Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

 
 

Updating Contrail Networking using the Zero Impact Upgrade Process in an Environment using Red Hat Openstack 13

This document provides the steps needed to update a Contrail Networking deployment that is using Red Hat Openstack 13 as it’s orchestration platform. The procedure provides a zero impact upgrade (ZIU) with minimal disruption to network operations.

If you are using Contrail Networking in an environment that is using a Red Hat Openstack 16-based release, see Updating Contrail Networking using the Zero Impact Upgrade Process in an Environment using Red Hat Openstack 16

When to Use This Procedure

This procedure is used to upgrade Contrail Networking when it is running in environments using RHOSP13.

The procedure in this document has been validated for the following Contrail Networking upgrade scenarios:

Table 1: Contrail Networking with RHOSP13 Validated Upgrade Scenarios

Starting Contrail Networking Release

Target Upgraded Contrail Networking Release

1912.L0 or later 1912.L release

1912.L1 or later 1912.L release

Note:

You can upgrade from any 1912.L release to any other 1912.L release.

1912.L0

2003

2003

2005

2005

2008

2008

2011

2011

Any 2011.L Release

Any 2011.L Release

Any 2011.L Release

You cannot run Contrail Networking in environments using Red Hat Openstack 13 orchestration using Contrail Networking Release 21-based releases. The first Contrail Networking Release 21-based release is Contrail Networking Release 21.3.

A different procedure is followed for upgrading to earlier target Contrail Networking releases in environments using RHOSP13 orchestration. See Upgrading Contrail Networking with Red Hat Openstack 13 using ISSU.

If you want to use this procedure to upgrade your Contrail Networking release to other releases, you must engage Juniper Networks professional services. Contact your Juniper representative for additional information.

Prerequisites

This document makes the following assumptions about your environment:

  • A Contrail Networking deployment using Red Hat Openstack version 13 (RHOSP13) as the orchestration platform is already operational.

  • The overcloud nodes in the RHOSP13 environment have an enabled Red Hat Enterprise Linux (RHEL) subscription.

  • Your environment is running Contrail Release 1912 and upgrading to Contrail Release 1912-L1 or to Contrail Release 2003 or later.

  • If you are updating Red Hat Openstack simultaneously with Contrail Networking, we assume that the undercloud node is updated to the latest minor version and that new overcloud images are prepared for an upgrade if needed for the upgrade. See the Upgrading the Undercloud section of the Keeping Red Hat OpenStack Platform Updated guide from Red Hat.

    If the undercloud has been updated and a copy of the heat templates are used for the deployment, update the copy of the heat template from the Red Hat’s core heat template collection at /usr/share/openstack-tripleo-heat-templates. See the Understanding Heat Templates document from Red Hat for information on this process.

Before You Begin

We recommend performing these procedures before starting the update:

  • Backup your Contrail configuration database before starting this procedure. See How to Backup and Restore Contrail Databases in JSON Format in Openstack Environments Using the Openstack 13 or Ansible Deployers.

  • Each compute node agent will go down during this procedure, causing some compute node downtime. The estimated downtime for a compute node varies by environment, but typically took between 12 and 15 minutes in our testing environments.

    If you have compute nodes with workloads that cannot tolerate this downtime, consider migrating workloads or taking other steps to accommodate this downtime in your environment.

  • If you are updating Red Hat Openstack simultaneously with Contrail Networking, update Red Hat Openstack to the latest minor release version and ensure that the new overcloud images are prepared for the upgrade. See the Upgrading the Undercloud section of the Keeping Red Hat OpenStack Platform Updated guide from Red Hat for additional information.

    If the undercloud has been updated and a copy of the heat templates are used for the deployment, update the Heat templates from Red Hat’s core Heat template collection at /usr/share/openstack-tripleo-heat-templates. See the Understanding Heat Templates document from Red Hat for additional information.

Updating Contrail Networking in an Environment using Red Hat Openstack

To update Contrail Networking in an environment that is using Red Hat Openstack as the orchestration platform:

  1. Prepare your docker registry. The registry is often included in the undercloud, but it can also be a separate node.

    Docker registry setup is environment independent. See Docker Registry from Docker for additional information on Docker registry setup.

  2. Backup the Contrail TripleO Heat Templates. See Using the Contrail Heat Template.
  3. Get the Contrail TripleO Heat Templates (Stable/Queens branch) from https://github.com/Juniper/contrail-tripleo-heat-templates.
  4. (Optional) Update the Contrail TripleO Puppet module to the latest version and prepare Swift Artifacts, as applicable.

    Below are sample commands entered in the undercloud:

  5. Update the parameter ContrailImageTag to the new version.

    The location of the ContrailImageTag variable varies by environment. In the most commonly-used environments, this variable is set in the contrail-services.yaml file.

    You can obtain the ContrailImageTag parameter from the README Access to Contrail Registry 20XX.

  6. (Recommended) If you are upgrading to Contrail Networking Release 2005 or later, check and, if needed, enable kernel vRouter huge page support to support future compute node upgrades without rebooting.

