1

I encounter a network problem. My clients can ping the Oracle ERP server but experience frequently session timeout when access to the ERP application. I have configured two port channel to the oracle switches and the two port-channel are configured in the same setting. I have found that message in the cisco siwtch log,%SW_MATM-4-MACFLAP_NOTIF: Host 0010.e05c.ac14 in vlan 3 is flapping between port Po4 and port Po5 (core switch)

Network diagram

Running config on cisco core switch:

interface 

Port-channel4

description To Oracle Switch 2

switchport mode trunk

!

interface 

Port-channel5

description To Oacle Switch 1

switchport mode trunk

!

interface GigabitEthernet1/0/11

 description To Oracle switch 1 

 switchport mode trunk

 channel-group 5 mode active

!

interface GigabitEthernet1/0/12

 description To Oracle Switch 2 

 switchport mode trunk

 channel-group 4 mode active

!

interface GigabitEthernet2/0/11

 description To Oracle switch 1 

 switchport mode trunk

 channel-group 5 mode active

!

interface GigabitEthernet2/0/12

 description To Oracle Switch 2 

 switchport mode trunk

 channel-group 4 mode active

!

How can I deal with the mac address flapping problem? I appreciate any suggestions and response:)

Improve this question
1
  • Did any answer help you? If so, you should accept the answer so that the question doesn't keep popping up forever, looking for an answer. Alternatively, you could provide and accept your own answer. – Ron Maupin♦ Aug 11 '17 at 16:19
2

This is entirely due to the host side redundant configuration. In this mode, only one NIC can be in use at a time. If it balances traffic across both NICs using the same MAC, you will have this problem.

(I'm pretty sure your hardware doesn't support Multi-chassis LACP.)

Improve this answer
1
  • As Ricky states, the problem is the host configuration. You need to use a configuration that either connects source servers to outbound interfaces (VMware would use hash based on source-mac) or an active/standby configuration where the end-host does not attempt to send packets up both links. – cpt_fink Apr 15 '15 at 5:05
0

Can you run a show port-channel summary from the cisco switch and see if the port actually came up? You may need to interface port-channel# and then no shutdown (for each interface). If the interface is up, make sure that you're using the same type of aggregation that the oracle switch expects.

From your diagram it's also possible that your cross-connect is to blame. How are you managing failover on the Oracle switches? If all the links are live then that can cause your flapping problem.

Improve this answer
-2

IF this is NOT a nexus switch:

set your port channel to mode on

Port-channel5 mode on

Improve this answer
2
  • You should add some explanation to your answer, and if there is any documentation, you should quote the relevant parts and link to it. – Ron Maupin♦ Dec 12 '15 at 17:10
  • You could improve your answer by editing it to add more details to make it more useful both for the original poster and future users. Typically short answers like this could provide reasoning why you believe this is the answer, more explanation about the concepts mentioned, references/links to supporting resources, or applicable examples. – YLearn♦ Dec 18 '15 at 2:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.