Migration when you use the IBM Storage Scale mmbackup command and IBM Storage Protect

Edit online

The following information describes how the IBM Storage Archive EE eeadm migrate --mmbackup command works with the IBM Storage Scale mmbackup command and the IBM Storage Protect backup-archive client.

Note: There are configuration considerations for the IBM Storage Protect backup-archive client when it runs on the same node as IBM Storage Archive EE.

When the IBM Storage Archive EE product software is installed, it includes a customized version of the IBM Storage Protect client. Continue to use that version of the IBM Storage Protect software to avoid version conflicts.

  • If you were using the IBM Storage Protect backup-archive client before you installed IBM Storage Archive EE, be aware that the backup-archive client is overwritten by IBM Storage Archive EE, and the configuration files (dsm.sys and dsm.opt) must be adjusted. Always use the version of the IBM Storage Protect backup-archive client that comes with IBM Storage Archive EE. IBM Storage Archive EE version 1.3.0.6 comes with a portion of IBM Storage Protect version 8.1.10 already installed on the IBM Storage Archive EE node for the use of IBM Storage Archive EE.
  • If you were using IBM Storage Archive EE first, and then want to enable the IBM Storage Protect backup-archive client, do not install the backup-archive client that comes with IBM Storage Protect. And you still need the configuration file to point to the correct IBM Storage Protect server.
  • In either case, you need to separately purchase the IBM Storage Protect backup-archive client license for using the function on the IBM Storage Archive EE node, even though there is not a separate installation process
If you need to manually modify the HSM configuration files (dsm.sys and dsm.opt) for the use of the backup-archive client, you must place the following lines at the beginning of the configuration file. Otherwise, IBM Storage Archive EE cannot migrate or recall files.
HSMBACKENDMODE TSMFREE
ERRORLOGNAME /opt/tivoli/tsm/client/hsm/bin/dsmerror.log
MAXRECALLDAEMONS 64
MINRECALLDAEMONS 64

When the IBM Storage Protect backup-archive client is used for a file system that is managed by IBM Storage Archive EE, it provides additional protection for the data. It can retrieve an older version of a file after an unexpected modification of archived data, or the unexpected deletion of a file. To create the backup files, configure the IBM Storage Protect backup-archive client on an IBM Storage Scale node. Then, direct the mmbackup command to transfer copies of the files to an IBM Storage Protect server.

When you configure IBM Storage Protect to create backup copies on a file system that is managed by IBM Storage Archive EE, you must schedule migration processing to start after the files are backed up. If migration processing starts before the files are backed up, unexpected file recalls from the IBM Storage Archive EE tapes occur when the mmbackup process accesses the migrated files to back them up. The file recalls cause backup processing to take longer to complete. Using the --mmbackup option of the IBM Storage Archive EE eeadm migrate command ensures that the order of the backup and migration processing is correct. The option checks whether a file was already backed up by the mmbackup command. It does not migrate a file that is not yet backed up.

Before you use the eeadm migrate --mmbackup command, you must be familiar with the IBM Storage Scale mmbackup command. For more information about the IBM Storage Scale mmbackup command, see the topics Protecting data in a file system using the mmbackup command and Back up your data in the online IBM Storage Scale documentation, which is located in the IBM Documentation.

How the eeadm migrate --mmbackup command works

The --mmbackup option of the eeadm migrate command works with the IBM Storage Scale tsbuhelper program to prevent files that are not yet backed up, or files that were modified since they were backed up, from being migrated.

The most current information about GPFS files that are not yet backed up, or that were modified since they were backed up, is contained in an IBM Storage Scale mmbackup shadow database.

The --mmbackup option of the eeadm migrate command requires that you specify the path to the IBM Storage Scale shadow database (--mmbackup <shadow_db_path>) so that file migration lists can be compared to file backup dates and file modification dates. Any files that require backup processing are not migrated.

The path to the shadow database is defined by the IBM Storage Scale mmbackup command. It can also be specified by the MMBACKUP_RECORD_ROOT environment variable for the mmbackup command. (Do not confuse the IBM Storage Scale mmbackup command with the --mmbackup option of the eeadm migrate command.)
  • When the path to the shadow database is defined by the IBM Storage Scale mmbackup command, it requires a parameter of either Device or Directory. If the Directory parameter is used, the directory is the path for the mmbackup shadow database. If the Device parameter is used, the mount point where the file system is mounted is the path for the mmbackup shadow database.
  • When the environment variable MMBACKUP_RECORD_ROOT is used, the variable specifies the path to the shadow database.
For more information, see mmbackup command in the IBM Storage Scale documentation on the IBM® Documentation.

Using the --mmbackup option

The eeadm migrate --mmbackup command can be manually started, or it can be started with a policy that is run with an mmapplypolicy command. To start the option with a policy, add the --mmbackup <shadow_db_path> command to the rule that defines the external pool. The <shadow_db_path> specifies the path of the mmbackup shadow database. If the mmbackup command is run for a file system with the path of /ibm/gpfs, then the <shadow_db_path> is /ibm/gpfs, as shown in the following example.
RULE EXTERNAL POOL 'md1'
EXEC '/opt/ibm/ltfsee/bin/eeadm' 
OPTS '-p pool1@lib2 --mmbackup /ibm/gpfs'
Be aware of the following conditions when you use the --mmbackup option.
  • If all target files for migration get filtered out because they are not yet backed up, no files remain as the target files. If the input for the migrate command is empty, the migrate command or the mmapplypolicy command might fail with errors.
  • If IBM Storage Scale is running the mmbackup command, and an mmapplypolicy starts the IBM Storage Archive EE eeadm migrate --mmbackup function simultaneously, the migration might fail. Do not run the two processes at the same time.