How it Works

How it works – TPVM Persistence

When using the Legacy method, the TPVM installation as well its configuration were configured through the ‘Exec‘ mode commands. The TPVM/Ubuntu was an independent Guest OS with its own RootFS and it will a make a copy or a static entry of the above executed configuration files. For example, if the TPVM Management Interface ‘eth0‘ is configured, from SLX, the TPVM Rootfs /etc/network/interface file will be edited to maintain the static entry for its future use.

The above model is good for any kind of TPVM reboot or powercycle of the SLX device and is insufficient, if TPVM is upgraded with a new image or SLX switch. The RMA is also installed with the new SLX switch, but with the old SLX switch running-configurations.

In the new way of installing TPVM, installing TPVM and its configuration is supported in the ‘Config‘ mode. Whereas all the TPVM configurations are also persisted at the SLX OS Configuration database, so they become part of the switch ‘running-configurations‘.
Note

Note

It can be either of the configuration modes, only one TPVM instance (named TPVM) is deployed or installed and its Debian image is expected to be available at the /tftpboot/SWBD2900 folder. The only major difference is being persistent at the SLX-OS, in the new Config mode.

The TPVM configurations are listed below as 3 types:

  • Install time configuration
    • Password
    • Interface Management
    • Interface Insight (to be implemented in the future release).
  • Run time configuration

    • auto-boot

    • hostname

    • timezone

    • dns

    • ntp

    • ldap host

    • ldap ca-cert

    • trusted-peer

    • disk (to be implemented in future release)

  • Action

    • deploy

The install time configuration should be always configured before the deploy configuration, else they will fail. As they are applied to the TPVM Instance RootFS during the installation time. The settings before the deploy configuration persist, as these configurations are present in the SLX-OS running-configuration database. Whenever the deploy is configured, at that time, these configurations are applied during the TPVM installation stage.

The post deploy configuration, if required, will stop the TPVM, by the command tpvm stop. Then the tpvm config submode is set at the install time configuration, followed by the tpvm start command.

The Run time configurations may be configured any time. If configured before ‘deploy‘, they are just persisted at SLX-OS running-configuration Database. If configured after ‘deploy‘ configuration, it implies that the TPVM is running and the configuration is first applied to the TPVM run-time. If they succeed only, then it is persisted to the SLX-OS running-configuration database.

The Deploy configuration implies the following:

  • It will look for TPVM Debian image package file under /tftpboot/SWBD2900 folder.
  • It will install the Debian image package.
      • poll for some timeout for install completion else report timeout
  • It will apply the Install time configuration if found persisted in the SLX-OS running-configuration database.
  • It will start the TPVM Guest OS.
    • The poll for timeout of the start completion, of the TPVM boot up, reports timeout. (The boot operation may still go on in the background).
  • It will apply the Run time configuration if found persisted in the SLX-OS running-configuration database.

Across normal SLX-OS reboot, there is no change in the configuration especially in the deploy configuration. Post boot action is needed and the TPVM will reboot with its configurations and if auto-boot is set manually through the tpvm start command.

Note

Note

The interface insight and the disk configurations, are currently not supported by the SLX-OS backend and it will report failure if tried configuring. By default for the first time the disk vdb is created and mounted to the /apps TPVM legacy. The subsequent deploy or install will reuse the existing disk vdb. Commands like “tpvm uninstall force” is needed to purge the vdb.

How it works – across RMA

When the SLX switch is RMA and its replaced by another SLX switch, it's expected that the new switch will have the desired SLX-OS image as well as the TPVM Debian image file.

The default-config new switch will not have any TPVM related configurations. The corresponding cleanup will be needed, if any switch boots with the default-config.

The Admin user will have the backup of the running-configuration last switch, and it will copy the same to the new switch running-configuration through the command:

For example - SLX# copy flash://old.cfg running-config.

or

SLX# copy scp://<username>:<password>@hostname/<filepath> running-config

If copied, the running-configuration will have the tpvm configurations and they will be applied.

How it works – across upgrade

The new tpvm upgrade command can be explained as 2 main steps:

First download the new Debian image package file and then as explained above, re-deploy the TPVM.

Re-deploying means stop and uninstall running TPVM instances. Hence it's like fresh deploying.

As explained for deploying, if any TPVM configurations are found in the SLX-OS running-configuration Database, they are also applied to TPVM.

9037309-01 Rev AA