Portchannel issue with ESXI

[grapevine]

We have 3 cisco switches in stack, two are IE-9320-26S2C and one IE-9320-24T4X with firmware 17.18.01(IE9K_IOSXE) . There are two esxi connected to this stack using port channel. One portchannel has ports from switch 1 and switch 3 and the other has ports from switch 2 and switch 3 in the stack. When we reboot one of the switches, let say switch 1, when it gets added to the stack, we lose connectivity to ESXI, ESXI has configured NIC bonding as active active and on the switch side it is channel group mode on. Please advise how to fix this issue.
We could see the mac on the switches but no arp where the layer vlan is created (firewall)

[Enes Simnica]

hello @grapevine. Since ESXi is running active/active, but the switch side is using a standard LACP/static port-channel, the hashing/forwarding can break during a member reload, which is why you see MACs but no ARPs getting through.

try some stuff like: 

• On ESXi, set the NIC teaming policy to Route based on IP hash (needed for true port-channel).
• On the switch side, make sure the port-channel is running LACP (mode active) and not just on.
• Double-check stack consistency and that all member switches have the same IOS XE version/config.

and check this: Layer 2 Configuration Guide, Cisco IOS XE Cupertino 17.7.x (Catalyst 9300 Switches) - Configuring EtherChannels [Support] - Cisco

-Enes
CCNP x2 Enterprise
Your Friendly Networking Ninja

[grapevine]

On ESXi,  the NIC teaming policy is set to Route based on IP hash.

We have standard vswitch so we cant do LACP

All switches have the same IOS XE version/config

[M02@rt37]

Hello @grapevine 

please share port-channel configuration (Po + interfaces) on your switches. If you have "mode on" please try "mode active".

On ESXi side, you have vsphere distributed switch or standard vswitch ?

[grapevine]

On ESXi side, we have a standard vswitch

[MHM Cisco World]

please reply me to your old post 

MHM

[M02@rt37]

OK @grapevine 

Thanks for that clarification 

[MHM Cisco World]

Hope issue is solve now

MHM

[grapevine]

Sorry for the late response, we have added spanning-tree portfast trunk to the etherchannel config, we will test again and let you know if this fixes the issue

[MHM Cisco World]

How you will test it?

This issue always happened?

MHM