Quick Deployment and Upgrade Guide

This quick deployment and upgrade guide gives only the essential steps, and assumes that you understand the deployment and the upgrade of an IBM Storage Scale System. You can deploy or upgrade an IBM system from a management server, such as Power9 or a management server virtual machine (MSVM) for x86 systems. The MSVM runs on IBM® Storage Scale System Utility Node, which is a bare metal.

Prerequisites

Complete the following steps before the deployment and the upgrade:
  1. Download the IBM Storage Scale System package from IBM Fix Central.
    Scale_System_UNIFIED-7.0.1.x-EMS.tgz
  2. Populate the /etc/hosts file to the management server.
  3. Extract the installation package.
    cd /home/deploy
    
    xz --decompress <xz package>
  4. Start the container.
    sh <untarred binary file>
  5. Import the system configuration and define nodes.
    essrun -N MS1,IOnode1,IOnode2,IOnoden,prt01,prt02 config load -p
    where:
    MS1
    Specifies a management server, such as Power9, MSVM for x86 systems.
    IOnode1,IOnode2,IOnoden
    Specifies one or more I/O nodes.
    prt01,prt02
    Specifies one or more protocol nodes.
  6. Check the defined nodes before the deployment and the upgrade.
    essrun -N MS1,IOnode1,IOnode2,IOnoden,prt01,prt02 update --precheck

Deployment from the Power9 management server

  1. Complete all steps in the Prerequisites.
  2. If the Power9 server needs to be updated, run the following command:
    Note: You might need to run the following command again, if the kernel is changed.
    essrun -N MS1 update --offline
  3. If canisters, I/O nodes, or protocol nodes need to be updated, run the following command:
    essrun -N IOnode1,IOnode2,IOnoden,prt01,prt02 update --offline
  4. Create network bonds.
    essrun -N MS1,IOnode1,IOnode2,IOnoden network --suffix=-hs
  5. Create a cluster.
    essrun -N IOnodeclasses cluster --suffix=-hs
  6. Add the Power9 server to the cluster.
    essrun -N IOnode1 cluster --add-ems MS1 --suffix=-hs
  7. Create a file system.
    essrun -N IOnodeclasses filesystem --name fs1 --suffix=-hs
    Start of changeWhere the I/O node classes can be similar to the ones in the following examples:
    • IBM Storage Scale System 3000: ess30001a and ess30001b
    • IBM Storage Scale System 3500: ess35001a and ess35001b
    • IBM Storage Scale System 6000: ess60001a and ess60001b
    End of change
  8. Add a protocol node to the cluster. For more information, see Creating a shared root file system for CES.
  9. Log in to each node and check whether nodes are installed properly.
    essinstallcheck -N localhost
  10. From the Power9 management server do the final health check.
    gnrhealthcheck
    mmhealth node show -a
  11. Set a time zone and set up Chrony. Refer to Setting a time zone.
  12. Set up call home. For more information about the call home configuration, see Configuring call home.
  13. Configure the GUI for hardware monitoring.
    essrun -N MS1,IOnode1,IOnode2 gui --configure

Deployment from an MSVM for x86 systems

Complete the following steps to deploy a system from an MSVM that runs on the Utility Node.

  1. Download the Scale_System_VM-x.x.x.x-x86_64-UTILITY-EMS.tgz package on the /serv directory and extract the package in the /serv directory.
  2. Copy all downloaded software to the /serv directory on the Utility Node.
  3. Extract the EMSVM script package in the /serv directory.
  4. Check the Utility Node.
    cd /serv/EMSVM-main ; ./emsvm --check-host
  5. Start the Utility Node.
    ./emsvm --start-EMS
  6. Connect to the Utility Node.
    ./emsvm --connect-EMS
  7. Log in to the virtual machine by using the default root/ibmesscluster credentials and change the root password immediately for security.
  8. Run the ess_ssr_setup tool and follow the instructions.
  9. Set the fully qualified domain name (FQDN).
    hostnamectl set-hostname FQDN
  10. Complete all steps in the Prerequisites.
  11. If the MSVM needs to be updated, run following command:
    Note: You might need to run the following command again, if the kernel is changed:
    essrun -N EMSVM update --offline
    Where, EMSVM specifies an MSVM that runs on Utility Node.
  12. If canisters, I/O nodes, or protocol nodes need to be updated, run following command:
    essrun -N IOnode1,IOnode2,IOnoden,prt01,prt02 update --offline
  13. Create network bonds.
    essrun -N EMSVM,IOnode1,IOnode2,IOnoden network --suffix=-hs
  14. Create a cluster.
    essrun -N IOnodeclasses cluster –suffix=-hs
  15. Add the MSVM to the cluster.
    essrun -N IOnode1 cluster --add-ems EMSVM --suffix=-hs
  16. Create a file system.
    essrun -N IOnodeclasses filesystem --name fs1 --suffix=-hs
    where, fs1 specifies a file system.
  17. Create a cesSharedRoot file system.
    essrun -N IOnode1,IOnode2 filesystem --ces
  18. Set up the Cluster Export Services (CES) by following instructions in Adding and configuring protocols in an exiting cluster by using IBM Storage Scale installation toolkit.
  19. Configure GUI for hardware monitoring.
    essrun -N EMSVM,IOnode1,IOnode2 gui --configure
Start of change

Precheck nodes

Check the system health before upgrade. You can use the precheck tool to check all nodes on a cluster for any issues. For more information, see Checking system health.

