CSS REDUNDANCY DO NOT WORKS

[abelleli71]

We experience in a  crash occurred on one of the switches 6K .

The Master CSS  which is physically connected to the involved Core switch, changed its status to the Backup.

The Backup CSS which is physically connected to the other Core switch, that was working fine and had no problems, changed its status to the Master one, but was not able activate correctly the uplink to the main network.

Here some output of the CSS backup

Thanks for any helps

CSS_B# sh log
Displaying Log events.

Press any key to abort...

CSS_B#
CSS_B# sh interface Ethernet-Mgmt
Name             ifIndex    Type      Oper  Admin     Last Change
----             -------    ----      ----  -----     -----------
Ethernet-Mgmt       2       fe        Up    Enable    11/09/2010 20:31:39

CSS_B#
CSS_B#     sh critical-reporters

No Critical-Reporters

CSS_B#
CSS_B#     sh critical-services

No Critical-Services

CSS_B#
CSS_B# sh redundancy
Redundancy:               Enabled    Redundancy Protocol:      Running  
Redundancy State:         Master     MasterMode:               No       
Number of times redundancy state changed to Master:            1        
                                         to Backup:            1        
Redundancy interface:     172.7.X.X
Current State Duration:   0 days 01:51:56
Last Fail Reason:         No Fail   
VRID:                     128        Priority:                 100       
Uplink Enabled:           1          Number Alive              1         
Service Name:             Service State:           
-------------             --------------           
uplink1                   Alive 

[Gilles Dufour]

What do you expect us to do with this info ?

You showed the management interface which is not involved in any loadbalancing action.

And the redundancy command shows that the box is active with the uplink alive.  So everything looks good.

You should have checked the arp table on upstream catalyst.

See if the CSS vip associated mac-address points to the right CSS.

When there is failover a G-Arp is normally sent to update L2 table on all surrounding devices.

If this packet was lost, the catalyst will continue to send traffic to the primary CSS which is now backup.

Check my blog first entry

Gilles.

[abelleli71]

Hi

maybe the trap in attach could help ..

What happen

1. The core switch uplink where is attached the CSS flapping UP& down many times

2. When finish to flap The CSS did noth change his election

The core switch send this message

Nov  9 19:40:52.829 MET: %LINEPROTO-SP-5-UPDOWN: Line protocol on Interface GigabitEthernet2/46, changed state to up
000875: Nov  9 19:41:50.393 MET: %CSM_SLB-4-REDUNDANCY_WARN: Module 3 FT warning: FT configuration back in sync