Interface Line Protocol Flapping withtin LACP/Pagp Etherchannel

[moritz.orphall]

Hello there,

suddenly some TenGigabitEthernet interfaces on an uplink of an access switch (C9200l-48P-4X ver. 17.3.4b) begin to flap, which means that the line protocol goes down and up again after approx. 2 seconds. The uplink leads towards an distrubition router (C9500). Those interfaces operate in an etherchannel of 2 links.

When removing interfaces from the etherchannel and let them operate as standalone, the flapping stops immediately. Thats why I would exclude a layer 1 problem like a faulty patch cable, bad gbic etc.... 

We have almost no input errors/CRC on counted on the concerning interfaces. 

The log does not give any more information about possible causes:

021047: Jun  7 09:13:14 CEST: %LINEPROTO-5-UPDOWN: Line protocol on Interface TenGigabitEthernet1/0/19, changed state to down

021048: Jun  7 09:13:16 CEST: %LINEPROTO-5-UPDOWN: Line protocol on Interface TenGigabitEthernet1/0/19, changed state to up

021049: Jun  7 09:13:56 CEST: %LINEPROTO-5-UPDOWN: Line protocol on Interface TenGigabitEthernet1/0/19, changed state to down

021050: Jun  7 09:13:57 CEST: %LINEPROTO-5-UPDOWN: Line protocol on Interface TenGigabitEthernet1/0/19, changed state to up

I would be thankful for every help. 

[Mark Elsen]

   - Can you check logs on the  distribution router too. Also check interface  counters, CRC  on that device equally.

 M.

[moritz.orphall]

Logs on the distri router are equal to the ones on the access switch. Just a flapping line protocol on the interfaces.

About 100 input errors a day can be find on the interfaces, which should not cause the links to go down. 

[Mark Elsen]

  - Could you try changing  , or use another then the current or default load balancing algorithm for lacp , check if that can help.

 M.

[MHM Cisco World]

check the show interface status,
when it down do you see err-disable ?

[moritz.orphall]

The Interfaces does not become err-disabeld, only line protocol gets down and after 2 seconds up again. 

[MHM Cisco World]

there are two end of the port channel, both end check 
show interface status 
check the err-disable, 

[Flavio Miranda]

Hi

  Logs does not say much about the problem. What I suggest is for you enable debug on the 9200 side with the command:

debug lacp  all 

debug pagp  all

Hopefully this can give more information about the problem.

[balaji.bandi]

is this stack ? what is the other side error ?

Can you post configuration both the sides ?

show run interface port-chan X

show run interface x/x

show interface x/x

[moritz.orphall]

interface Port-channel4
switchport trunk native vlan x
switchport trunk allowed vlan x-y
switchport mode trunk
switchport nonegotiate

Interface of Access Switch:

 switchport trunk native vlan x
switchport trunk allowed vlan x-y
switchport mode trunk
switchport nonegotiate
auto qos trust dscp
channel-group 6 mode active
service-policy input AutoQos-4.0-Trust-Dscp-Input-Policy
service-policy output AutoQos-4.0-Output-Policy

Interface of Distribution Router:

switchport trunk native vlan x
switchport trunk allowed vlan x-y
switchport mode trunk
switchport nonegotiate
auto qos trust dscp
channel-group 6 mode active
service-policy input AutoQos-4.0-Trust-Dscp-Input-Policy
service-policy output AutoQos-4.0-Output-Policy

[Mark Elsen]

            - Does the problem persist when the QoS-related commands are not used ?

 M.

[moritz.orphall]

I removed the qos-related commands. Lets see if something changes

[moritz.orphall]

Line Protocol is still flapping, i removed the qos statements from the portchannel and the physical interfaces.

[Mark Elsen]

 -  Could you try changing  , or use another then the current or default load balancing algorithm for lacp , check if that can help.

 M.

[moritz.orphall]

Unfortunately i does not help. I have choosen src-dst-ip as load balancing mechanism for the portchannel.