Start of change

Backing up the object storage

All IBM Spectrum Scale™ Object Nodes and IBM Spectrum Protect™ client nodes must be available with the object file system mounted on each node when the backup is being created. The IBM Spectrum Protect server must also be available.

Store all relevant cluster and file system configuration data in a safe location outside your GPFS™ cluster environment. This data is essential to restoring your object storage quickly, so you might want to store it in a site in a different geographical location for added safety.

Follow these steps to back up the object storage manually:
Note: The sample file system used throughout this procedure is called smallfs. Replace this value with your file system name wherever necessary.
  1. Back up the cluster configuration information.
    The cluster configuration must be backed up by the administrator. The following cluster configuration information is necessary for the backup:
    • IP addresses
    • Node names
    • Roles
    • Quorum and server roles
    • Cluster-wide configuration settings from mmchconfig
    • Cluster manager node roles
    • Remote shell configuration
    • Mutual ssh and rsh authentication setup
    • Cluster UID
    Note: Complete configuration information can be found in the mmsdrfs file.
  2. Preserve disk configuration information.
    Disk configuration must also be preserved to recover a file system. The basic disk configuration information needed for a backup intended for disaster recovery is:
    • The number of disk volumes that were previously available
    • The sizes of those volumes
    To recover from a complete file system loss, at least as much disk space as was previously available is needed for restoration. It is only possible to restore the image of a file system onto replacement disks if the disk volumes available are of similar enough sizes to the originals that all data can be restored to the new disks. The following disk configuration information is necessary for the recovery:
    • Disk device names
    • Disk device sizes
    • The number of disk volumes
    • NSD server configuration
    • Disk RAID configurations
    • Failure group designations
    • The mmsdrfs file contents
  3. Back up the GPFS™ file system configuration information.
    In addition to the disks, the file system built on those disks has the following configuration information that can be captured using the mmbackupconfig command:
    • Block size
    • Replication factors
    • Number and size of disks
    • Storage pool layout
    • Filesets and junction points
    • Policy rules
    • Quota information
    • Other file system attributes
    The file system configuration information can be backed up into a single file using a command similar to the following:
    mmbackupconfig smallfs -o /tmp/smallfs.bkpcfg.out925
  4. Save the following IBM Spectrum Protect configuration files for each IBM Spectrum Protect client node in the same safe location outside of your GPFS cluster.
    /etc/adsm/TSM.PWD
    Contains the client password that is needed to access IBM Spectrum Protect. This file is present only when the IBM Spectrum Protect server setting of authentication is set to on.
    /opt/tivoli/tsm/client/ba/bin/dsm.sys and
    /opt/tivoli/tsm/client/ba/bin/dsm.opt
    Contains the IBM Spectrum Protect client configuration files.
  5. Issue the mmcesdr primary backup command to save the Swift configuration files for all protocol nodes.

    This command stores all of the Swift configuration data in an independent fileset created on the object storage file system. This must be done before creating the global snapshot to preserve and back up the configuration data.

  6. Back up the object storage content to an IBM Spectrum Protect server by running the mmbackup command:
    1. Create a global snapshot by running the following command:
      mmcrsnapshot <file system device> <snapshot name>.

      For example, create a snapshot that is named objects_globalsnap1 by running the following command:
      mmcrsnapshot smallfs objects_globalsnap1
    2. Create global and local work directories by running the following commands:
      mkdir -p /smallfs0/.es/mmbackupglobal
      mkdir -p /smallfs0/.es/mmbackuplocal
    3. Issue the following command to start the snapshot-based backup:

      mmbackup <file system device> -t incremental -N <TSM client nodes> \ -g <global work directory> \ -s <local work directory> \-S <global snapshot name> --tsm-servers <tsm server> --noquote

      The \ indicates the line wrap.
      For example:
      mmbackup smallfs -t incremental -N node1,node2 \
      -g /smallfs0/.es/mmbackupglobal \
      -s /smallfs0/.es/mmbackuplocal \
      -S objects_globalsnap1 -–tsm-servers tsm1 --noquote
      where
      -N
      Specifies the nodes that are involved in the backup process. These nodes must be configured for the IBM Spectrum Protect server that is being used.
      -S
      Specifies the global snapshot name to be used for the backup.
      --tsm-servers
      Specifies which IBM Spectrum Protect server is used as the backup target, as specified in theIBM Spectrum Protect client configuration dsm.sys file.

      There are several other parameters available for the mmbackup command that influence the backup process, and the speed with which its handles the system load. For example, you can increase the number of backup threads per node by using the -m parameter. For the full list of parameters available, see the mmbackup command.

    4. Issue the following command to remove the snapshot that was created in step 6a:
      mmdelsnapshot <file system device> <snapshot name>

      For example:
      mmdelsnapshot smallfs objects_globalsnap1
End of change