The Novell Transaction Tracking System™ (TTS™) protects simple database applications by backing out application-based transactions that are incomplete because of a system failure. For example, a banking application might require that all steps in a user’s session be complete before updating the database and ending the transaction. If a system failover occurred and the transaction did not complete all steps successfully, all of the steps would be rolled back to their condition before the transaction began. If all steps of the transaction are complete, the TTS replays the steps and completes the transaction successfully.
By default, TTS works with the Traditional file system. TTS cannot run on NSS and Traditional volumes at the same time on the same server. If you have both Traditional and NSS volumes on your system and you want to run TTS on NSS volumes instead of Traditional volumes, you must first disable TTS on all Traditional volumes, and then enable TTS on your NSS volumes.
For more information, see Using the Transaction Tracking System
in the NW 6.5 SP8: Server Operating SystemAdministration Guide.
NOTE:When working on OES 2 Linux, if your data requires content-level transaction tracking and you want to continue to control access to data with trustees, you can use NCP volumes on Linux Reiser or Ext3 file systems. Configure the journaling mode for the file system to the Journaling level, which provides content-level transaction tracking. For information about creating and managing NCP volumes on Linux, see Managing NCP Volumes
in the OES 2 SP2: NCP Server for Linux Administration Guide.
To enable TTS on an NSS volume, enter the following at the server console prompt:
nss /Transaction=volname
To disable TTS on an NSS volume, enter the following at the server console prompt:
nss /NoTransaction=volname
To disable TTS on all Traditional volumes, add disable tts to the autoexec.ncf file.
If a system failure occurs, TTS replays the TTS log file to recover lost application-based transactions as part of the failover recovery.
Before replaying the TTS log to recover lost transactions, TTS checks the integrity of the transaction file blocks in the log. If TTS detects no problems, the replay continues.
If TTS detects a bad block, it does not use the log file. TTS creates a new TTS log file to use as the system continues to operate.
IMPORTANT:Although the system continues to run, data might have been lost.
TTS renames the corrupted TTS log file to indicate the date and time that it detected the problem. For example, it renames
volume:\UXaction.log
to
volume:\XLog yyyymmddhhmmss.log
Corrupt TTS log files are relatively large, about 10 MB. TTS stores the corrupt TTS log file in the root directory of its transactioned volumes. If users report finding TTS log files in the root directory of their volumes, it indicates that a problem occurred and that TTS detected it.
NOTE:If a system failure causes corruption in the transaction log file, please report to Novell Support the network circumstances that lead to the corruption of the TTS log file. We want to examine the event to see if we can establish why the corruption might have occurred.