Reclone process

• Replicated objects cannot be resynchronized successfully. This could occur if a node was forced to TRACKING state when the other node was unavailable but had existing entries in the OTL. One way to reconcile the errors is to choose the source node and reclone all replicated objects from that node to the copy node.
• One or both Db2 Mirror nodes are permanently suspended. For information on permanent suspend, see Permanent suspend.
• The primary node has been TRACKING for an extended period of time.

There are three techniques for performing a reclone:

1. Hardware reclone

   Hardware reclone uses storage replication technology to make the copy node an identical copy of the source node. This form of reclone will replace the operating system, all licensed products and applications, as well as all data on the copy node.

   Hardware reclone may be appropriate if there are unrecoverable errors involving non-replicated programs or data on the copy node, or if it is desirable to start with an exact image of the source node.

2. Software reclone

   A software reclone uses the rules within the replication criteria list (RCL) to determine which objects on the source node are replicated. Those objects are saved on the source node and restored on the copy node, ensuring that all replicated objects are identical between the two nodes.

   Software reclone is strictly focused on replicated objects and has no impact to the operating system, licensed products, or other non-replicated objects that are established on the copy node.

   Software reclone includes preparation steps that highlight any situation that should be understood before proceeding with the reclone. These situations are explained in the software reclone section.

   A software reclone can be used if the systems are configured using internal storage or external storage without FlashCopy® or replication services. A software reclone may also be preferred if the source and copy nodes have significantly diverged in their non-replicated objects, the products they have installed, or the level of the operating system fixes or operating system version. A software reclone is also an option to return to active replication in the event that a node is marked permanently suspended due to main store loss. A software reclone will ensure that all replicated objects are equivalent while not replacing non-replicated objects or system objects on the copy node.

3. Reclone a single replicated object

   Recloning a single replicated object is used to copy an object and its network of dependent replicated objects from one Db2 Mirror node to the other node. The object being recloned and all dependent replicated objects are identified and deleted on the copy node before being replaced by versions of the objects saved from the source node.

   Recloning a single object can be used to correct replication problems such as recovering from a failure that was recorded for an object in the OTL. In addition, since recloning a single object relies upon save and restore technology, it may be appropriate to overcome a situation where a large percentage of active rows are out of sync.

   

When recloning both SYSBAS and database IASPs, a combination of hardware and software reclone is supported. If a hardware reclone is used for SYSBAS, then all database IASPs must also be recloned using hardware reclone. If SYSBAS is recloned using software reclone, then either software or hardware reclone can be used to reclone any database IASPs.

After a hardware reclone of *SYSBAS, the mechanisms used to make any IFS IASPs highly available is not automatically reconfigured. For information about making IFS IASPs highly available, see Adding IFS IASPs to Db2 Mirror.