VMware NSX

 View Only
  • 1.  NSX Edge Node Deletion.

    Posted Nov 02, 2023 09:59 AM

    HI All,


    We recently faced a storage failure in one of our secondary sites, which resulted into NSX edge node being completely deleted (the edge VMs no longer exists). As a result we had to redploy the edge node and reconfigure the NSX on the hosts. For the task, we removed previously used T1, T0 and were trying to remove the failed Edge nodes but when we initiated the edge node deletion, it could not be deleted and the status is set to Deletion Failed.

    We did try the KB https://kb.vmware.com/s/article/89283 but the api call does not do anything, the Failed Edge nodes still exists.

    I will be very grateful for any suggestions.

    NSX-T Version:

    Thank You.

  • 2.  RE: NSX Edge Node Deletion.

    Posted Nov 10, 2023 09:50 AM

    You deleted T0 and T1, and lose an Edge?

    Why don't you redeploy NSX enviroments? It's more quickly than troubleshooting.

  • 3.  RE: NSX Edge Node Deletion.
    Best Answer

    Broadcom Employee
    Posted Nov 17, 2023 10:54 AM


    I can suggest you to remove the stale entries from the NSXT backend corfu database which is usually done with help of VMware support and we dont suggest it to be done by our customers. But prior to this you would need to collect the support bundles of all the nsxt manager nodes and also take a backup of the NSXT cluster following the official documentation: https://docs.vmware.com/en/VMware-NSX-T-Data-Center/3.0/administration/GUID-A0B3667C-FB7D-413F-816D-019BFAD81AC5.html

    Below are some steps you could follow, but please involve VMware support for these operations and advised not to attempt on your own. You could do the following as below:-

    1. Check for presence of ReplacementInfo corresponding to the Edge nodes that fail to delete in the below corfu table:  a)/opt/vmware/bin/corfu_tool_runner.py -o showTable -n nsx -t ReplacementInfo. b) Copy the key value corresponding to deleted edges.

    2. Determine edge nodes with "markedForDelete": true in the exported table EdgeTransportNode using
    /opt/vmware/bin/corfu_tool_runner.py -o showTable -n nsx -t EdgeTransportNode.

    3. Stop all services: service proton stop; service corfu-server stop. Start only Corfu: service corfu-server start

    4. Delete stale records where table contains key in format Edge path ending with UUID. Table names are mentioned in bold below:
    /opt/vmware/bin/corfu_tool_runner.py -o deleteRecord -n nsx -t ReplacementInfo --keyToDelete '{"stringId": "/infra/sites/default/enforcement-points/default/edge-transport-node/<edge-id>"}'
    /opt/vmware/bin/corfu_tool_runner.py -o deleteRecord -n nsx -t EdgeNodeExternalConfig --keyToDelete '{"stringId": "/infra/sites/default/enforcement-points/default/edge-transport-node/<edge-id>"}'
    /opt/vmware/bin/corfu_tool_runner.py -o deleteRecord -n nsx -t EdgeNodeInstallInfo --keyToDelete '{"stringId": "/infra/sites/default/enforcement-points/default/edge-transport-node/<edge-id>"}'
    /opt/vmware/bin/corfu_tool_runner.py -o deleteRecord -n nsx -t EdgeNodeConfigInfo --keyToDelete '{"stringId": "/infra/sites/default/enforcement-points/default/edge-transport-node/<edge-id>"}'
    /opt/vmware/bin/corfu_tool_runner.py -o deleteRecord -n nsx -t GenericPolicyRealizedResource --keyToDelete '{"stringId": "/infra/realized-state/enforcement-points/default/edge-transport-node/<edge-id>-edgetn"}'
    /opt/vmware/bin/corfu_tool_runner.py -o deleteRecord -n nsx -t EdgeTransportNode --keyToDelete '{"stringId": "/infra/sites/default/enforcement-points/default/edge-transport-node/<edge-id>"}'
    /opt/vmware/bin/corfu_tool_runner.py -o deleteRecord -n nsx -t DeletedVm --keyToDelete '{"stringId": "/infra/sites/default/enforcement-points/default/edge-transport-node/<edge-id>"}'

    This is not a full list of the steps but some of the things that could be done, but not without VMware support involved.