    You can enable or verify kernel-mode vRouter huge page support in the contrail-services.yaml file using either the ContrailVrouterHugepages1GB: and ContrailVrouterHugepages2MB: parameters:

    Notes about kernel-mode vRouter huge page support in Red Hat Openstack environments:

    • Kernel-mode vRouter huge page support was introduced in Contrail Networking Release 2005, and is configured to support 2 1GB huge pages by default in Contrail Networking Release 2005 or later.

      A compute node has to be rebooted once for a huge page configuration to finalize. After this initial reboot, the compute node can perform future Contrail Networking software upgrades without rebooting.

      Notably, a compute node in an environment running Contrail Networking 2005 or later has not enabled huge page support for kernel-mode vRouters until it is rebooted. The 2x1GB huge page support configuration is present by default, but it isn’t enabled until the compute node is rebooted.

    • We recommend only using 1GB or 2MB kernel-mode vRouter huge pages in most environments. You can, however, simultaneously enable 1GB or 2MB kernel-mode vRouter huge pages in Red Hat Openstack environments if your environment requires enablement of both huge page options.

    • Changing vRouter huge page configuration settings in a Red Hat Openstack environment typically requires a compute node reboot.

      • 1 GB pages: Reboot required.

      • 2 MB: Reboot usually required. The reboot is sometimes avoided in environments where memory isn’t highly fragmented or the required number of pages can be easily allocated.

    • We recommend allotting 2GB of memory—either using the default 1024x2MB huge page size setting or the 2x1GB size setting—for huge pages in most environments. Some larger environments might require additional huge page memory settings for scale. Other huge page size settings should only be set by expert users in specialized circumstances.

    • If the ContrailVrouterHugepages1GB: and ContrailVrouterHugepages2MB: parameters are set to empty value in the contrail-services.yaml file, vRouter huge pages are disabled.

  7. Update the overcloud by entering the openstack overcloud update prepare command and include the files that were updated during the previous steps with the overcloud update.

    Example:

  8. Prepare the overcloud nodes that include Contrail containers for the update.
    • Pull the images in the repository onto the overcloud nodes.

      There are multiple methods for performing this step. Commonly used methods for performing this operation include using the docker pull command for Docker containers and the openstack overcloud container image upload command for Openstack containers, or running the tripleo-heat-templates/upload.containers.sh and tools/contrail/update_contrail_preparation.sh scripts.

    • (Not required in all setups) Provide export variables for the script if the predefined values aren’t appropriate for your environment. The script location:

      The following variables within the script are particularly significant for this upgrade:

      • CONTRAIL_NEW_IMAGE_TAG—The image tag of the target upgrade version of Contrail. The default value is latest.

        If needed, you can obtain this parameter for a specific image from the README Access to Contrail Registry 20XX.

        Note:

        Some older deployments use the CONTRAIL_IMAGE_TAG variable in place of the CONTRAIL_NEW_IMAGE_TAG variable. Both variables are recognized by the update_contrail_preparation.sh script and perform the same function.

      • SSH_USER—The SSH username for logging into overcloud nodes. The default value is heat-admin.

      • SSH_OPTIONS—Custom SSH option values.

        The default SSH options for your environment are typically pre-defined. You are typically only changing this value if you want to customize your update.

      • STOP_CONTAINERS—The list of containers that must be stopped before the upgrade can proceed. The default value is contrail_config_api contrail_analytics_api.

    • Run the script:

      CAUTION:

      Contrail services stop working when the script starts running.

  9. Update the Contrail Controller nodes:
    • Run the openstack overcloud update run command on the first Contrail controller and, if needed, on a Contrail Analytics node. The purpose of this step is to update one Contrail Controller and one Contrail Analytics node to support the environment so the other Contrail Controllers and analytics nodes can be updated without incurring additional downtime.

      Example:

      Ensure that the contrail status is ok on overcloud-contrailcontroller-0 before proceeding.

      If the analytics and the analyticsdb nodes are on separate nodes, you may have to update the nodes individually:

    • After the upgrade, check the docker container status and versions for the Contrail Controllers and the Contrail Analytics and AnalyticsDB nodes.

    • Update the remaining Contrail Controller nodes:

      Example:

  10. Update the Openstack Controllers using the openstack overcloud update run commands:

    Example:

  11. Individually update the compute nodes.
    Note:

    The compute node agent will be down during this step. The estimated downtime varies by environment, but is typically between 1 and 5 minutes.

    Consider migrating workloads that can’t tolerate this downtime before performing this step

    Reboot your compute node to complete the update.

    Note:

    A reboot is required to complete this procedure only if a kernel update is also needed. If you would like to avoid rebooting your compute node, check the log files in the /var/log/yum.log file to see if kernel packages were updated during the compute node update. A reboot is required only if kernel updates occurred as part of the compute node update procedure.

    Use the contrail-status command to monitor upgrade status. Ensure all pods reach the running state and all services reach the active state.

    This contrail-status command provides output after a successful upgrade:

    Note:

    Some output fields and data have been removed from this contrail-status command sample for readability.

  12. Enter the openstack overcloud update converge command to finalize the update.
    Note:

    The options used in the openstack overcloud update converge in this step will match the options used with the openstack overcloud update prepare command entered in 7.

    Monitor screen messages indicating SUCCESS to confirm that the updates made in this step are successful.