RDF System Management Manual

Managing RDF
HP NonStop RDF System Management Manual—524388-003
5-12
File Recovery on the Primary System
Volume Recovery Failure
RDF cannot recover from a TMF subsystem crash if TMF cannot successfully perform
volume recovery. After the TMF crash has been resolved, you must perform the
following tasks:
1. Resynchronize the primary and backup databases, following procedures explained
in sections 6 and 7.
2. Reinitialize RDF.
File Recovery on the Primary System
A file recovery operation occurs whenever a TMFCOM RECOVER FILES command is
issued at the primary system. A simple file recovery operation does not affect RDF nor
does it require database synchronization. A file recovery operation to a timestamp, a
first purge, or TOMATPOSITION, however, does require you to stop RDF, reinitialize,
and resynchronize the affected files.
The file recovery TOMATPOSITION is a special usage that achieves synchronization
itself. If the primary system has failed, you have executed an RDF takeover operation
on your backup system without RDF/ZLT, and you have subsequently brought you
primary system back online, you can resynchronize the database on your recovered
primary system with file recovery TOMATPOSITION. When the takeover has
completed on your backup system, RDF normally logs an RDF event 888. This event
provides you with a master audit trail sequence number and relative byte address that
you can use for file recovery TOMATPOSITION on your recovered primary system.
The result of this operation puts the database on your primary system into
synchronization with the database on your backup system at the time when the
takeover operation completed. If you started application processing on your backup
system after the completion of the takeover operation, you then need to configure a
new RDF subsystem to replicate all changes made to the database on your backup
system to the database on your primary system.
File Recovery on the Backup System
You are encouraged to take online dumps on your backup database on a regular basis
for the following reasons:
•
If you have lost your primary system and have taken over on your backup system,
the online dumps can be used for any type of file recovery operation provided the
redo end point is located after all audit data that was generated during the RDF
takeover. For example, a file recovery to a timestamp must be to a timestamp after
the time when the RDF takeover completed.
•
If RDF is running from your primary to your backup system and you lose one or
more disks on your backup system, you should stop the RDF updating, perform a
simple file recovery on the backup system to recover the files on the affected disks,
and then restart RDF updating.