While it is useful and relatively simple to collect only the log writer IO statistics, the IO subsystem administrators may want to see a fuller perspective. Through only Oracle's v$sysstat view, all Oracle read and write requirements can be determined for both the background processes and server processes, broken out by reads and writes, and also by IOPS and MB/s. Figure 8-16 shows the output from OraPub's OSM script iosumx.sql. The log writer portion of the script gathers the same information as shown in Figure 8-14.

Figure 8-16. Shown is OraPub's OSM iosumx.sql script, which gathers and displays all Oracle IO details at the instance level. This can be very helpful when working with IO administrators who need to understand Oracle's IO requirements.

When focusing on the application, identify the redo-producing SQL, and then find ways to reduce the volume of redo generated over the period of concern. The metric we are interested in is the redo generated per second; that is, the redo generation rate. When focused on the application, we are more interested in the bytes of redo generated, rather than the number of redo entries or the number of log writer writes.