Copyright © Cloud Software Group, Inc. All Rights Reserved |
To develop a customized plan, consult TIBCO Support. The numbers in this diagram correspond to the following steps:
1. Prime the continuous backup process with an initial set of page images using hrntlbps (backup.000).
The MetaStor (segment 0) must be included in your backup. The MetaStor identifies in which segment each table resides. Always run the hrnbrptr utility to ensure valid database pointers.
2. All page update images are saved in JOURNAL1 and JOURNAL2. When one fills, the Data Object Broker automatically starts spin01 or spin02 to save the journal data.
− For Windows, spin01 and spin02 start a C program called spinctrl. Spinctrl copies journal update images into journal accumulation files called spinout.001, spinout.002, spinout.003, and so on
− For Solaris, spin01 and spin02 copy journal update images into journal accumulation files called spinout.001, spinout.002, spinout.003, and so on
3. The defined limit of the SPINLIM variable is checked each time spinctrl (for Windows) or spin01/spin02 (for Solaris) is started. When the number of spinout files reaches the limit, spin03 consolidates them into one file, spinout.000. All other spinout files are deleted. The hrntlmrg utility keeps only the most current page images.
When spin03 is run again, it merges spinout.000 with any new spinout files to create a new spinout.000.
4. The information in JOURNAL1 and JOURNAL2 is merged with existing spinout files by executing the spin04 batch file/script. The Data Object Broker must be shut down for spin04 to be successful.
5. Run spin05 on a regular basis to consolidate all updated journal page images into the continuous backup file backup.000. After spin04 and spin05 have run, backup.000 is identical to a complete TIBCO Object Service Broker backup of all segments using hrntlbps.TIBCO Object Service Broker for Open Systems Utilities for information on using the hrntlbps and hrntlmrg utilities.
Copyright © Cloud Software Group, Inc. All Rights Reserved |