What disaster recovery (DR) options are available for the Db2 pureScale Feature?

There are various disaster recovery options that are currently supported for Db2® pureScale®. The variety of the options allows for different solutions to meet the needs of all of users, and the choice is largely dependent on your business needs.

The different disaster recovery options are:
  • Geographically dispersed pureScale cluster (GDPC)
    • This option is available for businesses that require an active/active DR solution with ZERO down time in site failure. Essentially, this is a split pureScale cluster stretched over several kilometers at most. It can be deployed in different rooms or floors within the same building, between buildings, or cities within public transportation distance. The limitation of the distance is to ensure performance delay due to distance is within the acceptable range for a reasonable workload.
    • For more information on GDPC, refer to Geographically dispersed Db2 pureScale cluster (GDPC).
  • Db2 HADR support with pureScale
    • This option combines the continuous availability of Db2 pureScale with the robust disaster recovery capabilities of HADR providing an integrated zero data loss (i.e. RPO=0) disaster recovery solution. All four sync modes are supported as well as Log Spooling and Time Delayed Apply. It can span thousands of kilometers between the two sites depending on the sync mode.
    • For more information on Db2 HADR support with pureScale, refer to High availability disaster recovery (HADR).
  • Storage replication
    • This option refers to data copying technologies that can constantly update a second copy of a storage volume to match changes that are being made to a source storage volume. It is an active/passive solution where transaction runs against the primary site only. The DR site must be defined with the same instance, topology, and storage structure. The storage associated with the database’s data and logs must be mirrored. This is available in both synchronous mode (where it guarantees no data loss) or asynchronous mode. Depending on the storage vendor solution, there might be a maximum distance imposed. Check with storage vendors for details.
  • Q Replication / InfoSphere Change Data Capture (CDC)
    • This option provides a high-throughput, low latency replication solution for sites that can be thousands of kilometers away. Q Replication is logical in nature, in that the underlying physical aspects of the database do not come into play during replication. Only changes to specific tables are captured and they are replayed using SQL on the DR site. It is asynchronous and it is therefore possible that changes that get made to a primary site are lost if something happens to that site and failover to the DR sites has to take place. Active/active, bi-directional support will require a strategy for conflict avoidance and resolution.
  • Log shipping
    • This option is a "user-managed" active/passive DR solution. It involves copying whole log files to a standby system from a primary system and rolling forward through those log files on the standby. Only log files that are no longer actively being written to on the primary can be used.
    • For more information on log shipping, refer to High availability through log shipping.