FP5: First occurrence data collection supports new collection types and collection triggered by user-defined thresholds

First occurrence data capture (FODC) collects diagnostic information about your DB2® data server when a problem occurs. In Version 9.7 Fix Pack 5 and later fix packs, FODC supports additional manual collection types and supports triggering automatic diagnostic data collection when a user-defined threshold condition is exceeded.

You can invoke FODC manually with the db2fodc command whenever you suspect a problem, or FODC can be invoked automatically whenever a predetermined scenario is detected. For manual FODC, the -hang and -perf parameters, which have been available since before Fix Pack 5, collect end-to-end diagnostic data. However, these parameters often collect more diagnostic data than is required for troubleshooting, at the cost of additional processor usage and disk space requirements. On a system that is already resource constrained, any additional demand on resources might not be acceptable, even if it is required to collect diagnostic data. The following new FODC collection types, which apply to more specific performance problem scenarios, help address the problem of additional overhead:

-cpu

If you observe unusually high processor utilization rates, a high number of running processes, or high processor wait times, you can use the -cpu parameter to collect processor-related performance and diagnostic data.

-memory

If you determine that there is no free memory available, that swap space is being used at a high rate, or that excessive paging is occurring or if you suspect a memory leak, you can use the -memory parameter to collect memory-related diagnostic data.

-connections

If you determine that there is a spike in the number of applications in the executing or compiling state or that new database connections are being denied, you can use the -connections parameter to collect connection-related diagnostic data.

With the new -detect parameter for the db2fodc command, you can now specify your own threshold rule for a specific condition and trigger diagnostic data collection when the condition is exceeded. The -detect parameter supports detecting trigger conditions one or several times, at a regular interval that you specify. If the number of times that the threshold condition is detected matches the specified value, diagnostic data collection is triggered. Other options for detecting threshold conditions are also available, such as how many iterations of threshold detection and diagnostic data collection are performed, and how long threshold detection continues.

You can also configure the new user-defined thresholds to only detect and a specific problem condition but not collect diagnostic information. If a problem condition is detected, only a log record is added to the db2diag log files.