For complete documentation, troubleshooting guidance, and timeout configuration, see the README.md file in the GitHub repository (https://github.com/IBM/Scale_System_PRECHECK_TOOL).

End of change

Upgrade from the Power9 management server

You can upgrade online or offline from the Power9 management server.

  1. Complete all steps in the Prerequisites.
  2. Update the Power9 server from the container.
    Note: You might be prompted to run the following command again, if the kernel is changed:
    essrun -N MS1 update --offline
  3. If a cluster exists in the environment, restart the performance monitoring and GUI.
    systemctl restart pmsensors
    systemctl restart gpfsgui
  4. If the Power9 firmware needs to be updated in the management server, run the following commands:
    cd /install/ess/otherpkgs/rhels8/ppc64le/firmware/
    update_flash -v -f 01VL950_131_045.img
    update_flash -f 01VL950_131_045.img
    update_flash -c
  5. After the management server node is updated from the container, run the following command:
    • Offline mode:
      essrun -N IOnode1,IOnode2,IOnoden update --offline
    • Online mode:
      essrun -N IOnode1,IOnode2,IOnoden update
  6. Restart the performance monitoring service on each upgraded I/O node.
    systemctl restart pmsensors
  7. Update enclosures firmware.
    • Offline mode:
      mmchfirmware --type storage-enclosure --fast-offline
    • Online mode:
      mmchfirmware --type storage-enclosure
  8. Update the drive firmware.
    • Offline mode:
      mmchfirmware –type drive --fast-offline
    • Online mode:
      mmchfirmware --type drive
  9. If the Power9 firmware needs to be updated in the I/O nodes, run the following commands:
    cd /install/ess/otherpkgs/rhels8/ppc64le/firmware/
    update_flash -v -f 01VL950_131_045.img
    update_flash -f 01VL950_131_045.img
    update_flash -c
  10. Update IBM Storage Scale in the protocol nodes.
  11. Upgrade protocol nodes operating system and other components by using the container.
    • Offline mode:
      essrun -N prt1,prt2 update --offline
    • Online mode:
      essrun -N prt1,prt2 update
  12. If the Power9 firmware needs to be updated in the protocol nodes, run the following commands:
    cd /install/ess/otherpkgs/rhels8/ppc64le/firmware/
    update_flash -v -f 01VL950_131_045.img
    update_flash -f 01VL950_131_045.img
    update_flash -c
  13. Run a health check on each node that was upgraded.
    essinstallcheck

Upgrade from an MSVM for x86 systems

You can upgrade online or offline from an MSVM that runs on the Utility Node.

  1. Complete all steps in the Prerequisites.
  2. Update the MSVM from the container.
    Note: You might be prompted to run the following command again, if the kernel is changed.
    essrun -N EMSVM update --offline
  3. If a cluster exists in the environment, restart the performance monitoring and GUI.
    systemctl restart pmsensors
    systemctl restart gpfsgui
  4. After the MSVM is updated from the container, run the following command:
    • Offline mode:
      essrun -N IOnode1,IOnode2,IOnoden update --offline
    • Online mode:
      essrun -N IOnode1,IOnode2,IOnoden update
  5. Restart the performance monitoring service on each upgraded I/O node.
    systemctl restart pmsensors
  6. Update firmware of enclosures.
    • Offline mode:
      mmchfirmware --type storage-enclosure --fast-offline
    • Online mode:
      mmchfirmware --type storage-enclosure
  7. Update the drives firmware.
    • Offline mode:
      mmchfirmware --type drive --fast-offline
    • Online mode:
      mmchfirmware --type drive
  8. If the Power9 firmware needs to be updated in the I/O nodes, run the following commands:
    cd /install/ess/otherpkgs/rhels8/ppc64le/firmware/
    update_flash -v -f 01VL950_131_045.img
    update_flash -f 01VL950_131_045.img
    update_flash -c
  9. Update IBM Storage Scale on the protocol nodes.
    1. For more information about how to upgrade IBM Storage Scale on protocol nodes offline, see Performing offline upgrade or excluding nodes from upgrade by using installation toolkit.
    2. For more information about how to upgrade IBM Storage Scale on protocol nodes online, see Performing online upgrade by using the installation toolkit.
  10. Upgrade operating system of protocol nodes and other components by using the container.
    • Offline mode:
      essrun -N prtvm1,prt2 update --offline
      where, prtvm1 specifies a protocol node virtual machine.
    • Online mode:
      essrun -N prtvm1,prt2 update
  11. If the Power9 firmware needs to be updated on the protocol nodes, run the following commands:
    cd /install/ess/otherpkgs/rhels8/ppc64le/firmware/
    update_flash -v -f 01VL950_131_045.img
    update_flash -f 01VL950_131_045.img
    update_flash -c
  12. Verify whether all upgraded nodes were installed properly.
    essinstallcheck

Error handling

If you encounter any issues during the deployment, see Troubleshooting.