InfiniBand Cluster Bring-up Procedure

Transceiver Firmware Installation

Transceivers Firmware Installation

For manage switch you can upgrade the transciever FW using UFM or with NVOS (XDR systems only).

For unmanaged switch and servers you can upgrade using MFT.

Via UFM:

  1. Navigate to managed elements page.

  2. select the target switches and click on Upgrade Cable Transceivers option.

    cable-upgrade-version-1-modificationdate-1707037289260-api-v2.JPG
  3. A model will be shown containing list of the active firmware versions for the cables of the selected switches, besides the version number, a badge will show the number of matched switches:

    image2022-4-15_14-28-38-version-1-modificationdate-1707037300840-api-v2.png


    image2022-4-15_14-29-1-version-1-modificationdate-1707037300513-api-v2.png


  4. After the user clicks Submit, the GUI will start sending the selected binaries with the relevant switches sequentially, and a model with a progress bar will be shown (this model can be minimized):

    upgrade-cable-trans-version-1-modificationdate-1707037289613-api-v2.JPG


  5. After the whole action is completed successfully, you will be able to see the following message at the model bottom The upgrade cable transceivers completed successfully, do you want to activate it? by clicking the yes button it will run a new action on all the burned devices to activate the new uploaded binary image.

  6. Another option to activate burned cables transceivers you can go to the Groups page and right click on the predefined Group named Devices Pending FW Transceivers Reset or you can right click on the upgraded device from managed element page and select Activate cable Transceivers action.

    upgrade-cable-trans2-version-1-modificationdate-1707037290093-api-v2.JPG

    For further information, see UFM User Manual.

Via MFT:

  1. Query the Transceiver firmware information. 

    flint -d lid-1 --linkx --downstream_device_ids 1 q
    

    When the cluster has many switches, multiple hosts may be engaged in the upgrade process. Each device (NIC, Switch) can update only the modules connected directly to it, not the far end. Updating the far-end transceiver requires the same operation to be done at the far-end switch(es).

  2. Burn the cable using the Auto-update command.

    flint -d <device> --linkx--linkx_auto_update--download_transfer-i<binary file> b
    

     

  3. Activate the new firmware:

    flint -d lid-2 --linkx--linkx_auto_update–-activate b
    

More reading of flint for cables can be found in Cable Burn Command

Via NVOS - on XDR systems only:

  1. Show the Transceiver firmware information. 

    admin@croc-94-mgmt2:~$ nv show platform transceiver
    Transceiver  Identifier                     Vendor name  Vendor PN         Vendor SN      Vendor revision
    -----------  -----------------------------  -----------  ----------------  -------------  ---------------
    swA1         OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N001-FTF  MT2214VS03808  22             
    swA2         OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N001-FTF  MT2214VS03792  22             
    swA3         OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2139VS01287  33             
    swA4         OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2139VS01287  33             
    swA5         OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01231  33             
    swA6         OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01231  33             
    swA7         OSFP 8X Pluggable Transceiver  NVIDIA       MMS4X00-NV4       MT2220FT12808  11             
    swA8         OSFP 8X Pluggable Transceiver  NVIDIA       MMS4X00-NV4       MT2220FT14138  11             
    swA10        OSFP 8X Pluggable Transceiver  NVIDIA       NMA4OPT-XFJALML2  MT2416AC91448  11             
    swA11        OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01242  33             
    swA12        OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01242  33             
    swA13        OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01255  33             
    swA14        OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01255  33             
    swA15        OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00B      MT2234VS02570  33             
    swB1         OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01543  33             
    swB2         OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01543  33             
    swB3         OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01272  33             
    swB4         OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01272  33             
    swB5         OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2150VS00492  33             
    swB6         OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2150VS00492  33             
    swB7         OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01502  33             
    swB8         OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01502  33             
    swB9         OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01219  33             
    swB10        OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01219  33             
    swB11        OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01532  33             
    swB12        OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01532  33             
    swB13        OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01517  33             
    swB14        OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00A      MT2237VS01517  33             
    swB15        OSFP 8X Pluggable Transceiver  NVIDIA       MCP4Y10-N00B      MT2234VS02570  33     
    
  2. Note that we can have both optical & copper type transceivers - these can be identified by Vendor PN and running show explicitly.

For example: swA7 is optical and swA10 is copper. each transceiver type has different firmware image.

       

admin@croc-94-mgmt2:~$ nv show platform transceiver swA7 | grep "cable-type"
cable-type             : Optical module


admin@croc-94-mgmt2:~$ nv show platform transceiver swA10 | grep "cable-type"
cable-type             : Copper cable

3. In order to see the current firmware version for the transceiver - add firmware param as the following:


admin@croc-94-mgmt2:~$ nv show platform transceiver swA10 firmware
                      operational
--------------------  -----------
actual-firmware       39.230.20  
fw-upgrade-status     N/A        
fw-upgrade-error-msg  N/A       

4. The next step will be to fetch the right firmware image (and verify):

admin@croc-94-mgmt2:~$ nv action fetch platform firmware transceiver scp://root:12345@10.237.38.141/auto/sw_system_project/NVOS_INFRA/verification_files/transceiver_fw/Xodin/fw_39_230_00024_dev_signed.bin
Action executing ...
Fetching file ...
Action executing ...
File fetched successfully
Action succeeded  
admin@croc-94-mgmt2:~$ nv show platform firmware transceiver files 
Available Firmware Files              File Path                                                       
------------------------------------  ----------------------------------------------------------------
fw_39_230_00024_dev_signed.bin        /host/fw-images/transceiver/fw_39_230_00024_dev_signed.bin      


5. Install the fetched firmware image:

admin@croc-94-mgmt2:~$ nv action install platform transceiver swA10 firmware files fw_39_230_00024_dev_signed.bin
Action executing ... 1%
Installing firmware: fw_39_230_00024_dev_signed.bin
Action executing ... 14%
Installed FW: 39.230.24
Action succeeded

6. Verify the image was installed successfully

admin@croc-94-mgmt2:~$ nv show platform transceiver swA10 firmware
                      operational
--------------------  -----------
actual-firmware       39.230.24  
fw-upgrade-status     OK         
fw-upgrade-error-msg  N/A   


Last updated: