Fatal problems generate error messages with severity levels 19 and higher. They break the user’s connection to Adaptive Server (some of the higher severity levels shut down Adaptive Server). To continue working, the user must restart the client program.
When a fatal error occurs, the process freezes its state before it stops, recording information about what was happening. It is then killed and disappears.
When the user’s connection is broken, he or she may or may not be able to reconnect and resume working. Some problems with severity levels in this range affect only one user and one process. Others affect all the processes in the database. In some cases, it will be necessary to restart Adaptive Server. These problems do not necessarily damage a database or its objects, but they can. They may also result from earlier damage to a database or its objects. Other problems are caused by hardware malfunctions.
A backtrace of fatal error messages from the kernel is directed to the error log file, where the System Administrator can review it.
