Summary information |
|
|---|---|
Default value |
500 |
Range of values |
0–2147483647 |
Status |
Dynamic |
Display level |
Comprehensive |
Required role |
System Administrator |
cpu grace time, together with time slice, specifies the maximum amount of time that a user process can run without yielding the CPU before Adaptive Server preempts it and terminates it with a time-slice error. The units for cpu grace time are time ticks, as defined by sql server clock tick length. See “sql server clock tick length” for more information.
When a process exceeds cpu grace time Adaptive Server “infects” it by removing the process from the internal queues. The process is killed, but Adaptive Server is not affected. This prevents runaway processes from monopolizing the CPU. If any of your user processes become infected, you may be able to temporarily fix the problem by increasing the value of cpu grace time. However, you must be sure that the problem really is a process that takes more than the current value of cpu grace time to complete, rather than a runaway process.
Temporarily increasing the cpu grace time value is a workaround, not a permanent fix, since it may cause other complications; see “time slice”. Also, see Chapter 3, “Using Engines and CPUs,” and “Execution task scheduling” on page 29 of the Performance and Tuning Guide for a more detailed discussion of task scheduling.
