port52"
next
port51 Port Stats:
Rx Bytes: 3859959205219
Rx Packets: 4102187506
Rx Unicasts: 4091427311
Rx NUnicasts: 10760195
Rx Multicasts: 5318619
Rx Broadcasts: 5441581
Rx Discards: 4099935 <==== Packet drop
port52 Port Stats:
Rx Bytes: 1900858591287
Rx Packets: 2419162739
Rx Unicasts: 2405248197
Rx NUnicasts: 13914542
Rx Multicasts: 6741143
Rx Broadcasts: 7173398
Rx Discards: 47149 <- Packet drop.
Due to FortiSwitch software design, FortiSwitch will include lots of legitimate packet discards, such as ICL egress blocking, STP blocking, FortiLink protocol packet blocking, ACL blocking, etc. in Rx/Tx Discard count. As a result, it will trigger a false alarm in a network monitoring tool or cause concern for network administrators.
It is normal to observe less than 1% of the total packet process count as a value on RX packet drops. It is because these packets are processed by CPU instead of being transmitted back to the network.
The amount of packet discards will be very dependent on the port configuration, but it is expected that uplink ports and other ports connected to third-party devices using those protocols will display this behavior.
It is recommended to adjust the threshold baseline on monitoring tools accordingly or just monitor CRC errors instead with the following commands, in case any potential physical damage, the Input and Output error will be increasing, on this case is time to replace the SFP, change the ethernet cable, or use different Switch ports:
diagnose switch physical-ports set-counter-zero portNumber
diagnose switch physical-ports list portNumber
|
