Section 8.5. Exercises

---

8.5 Exercises

1. If the thought experiment in Section 8.3.7.3 had specified that each of the 100 users had simultaneously requested one second of CPU capacity, what would the resource profile look like for the 100-second interval?

2. Using V$SQL consumes less server capacity than using V$SQLAREA . Use extended SQL trace data to explain why.

3. Experiment with vprof . Try the following time scope experiments:

   • Mark t and then wait several seconds before executing the first database call in the targeted session.

   • Mark t immediately preceding the first database call in the targeted session.

   • Mark t in the midst of a long-running SQL statement's execution in the targeted session.

   • Mark t ₁ immediately after the conclusion of the final database call in the targeted session.

   • Mark t ₁ several seconds after the conclusion of the final database call in the targeted session.

   • Mark t ₁ in the midst of a long-running SQL statement's execution in the targeted session.

4. Describe the challenges that prevent us from constructing a utilization report that shows utilization by Oracle kernel wait event. For example, imagine a system with the following three indications :

   • There have been 1,000,000 seconds of waiting for disk I/O events since instance startup.

   • The instance startup occurred 500,000 seconds ago.

   • The system has six disks, so how much can we know about per-disk utilization since instance startup?

   What conclusions can you draw from these observations?

---

Top