UFM Enterprise Appliance Software Upgrade

NVIDIA UFM Enterprise Appliance Software User Manual v1.2.1

The installation process consists of replacing the containers/packages with the new version and upgrading the UFM data.

  1. Copy the tarball file to the UFM Enterprise Appliance, for example, to the /tmp folder.

  2. Connect to the UFM Enterprise Appliance via SSH.

  3. Stop the UFM service/cluster before upgrading.

    In SA mode, run:

    Copy
    Copied!
                

    #systemctl stop ufm-enterprise.service

    In HA mode, run:

    Copy
    Copied!
                

    # ufm_ha_cluster stop

  4. Extract the tarball file and run the installer for upgrade. Run:

    Copy
    Copied!
                

    # cd /tmp # tar xvf ufm-appliance-sw-<version>.tar # cd ufm-appliance-sw-<version> # ./install.sh

    Installer Options:

    Copy
    Copied!
                

    -q|--quiet             Upgrade UFM without prompt

    Warning

    In HA mode, this step should be performed on both servers.

  5. After the upgrade, start the UFM service/cluster.

    In SA mode, run:

    Copy
    Copied!
                

    # systemctl start ufm-enterprise.service

    In HA mode, run:

    Copy
    Copied!
                

    # ufm_ha_cluster start

  6. Wait one minute for the service to come up.

  7. Ensure the service health. Run:

    Copy
    Copied!
                

    # ufm_enterprise_sanity.sh Checking Service... Done Checking Images... Done Checking Containers... Done Checking ufm REST server... Done Sanity tests completed successfully!

Stop the UFM service before upgrading. Run:

Copy
Copied!
            

systemctl stop ufm-enterprise.service

For detailed information on upgrading the UFM docker in standalone mode, please refer to UFM Docker Upgrade.

Stop the UFM cluster before upgrading. Run:

Copy
Copied!
            

ufm_ha_cluster stop

For detailed information on upgrading the UFM docker in high availability mode, please refer to UFM Docker Upgrade.

  1. Stop the UFM cluster before upgrading. Run:

    Copy
    Copied!
                

    ufm_ha_cluster stop

  2. Download the UFM-HA package on both servers using the following command:

    Copy
    Copied!
                

    https://www.mellanox.com/downloads/UFM/ufm_ha_4.0.0-8.tgz

  3. On both servers, extract the downloaded UFM-HA package under /tmp/

  4. On both servers, go to the extracted directory /tmp/ufm_ha_XXX and run the installation script:

    Copy
    Copied!
                

    ./install.sh -l /opt/ha_data -d /dev/sda4 -p enterprise

    Option

    Description

    -l

    UFM enterprise files directory used by DRBD

    -d

    Partition name for DRBD

    -p

    UFM product name to run in HA mode

  5. On both servers, set a specific location (i.e. /opt/ufm/files) in the files /usr/local/bin/ha_utils/ufm-ha.conf and /usr/local/bin/ufm_ha_watcher.

    Copy
    Copied!
                

    sed -i -e "s#^pre_config.*#pre_config=touch /opt/ufm/files/config_ha#g" \            -e "s#^post_cleanup.*#post_cleanup=touch /opt/ufm/files/clean_ha#g" \          -e "s#^settings_dir.*#settings_dir=/opt/ufm/files/ufm_ha#g" /usr/local/bin/ha_utils/ufm-ha.conf

    Copy
    Copied!
                

    sed -i -e "s#^FAILOVER_FILE_PATH.*#FAILOVER_FILE_PATH=\"/opt/ufm/files/conf/failover.flag\"#g" /usr/local/bin/ufm_ha_watcher

  6. After the upgrade, start the UFM HA Cluster. Run:

    Copy
    Copied!
                

    ufm_ha_cluster start

© Copyright 2023, NVIDIA. Last updated on Sep 5, 2023.