Troubleshooting during the migration process.

Logs

You can find logs in /migrationVolume/logs/app.log file for troubleshooting. Logs are rotated on 10 MB and 20 log files are kept in the log location.

Timeouts

You can control the timeouts for each of the phases which includes upload template, deploy VM, and thread_pool size by changing values in conf/app.conf in the migration container.

If there is any timeout or slow environment for example, ovf is taking time to upload, then you can modify the timeouts and retry the CLI commands.

[App]

DEPLOY_PHASE_TIMEOUT_IN_SEC = 1800
VM_BACKUP_PHASE_TIMEOUT_IN_SEC = 1800
MIGRATE_PHASE_TIMEOUT_IN_SEC = 3600
REVERT_PHASE_TIMEOUT_IN_SEC = 1800 

OVF_TOOL_DEPLOY_TIMEOUT_IN_SEC = 1800 
OVF_TOOL_UPLOAD_TIMEOUT_IN_SEC = 1800 
MAX_THREAD_POOL_SIZE = 6

Windows

  1. To test the Migration in windows 2019 server edition:

    1. Enable containers by navigating to ServerManager > Manage > Add Roles and Features > Server Roles > Feautres.

    2. Enable Hyper-V by navigating to ServerManager > Manage > Add Roles and Features > Server Roles.

  2. Install Docker as per DockerversionRequirements.

    1. After installation, log out and log in.

    2. Perform migration steps as mentioned above.

  3. Command window stops responding during the migration process. However, the migration process continues in the background. Press return a few times and migration will be visible.

Docker issues

  • If you see the following error when you run docker, upgrade docker version as recommended in the Docker Version Requirements section.

    admin [ /migration ]$ ls
ls: cannot access ‘tools’: Operation not permitted
ls: cannot access ‘appliance’: Operation not permitted
ls: cannot access ‘appliance_data_template.json’: Operation not permitted
ls: cannot access ‘cli’: Operation not permitted
ls: cannot access ‘post-restore-template.yaml’: Operation not permitted
ls: cannot access ‘service_handlers’: Operation not permitted
ls: cannot access ‘util’: Operation not permitted
ls: cannot access ‘vcenter’: Operation not permitted
ls: cannot access ‘vm_cn_migration.py’: Operation not permitted
appliance  appliance_data_template.json  cli  post-restore-template.yaml  service_handlers  tools  util  vcenter  vm_cn_migration.py

Permission

If you see the following error while migrating the appliance, ensure that proper permissions are set for the migrationVolume:

22:03:34 PermissionError: [Errno 13] Permission denied: '/migrationVolume/migration.json'

Disk space issue

If MigrationVolume is filled up, run e2e-migrate -c command or backup -c command to cleanup before taking backups for such appliances. For the appliances which are successfully migrated, delete the artifacts manually.

Airgap issue

  • In Airgap VM, if you are unable to run the exec command into the container, check the status of the service and restart by performing the following steps:

    1. SSH to the appliance and su to the root.

    2. Check the service status systemctl status vmcnm.service.

    3. Run systemctl restart vmcnm.service.

    4. Execute into docker. If it still fails, refresh image.

    5. If exec still fails, refresh the image by performing the following steps. This is an advanced setup. If you encounter an issue with the vm-cn-migration image, you can forcibly load it from the disk:

      1. SSH to the appliance and cd /data/vmcnm_service.

      2. Run ./setup.sh true.

Deploying your own OVA

To use your own OVA:

  1. Upload your ova with the following format in the Vcenter. For example, VMware-Telco-Cloud-Automation-3.0.0-22803370-<Small/Medium/Large>.ova

  2. Specify the ovaurl in either of the formats given below:

    1. Example: "ovaUrl": "VMware-Telco-Cloud-Automation-3.0.0-22803370.ova",

  3. Run migration with the following command: python3 vm_cn_migration.py e2e-migrate <input-json file>.

    1. This will skip the deploy mode and perform the backup and migrate steps.

Migration failed with "Key Error: 'kubernetesVersion'"

In case the status of any Management Cluster deployment is Pending or Not Active state (Not fully deployed) then force delete such clusters before retrying migration.