Resynchronization Function

The purpose of the resynchronization function is to complete updates when there has been an error, such as a telecommunications line going down, during sync point processing before all updates within a transaction have been committed or rolled back. Resynchronization tries to ensure all protected resources in the transaction are either committed or rolled back, therefore maintaining the data's integrity. Resynchronization uses data logged by the SPM and by the resynchronization manager to determine what actions the resynchronization function needs to take. This data is recorded on dual CRR logs to protect against losing the CRR log data. In the case of protected resources, resynchronization tries to communicate with the unavailable protected resource manager. In the case of protected conversations, resynchronization tries to communicate with the other CRR recovery server, which resides on the same system as the protected conversation partner. The other CRR recovery server then tries to communicate with the unavailable protected resource manager.

If resynchronization cannot initially complete the updates, resynchronization issues a timed wait. After a specified time interval, which is called the timed wait interval, the resynchronization function leaves the timed wait and automatically attempts to complete the updates. This is called the automatic periodic retry of resynchronization.

If the CRR operator knows the protected resources will not be available for a certain period of time, the CRR operator can enter the CRR SUSPEND command. The CRR SUSPEND command stops the automatic periodic retry of resynchronization. This saves computer resources from being unnecessarily consumed while trying to resynchronize a protected resource that is not available. The CRR operator can use the CRR RESUME command to restart the automatic periodic retry of resynchronization.

For more information, see Stop and Restart Automatic Periodic Retry of Resynchronization and Bypass and Change Timed Wait Interval.

The resynchronization function also reports an irrecoverable error if the resynchronization function cannot complete the transaction with integrity.

When all protected resources and protected conversation partners become available, the resynchronization function completes the coordinated transaction. Resynchronization may not be able to complete the updates for a coordinated transaction because the protected resources or protected conversation partner may not become available for a long period of time. If resynchronization cannot complete the updates for this coordinated transaction, CRR provides the CRR RESYNC command for a CRR operator to manually complete the resynchronization.

Resources that participate in CRR should provide commands to complete the coordinated transaction (with respect to its resources). For example, SFS provides the following commands for the SFS operator to manually complete a coordinated transaction:
  • QUERY PREPARED–Displays information about SFS prepared work and forced work in a coordinated transaction.
  • FORCE PREPARED–Forces SFS prepared work to be committed or rolled back.
  • ERASE LUNAME–Erases the history of previously forced SFS prepared work from the SFS logs.

See Participation in CRR (SFS only) and Problem Management for more information about operator intervention to complete resynchronization.