Migration of Platform nodes in cluster to different Data Center where IP and Network Change is needed

Products

• VCF Operations for Networks

Issue/Introduction

• Migration  of Platform nodes to different Data Center where IP Change is needed
• For this activity the entire Network stack would need change.
• Platform Nodes are in Clustered deployment.

Environment

• VCF Operations for Networks 6.13.0
• VCF Operations for Networks 6.14.0
• VCF Operations for Networks 6.14.1

Resolution

1. On the Source Data center/vCenter where the 5 Node Cluster is currently sitting, execute steps using the script mentioned in KB https://knowledge.broadcom.com/external/article/314428 to ensure the Cluster shut down procedure is followed and snapshots are taken.

   
    From the above mentioned KB make sure you follow the procedure mentioned in Step 4 Steps a to j.

 

2. Post snapshots are taken, move the 5 Node platform VMs to the Destination Datacenter and manually power on the moved 5 node platform cluster in order 1 to N.

 

3. Once all nodes are powered on you need to take SSH session on each node and manually make changes to /etc/ntp.conf  to update the IP address of the NTP which you wish to use on the destination datacenter.

   

   The command to use used is as below:

sudo vim /etc/ntp.conf

Use i key on keyboard and update the correct IP address of the NTP and save it by pressing escape key and then entering :wq!

This needs to be done on all the rest platform nodes one by one.

 

4. Once Step 3 is completed on all the 5 platform nodes then proceed with step K and use the script in above KB and run the below command to start-services

./vrni-cluster-shutdown-script.sh start-services 127.0.0.1 "/home/ubuntu/vrni-cluster-shutdown-script$(date +%s).log"

 Make sure above is executed on Platform Node 1 only using SSH/putty with username support.

 Running above command will unmask NTP service when executed on all platform nodes and will allow to ensure NTP service is running with the IP updated in Step 3 above. 

 

5. Wait for NTP to show state IN-SYNC on all 5 platform nodes.

 

6. Now follow the steps mentioned in documentation below:

 Clustering - General- How to change IP of any platform node or a collector in a cluster

 

Note:

This needs to be done manually on all the platform nodes in Cluster in order 1 to 5. Execute steps a to e only.

You need to take a SSH session on Platform Nodes using consoleuser and execute the steps in above given documentation.

Ensure the above procedure is followed and correct IP Address for the Cluster nodes is update along with other  networks details.
Entering incorrect information can result in incorrect entries of the IP address in Aria Operations for Networks database ( Foundation DB).

   

7. Once IP address and other networks details are submitted correctly, see step 8.

 

8. Pair all the collector to the platform node1 using below command. This needs to be executed using username consoleuser

pair-collector set-platform --ip-or-fqdn <New_Platform1_IP>

Please Note: Only platform node1 IP is needed to be pair on all the collector nodes, rest platform nodes IP in the deployed Cluster are not needed to be be paired.

   

9. Validate the all 5 Node Cluster services shows running and healthy using below command on Platform Node1 using SSH/putty using username support

./run_all.sh sudo /home/ubuntu/check-service-health.sh -p -d

10. Validate the all collector (s) in questions shows all services and health using below command

./check-service-health.sh -p -d

11. Repeat Step 10 for rest other collector(s) in the deployment to ensure the services are running and healthy.

 

12. Validate by login to VCF Operations for Networks UI and navigate to Settings>Infrastructure and support>Infrastructure  and updates page.

 

13. Validate if all is looking good on this page as this page will show you the New IP address for the 5 Node platform VMs and the collector which you have paired with the collector.

 

14. Once you have validated that all is looking good then make sure that snapshots taken in step 1 are now deleted from all the platform and collector nodes.