Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

 

Hitless Software Upgrade of Data Center Devices Overview

 

Contrail Networking Controller supports the automation of basic device management functions such as software image upgrade on the devices in the data center fabric. You can perform Contrail Networking Controller-assisted maintenance activities such as a hitless software image upgrade on the leafs and spines of the data center fabric devices managed by Contrail Networking, with zero packet loss.

Software image upgrade on a networking device in a data center is a time consuming task and might include rebooting the device. During upgrade, if user traffic is being routed through the device then the packets are lost which adversely affects the data center fabric performance.

During hitless upgrade, the devices are placed in a new mode called maintenance mode for the duration of the maintenance activity. The following sequence of steps are performed during hitless upgrade.

  • Initial Verification

    • Verifying that traffic can be routed from the selected device to another equally capable device. If no such device is present, then hitless upgrade cannot be performed because there will be traffic loss.

    • Verifying that the selected upgrade image is compatible with the devices.

    • Performing health checks on devices. Health checks are pre-configured parameters against which the devices are checked. If the health checks for devices fail, then the upgrade process for that device is aborted by default. However, you can change the default setting to not abort upgrade upon health check failure.

    If all the checks in the initial verification are cleared, Contrail Networking Controller places the device in the maintenance mode and performs the software upgrade.

  • Maintenance Mode

    • Before the device or devices are placed in maintenance mode, Contrail Networking Controller captures a snapshot of the existing state of the device. This snapshot is used to verify the operational state of the device when the maintenance activity or software upgrade is completed.

    • The traffic flowing through the device is rerouted through another equally capable device and the Contrail Networking Controller verifies that there is no traffic flowing through the device.

    • The device is then taken offline and placed in the maintenance mode.

    • The Contrail Networking Controller upgrades the software image to the required version on the device.

  • Final Verification

    • The device is taken out of the maintenance mode and traffic is routed through it again.

    • Contrail Networking Controller captures a snapshot of the operational state of the device to verify against the snapshot taken previously.

Note

For hitless software upgrade to work as per design and for zero packet loss, all devices must be redundantly connected. If any device is not redundantly connected, then you will have connectivity and packet loss when the device reboots.

Benefits of Hitless Software Upgrade

  • Maintenance activities can be performed on devices in a data center without a maintenance window.

  • No user traffic is lost is lost during image upgrade on devices in the data center.