Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

Upgrade Contrail Service Orchestration from Release 5.1.2 to Release 6.0.0

 
Summary

Follow this procedure to upgrade from CSO Release 5.1.2 to CSO Release 6.0.0.

Contrail Analytics Nodes (CAN) for CSO Release 6.0.0 run on CentOS version 7.7.1908. Upgrading CAN VM from CSO Release 5.1.2 to CSO Release 6.0.0 is not supported because of the database schema changes. As a workaround, after the upgrade procedure is completed, you must maintain CSO Release 5.1.2 CAN VMs for 30 days.

During this time, CSO fetches data from CSO Release 6.0.0 CAN and CSO Release 5.1.2 CAN VMs. You can remove CSO Release 5.1.2 CAN VMs after 30 days.

Upgrade Contrail Service Orchestration for KVM Hypervisor

Prerequisites

  • You must have at least 40 GB in the /root directory to run the upgrade script.

  • You must not delete previously installed CSO 5.1.2 folder from the startupserver VM.

  • Configure the three physical servers for HA setup. Ensure that all the prerequisites are met. For details, see Provision VMs on Contrail Service Orchestration Servers. Ensure that these three physical servers can connect to the existing CSO servers.

  • Create a bridge interface for KVM hypervisors. For details, see Create a Bridge Interface for KVM Hypervisors. You must assign new IP addresses.

  • Modify the main Apt sources configuration file on the new physical servers to connect the Debian sources.list to the Internet.

    root@host:~/# cp /etc/apt/orig-sources.list /etc/apt/sources.list

    You do not need to modify the file if Debian sources.list is connected to the Ubuntu repository.

  • Run the setup_bms.sh script on all the three new physical servers.

    Run the following commands from the CSO folder:

    root@host:~/Contrail_Service_Orchestration_6.0.0# cd ci_cd
    root@host:~/Contrail_Service_Orchestration_6.0.0# ./setup_bms.sh

    Verify that the package is installed successfully.

    root@host:~/# dpkg -l |grep libguestfs-tools
    Note

    If you run the setup_bms.sh script after creating the bridge interface, you may see an error-device br0 already exists; can't create bridge with the same name. You can ignore the error message.

Follow this procedure to upgrade from CSO Release 5.1.2 to CSO Release 6.0.0.

  1. Download the CSO Release 6.0.0 installer package from the CSO Downloads page to the startupserver1 VM.
  2. Log in to the startupserver1 VM as root.
  3. On the startupserver1 VM, extract the installer package.

    For example, if the name of the installer package is Contrail_Service_Orchestration_6.0.0.tar.gz,

    root@host:~/# tar –xvzf Contrail_Service_Orchestration_6.0.0.tar.gz

    The contents of the installer package are extracted in a directory with the same name as the installer package.

  4. Navigate to the CSO Release 6.0.0 directory in the startupserver1 VM.
    root@host:~/# cd Contrail_Service_Orchestration_6.0.0
  5. You can view the list of files in the Contrail_Service_Orchestration_6.0.0.
    root@host:~/Contrail_Service_Orchestration_6.0.0# ls

    The Contrail_Service_Orchestration_6.0.0.tar.tz file includes the upgrade.sh script.

  6. Run the upgrade.sh script.Warning

    Before you upgrade ensure that all ongoing jobs in the Administration Portal and Customer Portal are stopped; otherwise, the upgrade process will fail.

    root@host:~/Contrail_Service_Orchestration_6.0.0# ./upgrade.sh

    Enter the IPs and password for the three new physical servers, as prompted.

    Depending on your deployment, it may take 60 minutes to 120 minutes to complete this task.

    You can view the upgrade.log file which is available at root/Contrail_Service_Orchestration_6.0.0/logs folder.

    If an error occurs, you must fix the error and re-run the upgrade.sh script. When you re-run the upgrade.sh script, the script continues to execute from the previously failed step.

    If it fails after 2 attempts, contact Juniper Networks support for further assistance.

    You can run ./python.sh deploy_manager/utils/decrypt_password.py command to decrypt the passwords for each infrastructure component.

  7. Synchronize the Virtual Route Reflector (VRR). Use the admin token. Do not use the cspadmin token.
    1. Obtain the topo-uuid for the VRR. Use the topology-service API to find the UUID for the VRR.

      GET: https://<IP-address>/topology-service/device?q=fq_name:vrr*
    2. Synchronize the VRR using the POST https://<IP-address>/routing-manager/synchronize-vrr API. Use the UUID obtained in the previous step.

