RDF System Management Manual for H-Series RVUs (RDF 1.8)

5. Perform each operation on Enscribe files on the backup system and the corresponding
operation on the primary system.
6. Finally, resume application processing.
Resynchronizing Databases
There are two ways of resynchronizing your primary and backup databases: offline and online.
With offline resynchronization you must first stop your applications and TMF on the primary
system. With online resynchronization, however, you can resynchronize entire databases, selected
volumes, a single volume, or individual tables and files while your applications continue to run
on the primary system.
The remainder of this chapter describes how to do offline resynchronization. For information
about online resynchronization, see Chapter 7 (page 155).
For NonStop SQL/MP or NonStop SQL/MX tables with sys-key or cluster-keys, you must do
offline synchronization, as well as for Enscribe unstructured files or any kind of entry-sequenced
file.
To resynchronize the primary and backup databases, you need to make all backup database files
or tables logically identical to the primary database files or tables when there is no audit data to
be processed for the files or tables. If you know which files or tables are not synchronized,
resynchronize the databases only on the volumes that contain those files or tables.
There is no audit data to be processed for a volume at the following times:
• Immediately after TMF has been started for the very first time and no applications have
been started yet
• When the RTD time is zero for the volume’s updater process, and no audit data is being
generated by any application while the files or tables are being duplicated
• When TMF is stopped (without the ABRUPT option)
Make sure the primary and backup databases are synchronized if any of the following should
occur:
• A TMF file recovery operation to a timestamp or to first purge occurs, after which only the
affected database tables or files need to be resynchronized.
• Asterisks (****) appear in the final column of the STATUS RDF display, indicating that an
updater process has experienced an unexpected file-system error.
NOTE: Resynchronization is not always necessary, however, after a file-system error in
an RDF process. For example, an updater process reporting an error 122 will restart.
• An additional volume is configured into an existing RDF configuration while TMF is running,
in which case database tables and files on the updater’s primary volume might not be
synchronized with the corresponding tables and files on the backup volume. In this case,
you must only resynchronize that single volume.
• TMF is deleted and reconfigured, or RDF is reinitialized, after a STOP RDF command is
issued at the primary system.
If RDF fails and reports an event whose recovery text indicates that database resynchronization
is required, you must resynchronize the backup and primary databases.
Resynchronizing Entire Databases Offline
To resynchronize an entire database offline, you must stop TMF, initialize RDF to the TMF
shutdown timestamp, and then copy the complete database from the primary system to the
backup system.
Resynchronizing Databases 153