Start the Cutover Process to Redirect Agent Traffic

When the Splunk AppDynamics components reach the saturated state, you can initiate the cutover to redirect the agent traffic to Virtual Appliance. Before you begin the cutover phase, verify the migration status.

Redirect Agent Traffic to the Virtual Appliance

Set up a load balancer for your Virtual Appliance to receive agent traffic.

Ensure to complete the following tasks before starting the cutover:

Standard

  1. Configure DNS entries

  2. Disable the Default Certificate Property.

  3. Disable the Events Service SSL Property

Hybrid

  1. Disable hybrid setting in the globals.yaml.gotmpl file.

  2. Configure DNS entries

  3. Disable the Default Certificate Property.

  4. Disable the Events Service SSL Property

  1. Update the Classic On-Premises DNS to point to the Virtual Appliance load balancer. Map the classic on-premises port to VA port as follows
    Classic On-Premises Virtual Appliance
    8181 443
    9080 32105
    7001 80
  2. Restart the Synthetic server in Virtual Appliance.
  3. Shutdown the classic on-premises Controller.
    CODE 
    platform-admin.sh stop-controller-appserver
  4. Start the migration tool and wait until the process reaches saturation.
    CODE 
    nohup python3 ./migration_tool.py migrate all &
    This step transfers the remaining data received during the cutover from the classic on-premises environment to the Virtual Appliance. Once the transfer is complete, verify the migration status.
    Note: If all services reach a saturated state except for the Controller, perform the steps from Step 5.
  5. Verify whether the following logs exist run the commands to ensure the Controller MySQL data is saturated:
    CODE 
    coordinator - controller_mysql_worker_0 - WARNING - No new binlog data to process, waiting...
coordinator - controller_mysql_worker_0 - INFO - Incremental backup returned None (no new data) so skipping incremental backup

coordinator - controller_mysql_worker_1 - WARNING - No SQL, SQL.gz files or MySQL Shell dumps found in /mnt/nfs_controller/ Waiting for 60 seconds...
    1. Run the following command to navigate to the Controller folder:
      CODE 
      cd /mnt/nfs_controller/
    2. Run the grep commands:
      CODE 
      ls -larth | grep inc | wc -l

ls -larth | grep inc | grep restored | wc -l
      If the grep commands result with the same count, it indicates that the data migration for all services is complete.
  6. Stop the Migration Tool.
  7. Update the Resource Permissions for EUM and Synthetic Services.
Synthetic jobs might be missing in the Controller UI after data migration. To troubleshoot this issue, see Missing Synthetic Jobs in the Controller UI After Data Migration.

Once these steps are completed, you can begin using the Virtual Appliance to monitor your applications with the data previously present in Classic On-Premises.

Decommission the NFS server

Monitor the data migration for few days and then decommission the NFS server which is created to do the data migration.

Run these commands on each Virtual Appliance server and classic on-premises servers.

  1. Fetch the mount points.
    CODE 
    mount | grep mnt
sudo umount -f <each_path_from_above_output>
  2. Decommission the NFS server.