NVIDIA NVOS User Manual for InfiniBand Switches

Upgrading Operating System Software

Important Notes Before Upgrading OS Software

Consider the following items prior to upgrading the operating system:

  • The system becomes unavailable while OS upgrade is in progress.

  • The upgrade procedure burns the software image as well as the firmware.

  • Before upgrading the software image, make sure to close all CLI sessions besides the one used to run the upgrade process.

 

In Service Software Upgrade (ISSU):

ISSU is a method to update the switch NOS software bundle, to a newer version then the currently running version, without disrupting the forwarding of traffic through the switch.

During the upgrade, the control and management planes incur unavailability while moving to the new OS version and refreshing the internal software databases.

However, the switch ASIC(s) are still forwarding traffic and keeping all physical and logical ports uninterrupted, even while the firmware inside the ASIC is being updated.

The ISSU flow involves interaction with the Subnet Manager (SM) to get its approval to launch ISSU.

The command to issue ISSU is the same command to perform regular software update with two additional flags

admin@nvos:~$ nv action install system image files new-nvos-image.bin ?

    [Enter]
    reboot                 Install an image and reboot system
    force                  Force system reboot after system image installation
    issu                   Install image in ISSU mode
    skip-sm                Skip SM confirmation for ISSU

Additionally on systems that support ISSU, its status will be included in nv show system image output.

admin@nvos:~$ nv show system image
             operational          pending
-----------  -------------------  -------
current      1                          
next         1                          
issu-status  no-issu                    
partition1                              
  build-id   nvos-25.02.2201
partition2        
  build-id   nvos-25.02.2202

 

adding issu flag would trigger this flow and avoid the reboot.

skip-sm should be used only when there is a need to perform ISSU while SM is not available. This needs special care in a production environment.

ISSU is disabled by default on OpenSM, and needs a special configuration to enable it, please refer to OpenSM release notes.


Upgrading OS Software

To upgrade NVOS, perform the following steps.

  1. Display the image (.bin file) that is currently available.

    admin@nvos:~$ nv show system image
                 operational          pending
    -----------  -------------------  -------
    current      1                          
    next         1                          
    issu-status  no-issu                    
    partition1                              
      build-id   nvos-25.02.2201
    partition2        
      build-id   nvos-25.02.2202            
    


  2. In case there are partition1 and partition2, Uninstall the image that isn't used as 'current' or 'next' prior to fetching the new image. Use the command nv action uninstall system image, for this purpose. 

    admin@nvos:~$ nv action uninstall system image
    

    Another option is to clean-up all the unused images.

    admin@nvos:~$ nv action uninstall system image force
    


  3. Upload the new software image to the system using fetch command.

    admin@nvos:~$ nv action fetch system image scp://<username>:<password>@<ip-address>/var/www/html/<new_image>
    


  4. Install the new image.

    admin@nvos:~$ nv action install system image files new-nvos-image.bin
    Operation will reboot the system.
    If you choose ‘y’, the system will install the image and reboot.
    If you choose ‘N’, the operation will abort without installing the image and without rebooting the system.
    Do you want to continue? [y/N]
    


  5. After reboot, display available images and verify that the new image now appears.

    admin@nvos:~$ nv show system image 
                 operational          applied
    ----------  -------------------   ------- 
    current     2                  
    next        2                                    
    partition1
      build-id  <old_image>
    partition2
      build-id  <new_image>
    


    • After software reboot, the software upgrade will also automatically upgrade the firmware version.

    • To recover from image corruption (e.g., due to power interruption), there are two installed images on the system. For more information, see the nv action boot-next system image command.



Last updated: