NSX CROSS VC NOT WORKING

[Swadeey]

Hi Gurus,

I  have been struggling to  get my  universal logical switch working , what i have  done is the following:

1. The universal switches created from the primary NSX manager  have been propagated  to secondary NSX manager .

2. After attaching the VM  on primary site to the universal switch called USAPP  and the  VM on secondary site to the same universal switch. The virtual machines are unable to ping each other  but the   virtual machines on the same site are able to ping each other.

3. Universal Transport Zone, Universal Segment ID  has been created as well.

4  VxVLAN  has been configured  , clusters in the respective site attached

Please can someone guide on where to start troubleshooting or to fix the issue .

Regards

Swapnil

[Swadeey]

Please can someone help.

[yak9]

Are VTEPs on site A reachable by VTEPs on site B?

What is MTU setting on site interconnect?

You can run checks with following command from site A host:

ping ++netstack=vxlan <site B host VTEP IP address> -s 1572 -d

[Swadeey]

Thanks for your response.

Could you please explain what do you mean by MTU setting on site interconnect?

Regards

Swapnil

[Matthew Larson]

The recommended MTU size for VXLAN traffic is 1600.  All devices through with VXLAN traffic flows, from VTEP to VTEP, should have an MTU of 1600 or greater.

This link proivdes a way to see if connectivity exists among VTEPs from within the logical switch.  It does the same command string that yak9 provided, except throught the GUI, and saves you from having to go from host to host to check connectivity.

NSX 6 Documentation Center

Also, try restarting the netcpa service on the hosts experiencing issues.

[rikherlaar]

and verify the /var/log/netcpa.log - clearly VTEP comms is not okay...

/r

[dverma20]

Just wondering, are you using hardware based VTEPs?

[GuyManDude]

Rather than starting a new thread I'm going to use this old one because I need to understand the correct answer provided to it.  I have Cross VC setup but hosts on the same Logical Switch cannot ping each other.  The VTEP's on site A are 20.x.x.x/8 and 30.x.x.x/8 on site B.  The correct answer suggested those are supposed to be able to ping each other but how would you accomplish that without some kind of router?

Thanks for your thoughts.

Original post and correct answer below for reference.

==========================================================================================

Hi Gurus,

I  have been struggling to  get my  universal logical switch working , what i have  done is the following:

1. The universal switches created from the primary NSX manager  have been propagated  to secondary NSX manager .

2. After attaching the VM  on primary site to the universal switch called USAPP  and the  VM on secondary site to the same universal switch. The virtual machines are unable to ping each other  but the   virtual machines on the same site are able to ping each other.

3. Universal Transport Zone, Universal Segment ID  has been created as well.

4  VxVLAN  has been configured  , clusters in the respective site attached

Please can someone guide on where to start troubleshooting or to fix the issue .

Regards

Swapnil

Correct Answerby yak9 on Jun 23, 2016 5:32 AM

Are VTEPs on site A reachable by VTEPs on site B?

What is MTU setting on site interconnect?

You can run checks with following command from site A host:

ping ++netstack=vxlan <site B host VTEP IP address> -s 1572 -d

[GuyManDude]

I added a router between them and it solved the problem.