NS and VNF Heal Support

NS and VNF Heal is a Life Cycle Management operation. A user can restart a VM which is already in the RUNNING state using the heal feature. If heal fails, then a user can do a heal-retry.

Note: Retry for instantiation, scale-out, scale-in and termination do not impact this feature.

SOL005 and SOL003 standards do not support VDU level healing.

For SOL003, it is not possible to identify if there is an SVNFM account as part of the service. If there is an expectation for ‘additionalParams’, then it is possible to pass those parameters. However, there is no standard and each SVNFM may behave differently.

For SOL005, RIFT.ware will pass VDURs in ‘additionalParams’ to force VDU level healing.

Note: The Heal supports OpenStack and VCD for RIFT VNFM accounts.

See the Or-Vnfm Account (ETSI GS NVF-SOL 003) and Os-Ma-Nfvo Interface (ETSI GS NVF-SOL 005) sections for additional details on healing for SOL003 and SOL005.

A user can deploy Aodh services in OpenStack for RIFT.ware to receive alarms from the VIM. These alarms indicate that an error occurred with the VM. After RIFT.ware receives an alarm, then Launchpad decides if it is necessary to bring down the previously successfully instantiated NS or trigger a heal operation. For details on deploying Aodh service in OpenStack, refer to the OpenStack Aodh Documentation.

Heal

An operator can restart a VM when a Network Service is in the Running state. When a user attempts to Heal the service, the status in the UI shows the VNF-Heal progress. RIFT.ware processes a Heal on each VDU. The heal options are restart vdu and reconfigure vdu.

  • Restart VDU: This heal option attempts to restart a VM on the running NS. This is the default operation.

  • Reconfigure VDU: The heal option means that only pre-heal primitives are run on the service. Before heal, pre_heal_service_primitive will run (not applicable for SOL003).

During the NSR Heal operation, the status can either change to Running (successful) or Failed. If one of the VNFRs in the service changes to the running state, then the NSR status can change to Running depending on how many other VNFRs are pending.

SOL003 supports both manual and auto-heal.

Heal-Retry

If the first heal attempt fails, then an operator can try to restart the VM again. This action can only be taken when the NSR is in the Failed state. A retry can also be made when scale instance state is in the failed-temp state. When a user attempts to retry the heal, the status in the UI shows the VNF-Heal progress. RIFT.ware processes a Heal on each VDU. The heal options are restart and no operation.

  • Restart VDU: This heal option attempts to restart a VM on the running NS. This is the default operation.

  • Reconfigure VDU: This heal option means that only pre-heal primitives are run on the service. Before heal, pre_heal_service_primitive will run (not applicable for SOL003). If you choose this option, you are reconfiguring the VDU. Reconfigure VDU does not restart the VM.

During the NSR Heal operation, the status can either change to Running (successful) or Failed. If one of the VNFRs in the service changes to the running state, then the NSR status can change to Running depending on how many other VNFRs are pending.

Heal attempt in the UI

There is now a Heal button in the SERVICES section in the UI. This option allows users to heal a VM.

  1. On the Launchpad menu, click SERVICES.

  2. Click the button next to a specific service to start the heal process. If you open a running NS, then click the button at the top of the screen.

  3. Choose Restart VDU or Reconfigure VDU in the SELECTION OPERATION drop down menu to attempt heal.

    • RESTART VDU – This heal option attempts to restart a VM on the running NS. This is the default operation.

    • RECONFIGURE VDU – This heal option means that only pre-heal primitives are run on the service. Before heal, pre_heal_service_primitive will run (not applicable for SOL003). If you choose this option, you are reconfiguring the VDU. This reconfigure option does not restart the VM.

  4. Click on xpath in the SELECT VDURS field and the possible vdurs appear.

  5. Choose the appropriate vdurs and click HEAL.

  6. During the heal attempt process, the Services tab shows the progress state – Heal in progress.

    Note: After performing a heal action, the heal status appears in the History tab. See Monitor the History of a Service for more details about a specific event.

  7. If the heal attempt fails, then the state changes to Vnf Heal failed.

  8. If the heal attempt fails, then an error message appears.

    If your service is in the failed-temp - configured or heal-in-progress - configured status, then the VNF-Heal progress view will not appear.

  9. Click and a Last Operation: heal-nsr dialog box appears.

  10. Choose retry and click EXECUTE to attempt the heal again.

    Note: There is no rollback option for a heal operation.

    The VNF-Heal state will no longer appear when the NS instantiation is successful.

    The heal state appears on the Services > Viewport screen. You can also check the Event Viewer to see more information about the success or failure of the heal attempt.

    If a VNF heal attempt fails, then see Retry and Rollback Support Improvements to recover using a scale out instance.

See also

Launchpad Viewport

Monitor the Network Through the Viewport