Database Reference
In-Depth Information
In the output following, dc_sequences shows 51.6% success in getting the next sequence number. This means
that 50% of the time the sequences cache was empty and had to be refreshed:
Script: MVRACPDnTap_rowcache.sql
SELECT inst_id INT,
parameter,
SUM(gets) gets,
SUM(getmisses) misses,
100 * SUM(gets - getmisses) / SUM(gets) pgets,
SUM(modifications) updates
FROM gv$rowcache
WHERE gets > 0
HAVING SUM(modifications) > 0
GROUP BY inst_id,
parameter
ORDER BY inst_id,
6 DESC; d
INT PARAMETER GETS MISSES PGETS UPDATES
---- ----------------------- -------- ------- ----- --------
1 dc_sequences 153802 126036 51.6 153802
dc_rollback_segments 3054224 258 100.0 167
dc_segments 31636 2325 92.7 158
dc_objects 180114 3333 98.1 156
dc_histogram_defs 87711 6224 92.9 55
dc_constraints 36 19 47.2 36
outstanding_alerts 267 237 11.2 21
dc_awr_control 334 5 98.5 6
dc_users 243669 186 99.9 2
2 dc_sequences 160079 126035 56.7 160079
dc_segments 24599 2258 90.8 399
dc_objects 108531 2847 97.4 334
dc_rollback_segments 2849803 367 100.0 296
dc_constraints 92 47 48.9 92
dc_histogram_defs 43679 5286 87.9 49
outstanding_alerts 298 254 14.8 47
dc_awr_control 334 7 97.9 4
dc_global_oids 39576 106 99.7 1
dc_files 98 14 85.7 0
Similar to the preceding output, the cluster level also indicates high requests and conflicts. It indicates high
cluster level contention due to concurrency issues.
 
Search WWH ::




Custom Search