Cisco Catalyst 2960x lacp Master reelection port disabled temporaily

[ashley_dew]

Hello,

i have a pair of Cisco Catalyst 2960x as top of rack switch connected to Core Switch 3850x via an etherchannel lacp.

During a power redundancy exercise with the master cisco catalyst 2960x was powered down.

During this moment, the etherchannel on the core switch went down with the interface on the other cisco catalyst 2960x went down temporarily.

When we checked the logs of the second 2960x member, it was found that after the master election the active cisco catalyst 2960x disabled all lacp members for about 5-6 seconds before reactivating the interfaces. This explains that the etherchannel went down completely.

My understanding is lacp and stack should be transparent but we had a complete outage for a few seconds.

from what i read this is not stack mac persistence related as it did not see any lacp negotiation errors with lacp master id change and so do not have any stp instance running on Core and top of rack.

i need your help to understand if this is normal or a means to minimise this complete outage.

Thanks,

Ashley

[balaji.bandi]

You need to provide the config of both devices, what code running, and what Logs you see on that time ? (if you have some)

is the Down, or disabled (error) ?

no port-channel standalone-disable

https://www.cisco.com/c/en/us/td/docs/switches/lan/catalyst4500/12-2/15-02SG/configuration/guide/config/channel.html#wp1076851

some troubleshoot tips :

https://www.cisco.com/c/en/us/support/docs/lan-switching/etherchannel/19642-126.html#po1a

[ashley_dew]

Hello,

The

interface goes down.

Switch2 was master on the stack of 2. Switch 2 was powered down

Etherchannel is made up of ports g1/0/52 and g2/0/52. Both were up before.

I masked some information but the concern is marked in Red. So it is not error disabled, the active interface is shutdown and up again.

Version for Catalyst 2960x:

C2960X Software (C2960X-UNIVERSALK9-M), Version 15.2(2)E5, RELEASE SOFTWARE (fc2)

Config on C2960X

Active mode on both ports and on the other side C3850.

Config on Core sw

bOTH LCAP active on both ends, so normally no problem

Log on Core SW:

Same logs been reflected on the Core, the active member goes down and then up and etherchannel is up.

[balaji.bandi]

One major issue i see is the uptime of the device 5+ years or port-channel, I am guessing you have a config issue on a 2960 stack of switches? or could be bug of the version of code running.

can you post from 2960

show version

show switch

Note: when all working, switch2 is active as master as per the logs, when switch2 is down, Switch 1 becomes master, and not sure why this is admin down. 

On another note, you mentioned on your not running any STP.

[ashley_dew]

Hello,

Thanks for your reply.

Here is the show switch below:

sh ver file is attached.

You are right the output of the etherchannel age is contradictory with the logs and the admin shut  from the logs was done during the switching process of master and we confirm no issuing the command.

It is weird. Looks more like a bug, however the impact is real. The etherchannel went down temporarily at both ends and do not see any lacp negociation errors leading to port going down on the Core switch.

[balaji.bandi]

Not sure at this stage - you may be hitting BUG :

when you get chance  - upgrade to  15.2.7E7 and monitor

[MHM Cisco World]

there is something in your PO config, 
I think the mode in both side, are you run 
ON/active or active/active or passive/active