Changing the secondary site

This topic lists the steps to follow when the secondary site stops functioning.

Complete the following steps if the secondary site stops functioning:

  1. Create a GPFS independent fileset on the new secondary site.
  2. Copy the data on the primary site to the new GPFS fileset by using ftp or scp. This copying data is known as outband trucking.
  3. Run the following command to establish a new secondary site for the primary site:
    mmafmctl Device changeSecondary-j FilesetName    
    --new-target NewAfmTarget [ --target-only |--inband ]          
    [-s LocalWorkDirectory]

    The --inband option is used depending on the method that is being used to truck data. If you want to change the mount path or the IP address in the target path, you can use the --target-only option. The new NFS server must be in the same home cluster and must be of the same architecture as the existing NFS server in the target path. The --target-only option must not be used if the home is changed completely, or if you want to change the protocol while you are using the same home cluster. While you are using the same secondary path, if you want to change only the target protocol from GPFS to NFS and vice versa, you can use mmafmctl changeSecondary command.

  4. Convert the GPFS fileset on the new secondary site to a secondary and set the primary ID by running the mmchfileset command or the mmafmctl command with convertToSecondary option.

    Ensure that the NFS export on the secondary site is accessible on all the gateway nodes on the primary cluster if NFS is used for defining AFM target on the primary site. If the GPFS protocol is used for the target, the secondary file system must be mounted on all gateway nodes on the primary site.

    After the primary and secondary sites are linked, the psnap0 queued from the primary fileset is played on the secondary fileset. The new secondary site is now linked to this primary site. Ensure that the primary file system is mounted on all gateway nodes and new secondary file system is mounted on all the nodes in the secondary cluster.