How NGT Recover works
Uses the same DB2 resources that are used by the DB2 RECOVER utility
Uses any available change accumulation files as input to forward recovery
Uses index image copies that are made by NGT Copy, the IBM DB2 COPY utility, NGT Recover, or DSN1COPY
Uses Instant Snapshot copies that are made by NGT Copy
Uses encrypted copies that are made by NGT Copy
Uses cabinet copies that are created by NGT Copy as part of the Recovery Management solution
Uses Online Consistent Copies , which you can create if you have the Recovery Management solution
Supports a wide range of recovery strategies by using a variety of commands and options
During a recovery, NGT Recover automatically optimizes recovery processing by performing the following tasks:
Optimizing block sizes
Extracting index keys, when possible and requested, during table space recovery
Detecting and ordering activities for the proper positioning of stacked tape data sets, change accumulation files, and input and output copies containing all partitions of a partitioned space
Using a subtask to preallocate and open VSAM data sets required during recovery
Automatically falling back, without restarting, to a previous copy or Instant Snapshot, if an invalid or unavailable image copy or Instant Snapshot is encountered during recovery
Log data sets and table space data sets are dynamically allocated to the job. NGT Recover also dynamically allocates the input image copy data sets and required change accumulation files. The target table and index spaces are identified through IBM-compatible syntax, and the target DB2 subsystem is identified in the EXEC statement utility parameters. As an option, NGT Recover provides support for dynamic allocation of output image copies.
NGT Recover parses all NGT Recover command statements before initiating recovery. NGT Recover constructs a recovery plan and produces a report on this plan. Other options allow control over other aspects of the recovery. For example, you can specify the device type (SORTDEVT) and number of temporary data sets (SORTNUM) for sorting, or you can specify the MAXKSORT option for multitasking index key sorts and index rebuilds. You can further refine your index recovery using the KSORTSHARE option. You use the MAXLSORT option to specify concurrent log sorts and parallel MERGE phases. NGT Recover provides the ANALYZE ONLY option to view recovery resources and preview recovery plans without actually executing a recovery job. NGT Recover also provides a LOGSCAN command to gather information about the recovery and provide information that you can use to choose sort parameters.
NGT Recover uses the BMC BMCSORT technology for sorts. BMCSORT provides NGT Recover with more control of the sort process than external sort routines provide. This added control helps prevent memory-related problems during the sort process. NGT Recover allocates the amount of resources to each sort process based on the amount of work that NGT Recover determines that the sort process will perform. NGT Recover also dynamically detects excess available memory and allocates a percentage of it to the sort processes.
The recovery simulation feature, which requires a BMC Recovery Management for DB2 solution password, provides a non-destructive mechanism to exercise all recovery resources, including image copies, log records, and change accumulation files. This feature allows customers to ensure that all recovery resources are available to recover a table space or index without actually recovering it.
NGT Recover controls spaces by altering the DB2 status of the objects as necessary. Synchronization with BMC utility operations is provided through DB2 tables that are delivered and installed as part of NGT Recover.
NGT Recover does not use the DB2 subsystem buffers; however, it uses the DB2 subsystem to find information for the recovery process.