Technical data

Optimize Performance for IMDB Cache
Troubleshooting Oracle In-Memory Database Cache 4-15
finished for interval 15000ms with error.
14:56:36.12 Err : ORA: 5988: ora-5988-4724-fresher01606: Autorefresh number 9
failed for cache groups with interval 15000 ms after 10 retries.
The error message for thread ID 4724 shows that the change log table, TT_03_81799_L,
is missing. The introduction to "Autorefresh not refreshing cache at the specified
interval" on page 4-16 has a table entry that describes what to do in this situation.
Diagnosing autorefresh performance problems
You can use the ttTraceMon utility to diagnose autorefresh performance problems.
See "AUTOREFRESH tracing" on page 1-16.
TimesTen tracing severely impacts application performance and consumes a great deal
of disk space if trace output is directed to a file. When you are finished, reset tracing to
the default values.
Using SNMP traps for alerts about autorefresh problems
Enable SNMP traps to alert you when autorefresh problems occur.The SNMP traps
related to autorefresh include:
■ ttCacheAutoRefQueFullTrap
■ ttCacheIncAutoRefFailedTrap
■ ttCacheValidationErrorTrap
■ ttCacheValidationWarnTrap
■ ttCacheValidationAbortedTrap
See "Diagnostics through SNMP Traps" in the Oracle TimesTen In-Memory Database
Error Messages and SNMP Traps.
Optimize Performance for IMDB Cache
The following recommendations optimize performance for the IMDB Cache:
■ Pin the IMDB Cache meta tables and cache group base tables in the SGA. by
executing the "ALTER TABLE <table_name> CACHE" statement to indicate to the
Oracle database that these tables should be stored in the keep portion of the SGA
buffer cache. Pinning IMDB Cache tables in the SGA increases the probability that
any given datablock needed for a IMDB Cache refresh operation will be available
in the SGA when the refresh is performed and will not force a disk read. This
minimizes physical disk reads executed during TimesTen cache refresh
operations. For more information about Oracle buffer cache management, see
"Configuring and Using Memory" in the Oracle Database Performance Tuning Guide.
■ Pin IMDB Cache triggers into the shared pool using the dbms_shared_
pool.keep procedure. Pinning triggers into the shared pool for applications
where updates to the cache group base tables are infrequent keeps the trigger from
having to be reloaded and reparsed. This is not necessary for highly volatile tables
where the trigger will be executed frequently and will remain in the shared pool
under any circumstances.
■ Enable parallel query. For very large base tables with 10 million rows or more,
consider using the Oracle database parallel query facility. The primary join query
between the log table and the base table is the kind of query which the Oracle
database parallel query is designed to handle. When parallel processing is