Figure 8-13 also shows that over the 60-second trace, the log writer made 18 getrusage calls. Back in Chapter 5, I explained that Oracle's system time model, introduced with Oracle Database 10g (v$sess_time_model and v$sys_time_model), gathers Oracle process CPU consumption by having the processes ask the operating system for their process's CPU consumption through a getrusage call, and then places that information into Oracle internal structures. As performance specialists, we can see this CPU consumption via the system time model views. Oracle states the CPU time will be accurate within 5 seconds; my tests have shown the actual time to be typically just over 6 seconds. In the 60-second interval shown in Figure 8-13, since the log writer background process made 18 getrusage calls, the log writer background process CPU consumption information should be accurate to within around 3.3 seconds.

Figure 8-13. Shown is a 60-second interval operating system summary trace of the log writer background process. Within the 60-second interval, Oracle issued 680 write calls (pwrite64), nearly 3,000 gettimeofday calls, and over 2,200 times calls. During this interval, the average write call reported from this operating system utility was 925 ms!

Just how much IO the log writer background process is generating can be determined from Oracle's instance activity view, v$sysstat. Most IO administrators want Oracle's IO requirements from either an IO operations per second (IOPS) perspective or a megabytes per second (MB/s) perspective. In effect, what they will need to do is ensure the IO subsystem has the capacity to process both Oracle's read and write requirements within acceptable service levels. While every system has its own service level requirements, common service levels are for Oracle reads to be satisfied within 10 ms and writes to be satisfied within 5 ms.