Successful Automatic Recovery Tips

c-tree Plus made available an extremely affordable multi-user data handling system for thousands of applications around the world. Most are still in use today. Our FPUTFGET technology allowed even the smallest of environments to share valuable data. Those systems can suffer from two important constraints, however.

First, FPUTFGET saturates quickly as the number of concurrent users increases. System resources become a bottleneck as data is constantly pushed to disk for integrity. More importantly, FPUTFGET doesn't always handle recovery from power failures, or other outages well. As a result, indexes frequently must be rebuilt following even a minor application crash.

Improve your recovery times with the c-tree Server! The c-tree Server guarantees your data is always secure when using transaction processing. Even in the most devastating of system failures, the c-tree Server is almost always capable of restoring your data and index files back to a consistent state, protecting your valuable time and data. The c-tree Server always undergoes automatic recovery on every startup validating your data files.

Occasionally, it appears that automatic recovery fails. Frequently, this is only a minor problem with certain files. Many times, these files are temporary files, for example, that are not required for automatic recovery to proceed. The c-tree Server makes available options to avoid these files and gain additional details on why an automatic recovery could not proceed, allowing you to still bring your system up in a reliable and consistent manner. In addition, other performance options can speed automatic recovery operations. Read on for several c-tree Server options that can put you at complete ease concerning your data integrity.

c-tree Server Automatic Recovery Options

SKIP_MISSING_FILES <YES | NO>

This keyword is available for special c-tree Server startup conditions. If a user file required by the c-tree Server during automatic recovery was deleted, an error 12 might be returned and the c-tree Server would not continue. By adding SKIP_MISSING_FILES to the default ctsrvr.cfg file, the error will be logged and the c-tree Server will successfully start up. However, SKIP_MISSING_FILES is not recommended as a permanent setting. Deleting files under transaction processing control adversely affects database integrity.

Default: NO

RECOVER_DETAILS <YES | NO>

This keyword sends detailed information about the Server automatic recovery process to the server console. The time spent for each phase of automatic recovery in addition to the number of transactions processed for each phase is provided. This keyword adds minimal overhead to c-tree Server operations.

Default: NO

RECOVER_FILES <number of files | NO>

RECOVER_FILES makes it possible to set separate limits on the number of files used during automatic recovery and regular operations. The reason automatic recovery may require more files than regular operations is that during recovery files opened stay open until the end of recovery. RECOVER_FILES takes as its argument the number of files to be used during recovery. If this is less than the number used during regular operation specified by the FILES keyword, the number of recovery files is set equal to the regular files and the keyword has no affect. If the number of recovery files is greater than the number of operational files, the number of files is adjusted downward at the end of automatic recovery freeing memory used by the additional control blocks, about 900 bytes per logical file.

Default: NO

RECOVER_MEMLOG  <# of logs to load | NO>

Loads one or more transaction logs into memory during automatic recovery to speed the recovery process. The argument for this keyword specifies the maximum number of memory logs loaded into memory during automatic recovery.

Default: NO

Note: Always check the c-tree Server status file, CTSTATUS.FCS, for complete information about your system, including any messages generated by the automatic recovery mechanism on server startup.