Replacing both cluster caching facilities

You can replace both cluster caching facilities using a rolling upgrade technique which ensures that your Db2® pureScale® instance will not experience an outage.

Before you begin

You must have root user authority to perform this task.

Ensure that you know which cluster caching facility is currently the primary, so that you don't inadvertently stop it. As well, the secondary cluster caching facility needs to be in PEER state. You can accomplish both of these items by querying the DB2_CF administrative view, as follows: SELECT * FROM SYSIBMADM.DB2_CF

Procedure

Use the following steps to replace both cluster caching facilities (CFs):

  1. Stop the secondary cluster caching facility using the command: db2stop CF CF-identifier
    For more information about stopping a cluster caching facility, see Stopping a cluster caching facility.
  2. Stop the instance on the secondary cluster caching facility to ensure no other Db2 processes are running, using the command: db2stop instance on host-name
  3. Drop the cluster caching facility using the command:db2iupdt -drop -cf host-name instance-name
    For more information about dropping a cluster caching facility, see Dropping a member or cluster caching facility.
  4. Add the new host using the command: db2iupdt -add -cf host-name:net-name instance-name
    For more information about adding a cluster caching facility, see Adding new members or an additional CF.
  5. Find out the ID for the newly added cluster caching facility using the statement: SELECT * from SYSIBMADM.DB2_CF
  6. Start the newly added cluster caching facility using the command: db2start CF CF-identifier
    For more information about starting a cluster caching facility, see Starting a cluster caching facility.
  7. Once the new cluster caching facility (now the secondary cluster caching facility) reaches PEER state, stop the primary cluster caching facility as in steps 1 and 2.
    Db2 cluster services will fail over the primary role to the secondary cluster caching facility.
  8. Repeat steps 3 - 6 to replace the old primary cluster caching facility and restart it as the secondary cluster caching facility.