EPs are in freeze state

[brlehigh]

Greetings, all,

I have an endpoint that keeps going into a freeze state and I can't figure out what is going on and how to resolve it.

The MAC address and IP addresses in question are physically assigned to the "IPROUTER2" host, but based on the EP Tracker output, it looks as though it's floating over to the PANASAS_VPC periodically.

Here's a basic diagram of how this is connected:

The "IPROUTER2" is a physical server with a pair of NICs.  The "smc3548" that's connected via the PANASAS_VPC is an L2 switch with a few hosts hanging off of it.  The PANASAS_VPC is configured as access (untagged).  Not trunking any VLANs.  

This should be the relevant info from the BD for this EPG:

I'm not sure that these EP freeze events are negatively impacting service in any way, but they continue to raise faults.

Any recommendations would be greatly appreciated.

[Sergiu.Daniluk]

Hi @brlehigh 

The EPs are in freeze state most likely because of the multiple relearns and moves between the two vPCs. If no moves are noticed in 10 minutes, the EPs should un-freeze.

I would concentrate my efforts in finding out why the IP/MAC was learned from PANASAS_VPC - maybe some network/STP changes.

Stay safe,

Sergiu

[brlehigh]

Thank you for the reply, @Sergiu.Daniluk .  Sorry for the delayed response.

I don't believe there are any STP / topology events occurring.  All the ports on the smc3548 switch (PANASAS_VPC) are up and stable.  Fortunately, these freeze events don't seem to be impacting connectivity.  

Since most of the state transition events look to be coming from the PANASAS_VPC, I'm curious if it might have something to do with it being a layer 2 switch?  

Do my BD settings look correct?  

Do you have any troubleshooting recommendations that I could perform within the fabric?

Thanks

[MaulaShaikh59472]

Kindly try to enable GARP in BD which will clear the mac address once the EPG are frequently moving between DC.