ARP Ageing Issue


I am having a strange issue with my NSX-V deployment.

Let me explain it briefly...

I am creating a `VM` and putting it to my `Logical Switch A` with IP address `` and mac address is ee:e1:e4:1c:16:e2.

Now this VM is accessible [pingable] from other VMs in the same or other logical switches.

And then I am killing this VM and creating new one with same IP and connecting it to same Logical Switch but mac address is changing on each VM deployment as expected.

Now new mac address is 46:61:df:0c:2f:49. IP is same, again.

But VM is inaccessible unless I wait 10 mins and then it starts working.

I am just guessing, this is related to `ARP Ageing`

Running `show tunables detailed` over NSX API returns

Desired states of tunables:

|| Tunable Name                             | Value  ||


| L2 Mac Ageing (Expire Time)              | 900    |

| L2 ARP Ageing (Expire Time)              | 300    |

| L3 ARP Ageing (Expire Time)              | 600    |

| L3 ARP Retry (Probe Interval)            | 30     |

| L3 ARP Preempt Timer (Unicast Probing)   | 450    |

| L3 rtDeferFlushTimer                     | 300    |

| Host event notification resend interval  | 30     |

| Host event notification status           | true   |

| L2 MAC Table Size                        | 2048   |

| L2 ARP Table Size                        | 512    |

However, I am unable to set anything below than default values.

How can I disable ARP caching mechanism or `flush` it?

My environment is kinda between dev and staging so I have to kill VMs and create them all the day [using PowerCLI]

Any help would be much appreciated.


0 Kudos
0 Replies