Catalyst 2960s: wrong value in ifHCOutOctets of 10Gbps ports

[igloo_kem]

Hello!
Sometimes my snmp monitorings gets strange value from ifHCOutOctets of 10Gbps ports (it happens randomly).
Such value makes calculation of bps on port like through port gone away 150Gbps..
Calculation of bps (time between checks - 60s): evaluate value as (value-prev_value)/(time-prev_time), where
value - current value
value_prev - previously received value
time - current timestamp
prev_time - timestamp of previous value
Value after calculation always ~150Gbps  (it looks like counter's value was increased onto very big and unreal number).
Switch: WS-C2960S-48TD-L with 15.2(1)E IOS (previous 15.0.2-SE5 had the same behaviour).

Is it known and resolved problem? I had try to find info about such problem, but without results.

[Vinod Arya]

It can be one of the issue:

>> Ideally devices takes longer than 10 seconds to update the SNMP ifTable/ifXTable objects. If  the traffic related SNMP MIB objects of ifTable/ifXTable are polled aggressively, then the counters would not be updated, or sometime it may show incorrect details as well.

This should not affect most customers as it is considered one of the best practices to fetch the data over 3-5 minute poling intervals. This is considered a best practice as the data being fetched for performance monitoring and capacity planning, without impacting the performance. 

>> There are certain know bugs on some of the platforms/IOS on the 10Gig ports, like :

CSCsd65958    Incorrect value for ifHCInOctets, ifHCOutOctets, ifHCInUcastPkts 

CSCsw29555    ifHCInOctets and ifHCOutOctets increase abnormally 

You can open a TAC case to confirm the bug ID you may be hitting and fix IOS versions.

-Thanks

Vinod

**Encourage Contributors. RATE Them.**

[igloo_kem]

I don't have access to TAC.

Anyway, thank you for your answer!