cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
1518
Views
0
Helpful
7
Replies

It seems a bug in 12.2(33).SXJ1

piskorsky
Level 1
Level 1

I see the twice counted outbound traffic on subinterfaces that has ISL encapsulation on my SUP720-3BXL with s72033-advipservicesk9_wan-mz.122-33.SXJ1.bin. I haven't seen this problem with s72033-advipservicesk9_wan-mz.122-18.SXF17.bin

Don't hit me for ISL, it is worked by historical reasons. I have already change ISL to 802.1q - and it seems look like a workaround. I can't find this problem in Bug Search Tool, so if you can please drop me a bug id.

Thanks for any replays.

1 Accepted Solution

Accepted Solutions

is it possible that the traffic is really going twice via your ISL subinterface? for some reason after updating?

View solution in original post

7 Replies 7

nkarpysh
Cisco Employee
Cisco Employee

Hi George,

Can you please paste the outputs you refer to and advise on how you understood that those are not correct.

Thanks

Nik

HTH,
Niko

Hi Nikolay,

It is very simple. I have tool, named Cricket, that based on RRD Tool and collect every 5 minutes SNMP-data from my network devices. So this tool collects ifHCInOctets and ifHCOutOctets and make picture from collected data. After I have changed IOS from s72033-advipservicesk9_wan-mz.122-18.SXF17.bin to s72033-advipservicesk9_wan-mz.122-33.SXJ1.bin I have seen that graphics that based on ifHCOutOctets on all of my ISL encapsulated interfaces on SUP720 and X6748-SFP cards have been counted twice.

Another conformation of the twice counted traffic in SNMP-data that all traffic that passed trouth ISL subinterface and than passed trourh 3750 not counted twice on 3750 interface.

And another one conformation is that after I have changed encapsulation type from ISL to 802.1q - all SNMP counters become right.

is it possible that the traffic is really going twice via your ISL subinterface? for some reason after updating?

I can't imagine how it could happen in my configuration... It is subinterface with /30 interface network and static route to neighbor. So I think it is not possible.

interface GigabitEthernet1/2

no ip address

ip flow ingress

load-interval 30

media-type rj45

snmp ifindex persist

interface GigabitEthernet1/2.100

encapsulation dot1Q 100

ip address 1.1.1.1 255.255.255.252

no ip redirects

ip flow ingress

P.S. Sorry, I have press "right answer" inadvertently.

piskorsky
Level 1
Level 1

About 2 hours ago I have tested the same IOS version on SUP32-GE-3B, and have seen the same problem with ifHCOutOctets counter on ISL subinterfaces. I have changed ISL to 802.1q and counter become right.

Hi George,

it would be a really strange bug, have you informed Cisco TAC allready?

Can you check the counters directly on the interface?

Have you tried to use a normal SNMP object ifOutOctets, just for testing?  Does it show the same result?

darshan99
Level 1
Level 1

IOS 12.2(33)SXJ is build to support WiSM2 cards, this is the minimum requirement to support WiSM2 cards.

In this firmware it wont allow to configure on individual gigabitethernet.

It requires autoconfigurations. IT DOES NOT COUNTAIN BUG THAT IS NOT ALLOWING TO CONFIGURE GIGABITETHERNET PORTS.

In order to configure for WiSM1 cards

wism (module) controller (number) native-vlan (---)

wism (module) controller (number) allowed-vlans (all those are allowed on supervisor).

It will solve issue configuring WiSM1 cards on 12.2(33)SXJ or 12.2(33)SXJ1

Review Cisco Networking products for a $25 gift card