In the previous article (Analyze Oracle event information), there was mention of what I/O latencies are "acceptable" or not. Often, expectations for "acceptable" are defined in comparison to the I/O latencies experienced with other databases which are not based on Delphix. So, if a VDB has I/O latencies of N ms and another database not using Delphix also has I/O latencies of N ms, then even though we may wish N were lower, N might be considered acceptable because that is the expectation for other non-production systems.
However, measuring I/O latencies between two databases can be very difficult due to many factors, such as generating the exact same workload, the effects of Oracle buffer caching, and the impact of other workloads on the testing.
Delphix developed the "sanity check" scripts to attempt to generate the same workload consistently, so that I/O latencies can be compared, either at different times within the same database, or between two completely different databases.
The "sanity check" script is called "sc.sql" and is available for download HERE. If you experience difficulty downloading this script from Google Drive, then please contact someone through the Delphix Community Forums to email the script to you directly.
Please note that the "sanity check" script is expected run under a SYSDBA account, because it will temporarily create a tablespace, create a table and an index within that tablespace, perform standard workload, and then drop that temporary-use tablespace. The script will prompt you for the location of the tablespace and the size in Mbytes.