Recovering from an existing CCR backup

When you have a CCR backup file, you can recover a cluster by using that backup file. The CCR backup file contains all files that are committed to the CCR at the time the mmccr backup command is launched.

When you restore the cluster from a CCR backup file, the cluster runs based on the configuration in this backup file. Therefore, it is recommended to create a CCR backup file in regular intervals or at least after important cluster configuration changes.
Note: The mmsdrbackup user exit can be used to create CCR backups automatically every time the mmsdrfs file changes. The mmsdrfs file contains basic cluster and file system configuration information. For more information, see mmsdrbackup command in IBM Storage Scale: Command and Programming Reference Guide.
Perform the following steps to create a CCR backup:
  1. Issue the mmccr backup to create the CCR backup as shown in the following example: 
     mmccr backup -A /root/CCRbackup_20210708115924.tar.gz
CCR archive stored under: '/root/CCRbackup_20210708115924.tar.gz
    Note: In this example, the entire /var/mmfs directory is deleted on all nodes in the cluster to simulate this case.
  2. Issue the mmsdrrestore command with -F and -a options as shown in the following example to restore backup: 
     mmsdrrestore -F /root/CCRbackup_20210708115924.tar.gz -a
    The system displays output similar to this: 
    Restoring CCR backup
Verifying that GPFS is inactive on quorum nodes
Node node-25.localnet.com was not recovered because it is not available
Node node-24.localnet.com was not recovered because it is not available
When the unreached nodes are available, they can be recovered with:
  mmsdrrestore -p  node-21.localnet.com
mmsdrrestore: Propagating the cluster configuration data to all
  affected nodes.  This is an asynchronous process.
CCR backup has been restored
[root@node-21 ~]# Thu Jul  8 16:00:55 CEST 2021: mmcommon pushSdr_async: mmsdrfs propagation started
Thu Jul  8 16:01:00 CEST 2021: mmcommon pushSdr_async: mmsdrfs propagation completed; mmdsh rc=0
  3. Issue the mmsdrrestore command with the -p option to retrieve the backup on the non-quorum nodes. The -p option must specify a healthy quorum node from which the necessary files can be transferred. The mmsdrrestore command must run on the node to be recovered as shown in the following example:
     mmsdrrestore -p node-21
genkeyData1                                                                                                                                                                                                100% 3529     2.0MB/s   00:00
genkeyData2                                                                                                                                                                                                100% 3529     3.3MB/s   00:00
Thu Jul  8 16:03:46 CEST 2021: mmsdrrestore: Processing node node-24.localnet.com
mmsdrrestore: Node node-24.localnet.com successfully restored.

     mmsdrrestore -p node-21
genkeyData1                                                                                                                                                                                                100% 3529     4.0MB/s   00:00
genkeyData2                                                                                                                                                                                                100% 3529     4.1MB/s   00:00
Thu Jul  8 16:05:04 CEST 2021: mmsdrrestore: Processing node node-25.localnet.com
mmsdrrestore: Node node-25.localnet.com successfully restored.
  4. Verify the GPFS state at the cluster level by using the mmgetstate command as shown in the following example: 
     mmgetstate -a

 Node number  Node name  GPFS state
-------------------------------------
           1  node-21    active
           4  node-24    active
           5  node-25    active
           6  node-22    active
           7  node-23    active
Note: Based on the age of the used CCR backup file, it is possible that the cluster might be recovered to an old cluster configuration. It is recommended to take regular backup of CCR.