After a successful upgrade, CSO is functional and you can log in to the Administrator Portal and the Customer Portal.

Upgrade Contrail Service Orchestration for ESXi Hypervisor

Prerequisites

  • You must have at least 40 GB in the /root directory to run the upgrade script.

  • You must not delete previously installed CSO 5.1.2 folder from the startupserver VM.

  • Configure the three physical servers for HA setup. Ensure all the prerequisites are met. For details, see Provision VMs on Contrail Service Orchestration Servers.

    Ensure these three physical servers can connect to the existing CSO servers.

  • Create new contrail_analytics VMs on each of the three new physical servers.

    1. Download the appropriate installer package from the CSO Downloads page on any of the servers.

      Use the Contrail Service Orchestration installer package if you have purchased Network Service Orchestrator and Network Service Controller licenses for a distributed deployment.

    2. Expand the installer package.
      root@host:~/# tar –xvzf cso<version>.tar.gz

      The expanded package contains ESXi-6.0.0.tgz under the /Artifacts folder.

      Extract the ESXi-6.0.0.tgz package.

      The ESXi-6.0.0.tgz package contains the ubuntu-16.04-server-cloudimg-amd64.ova file, the junos-vrr-x86-64-19.4R1.12.ova file and the centos-77.ova file.

    3. Provision the contrail_analytics VMs using the centos-77.ova file.

      The VMs must match the server requirements specified in Minimum Requirements for Servers and VMs.

      The default username is root.

    4. Assign a new IP address to the logical interface (ens192) associated with the contrail_analytics VM.

      For example:

      vi /etc/sysconfig/network-scripts/ifcfg-ens192
    5. Configure a valid hostname for the contrail_analytics VMs. and update the /etc/hostname file.Note

      The hostnames must start and end with an alphanumeric character. The hostnames can contain only the following special characters—hyphen (-) and period (.). The hostnames cannot contain uppercase letters.

    6. Update the /etc/hosts file on the contrail_analytics VMs.

      For example:127.0.1.1 <hostname>.example.net <hostname>

    7. Reboot the contrail_analytics VMs.

Follow this procedure to upgrade from CSO Release 5.1.2 to CSO Release 6.0.0.

  1. Download the CSO Release 6.0.0 installer package from the CSO Downloads page to the startupserver1 VM.
  2. Log in to the startupserver1 VM as root.
  3. On the startupserver1 VM, extract the installer package.

    For example, if the name of the installer package is Contrail_Service_Orchestration_6.0.0.tar.gz,

    root@host:~/# tar –xvzf Contrail_Service_Orchestration_6.0.0.tar.gz

    The contents of the installer package are extracted in a directory with the same name as the installer package.

  4. Navigate to the CSO Release 6.0.0 directory in the startupserver1 VM.
    root@host:~/# cd Contrail_Service_Orchestration_6.0.0
  5. You can view the list of files in the Contrail_Service_Orchestration_6.0.0.
    root@host:~/Contrail_Service_Orchestration_6.0.0# ls

    The Contrail_Service_Orchestration_6.0.0.tar.tz file includes the upgrade.sh script.

  6. Run the upgrade.sh script.Warning

    Before you upgrade ensure that all ongoing jobs in the Administration Portal and Customer Portal are stopped; otherwise, the upgrade process will fail.

    root@host:~/Contrail_Service_Orchestration_6.0.0# ./upgrade.sh

    Enter the IPs and password for the three new contrail_analytics VMs, as prompted.

    Depending on your deployment, it may take 60 minutes to 120 minutes to complete this task.

    You can view the upgrade.log file which is available at root/Contrail_Service_Orchestration_6.0.0/logs folder.

    If an error occurs, you must fix the error and re-run the upgrade.sh script. When you re-run the upgrade.sh script, the script continues to execute from the previously failed step.

    If it fails after two attempts, contact Juniper Networks support for further assistance.

    You can run ./python.sh deploy_manager/utils/decrypt_password.py command to decrypt the passwords for each infrastructure component.

  7. Synchronize the Virtual Route Reflector (VRR). Use the admin token. Do not use the cspadmin token.
    1. Obtain the topo-uuid for the VRR. Use the topology-service API to find the UUID for the VRR.

      GET: https://<IP-address>/topology-service/device?q=fq_name:vrr*
    2. Synchronize the VRR using the POST https://<IP-address>/routing-manager/synchronize-vrr API. Use the UUID obtained in the previous step.

After a successful upgrade, CSO is functional and you can log in to the Administrator Portal and the Customer Portal.