Maintenance of a node

This topic describes the procedures for taking off an IBM Storage Scale Erasure Code Edition node from service for maintenance.

For the general procedure for maintenance of an IBM Storage Scale Erasure Code Edition node, follow the steps that are described in Manual online upgrade of IBM Storage Scale Erasure Code Edition and replace the step "6. Upgrade the IBM Storage Scale Erasure Code Edition software" with "Operations of node maintenance".

Manually online upgrade OS/driver for IBM Storage Scale Erasure Code Edition node

  1. Follow the upgrading steps 1 through 5 as described in Manual online upgrade of IBM Storage Scale Erasure Code Edition.

  2. Disable GPFS autostart by issuing the following command:
    mmchconfig autoload=no
  3. Upgrade the OS or the driver.
    Note: If you have Mellanox OFED driver installed, uninstall it before the OS upgrade, and install the new driver according to the upgraded OS version. You might need to restart the node after the upgrade.
  4. After the upgrade, with the system that runs the new kernel or driver, issue the following command to build the GPFS portability layer:
    mmbuildgpl
  5. Perform steps 7 through 9 as described in Manual online upgrade of IBM Storage Scale Erasure Code Edition.
  6. After all the nodes are upgraded, change the autoload config back to "yes".
    mmchconfig autoload=yes