What does "ingress drop packets periodic" mean?

[SIMMN]

I ran into an issue that VM can not be vMotioned between two ESXi hosts (A&B) in two chassis connected to the same pair of leafs via individual vPC (PC1 for A; PC2 for B). They share the same vMotion network that is configured as an EPG. Both ESXi vMotion kernel addresses have been learned properly under EPG. There is no faults presented in APIC and no contract drops... ACI L2 MTU is set as 9000 and the same is used on the compute NIC setting.

I used Visibility & Troubleshooting tool in APIC to check the Endpoint to Endpoint flow, again there is no faults. But under "Drop/Stats", I found drops with reason "ingress drop packets periodic". Screenshot is below for reference...

Can someone please help guide me what does this even mean? Or if you have document regards or regarding the Visibility & Troubleshooting tool itself, please let me know as well.

[Robert Burns]

ACI will raise a fault related to Atomic Counters for traffic forwarded from one VPC peer to another vPC peer. This occurs because the vPC peer receiving the packet thinks the packet originated on the switch locally since it owns the source ip address [VIP] (along with the other vPC peer). These faults are cosmetic and this behavior is described in CSCuz99892
Workaround: Ignore/squelch the fault. The Atomic Counter policy can also be disabled.

Back to your vMotion issue, can you vMotion any other VMs between hosts, either in the same Chassis or to/from a remote Chassis?  Are there any more details in the VMW logs that might point to why it's failing? https://blogs.vmware.com/vsphere/2019/09/troubleshooting-vmotion.html

Robert

[M02@rt37]

Hello @SIMMN 

It is crucial to determine whether the issue is isolated to specific VMs or hosts, or if it's a broader issue affecting vMotion across different parts of your infrastructure.

Test Other VMs:

Same Chassis: Attempt to vMotion other VMs between the same two hosts that are in the same chassis. If these work, the issue might be related to the specific VM or its network configuration.

Remote Chassis: Try vMotioning a VM to or from a different chassis. If this fails too, it suggests a broader issue, possibly with the vMotion network or the ACI fabric.

Also, check the vmkernel.log on the ESXi hosts involved in the vmotion attempt. Look for specific error messages related to vmotion. These logs can give you insights into why vMotion is failing—such as timeouts, network issues, or authentication errors...

Note that vMotion can also fail if there are high latencies or I/O issues on the shared datastore between the hosts. Check the datastore performance metrics too.