In addition to the Oracle-documented
initialization parameters, there are varying numbers of undocumented
initialization parameters in every version of Oracle. These
undocumented initialization parameters are usually only used in
emergencies and should only be used under the direction of a senior
DBA or Oracle support. Source 6 shows a script for getting the
undocumented initialization parameters out of an Oracle 7.2 instance
and then out of a 7.3 or 8.0.2 instance.
SOURCE 6: Scripts for documenting undocumented
initialization parameters
(pre-7.3 and post-7.3).
The output for an Oracle 7.2 database looks
like Listing 13.
LISTING 13: Undocumented initialization
parameters for Oracle 7.2 not included in Oracle 7.3 and 8.0.2.
The undocumented parameters for 7.3 are listed
in Listing 13. Note that the descriptions for the parameters are
available starting with Oracle 7.3. The undocumented parameters new
for Oracle8 are shown in Listing 14.
LISTING 14: Undocumented initialization
parameters for Oracle 7.3 not included
in 8.0.5.
Table 14 shows all of the undocumented
initialization parameters for 8i.
Parameter |
Value |
Description |
_NUMA_instance_mapping |
Not Specified |
Set of nodes that this instance should run
on |
_NUMA_pool_size |
Not Specified |
aggregate size in bytes of NUMA pool |
_advanced_dss_features |
FALSE |
enable advanced DSS features |
_affinity_on |
FALSE |
enable/disable affinity at run time |
_all_shared_dblinks |
NULL |
treat all dblinks as shared |
_allocate_creation_order |
FALSE |
should files be examined in creation order
during allocation |
_allow_read_only_corruption |
FALSE |
allow read-only open even if database is
corrupt |
_allow_resetlogs_corruption |
FALSE |
allow resetlogs even if it will cause
corruption |
_always_star_transformation |
FALSE |
always favor use of star transformation |
_arch_io_slaves |
0 |
ARCH I/O slaves |
_average_dirties_half_life |
2097152 |
Half life in bytes of first dirties
running average |
_b_tree_bitmap_plans |
FALSE |
enable the use of bitmap plans for tables
w. only B-tree indexes |
_backup_disk_io_slaves |
0 |
BACKUP Disk I/O slaves |
_backup_io_pool_size |
262144 |
memory to reserve from the large pool |
_bump_highwater_mark_count |
0 |
How many blocks should we allocate per
free list on advancing HW |
_cleanup_rollback_entries |
20 |
no. of undo entries to apply per
transaction cleanup |
_close_cached_open_cursors |
FALSE |
close cursors cached by PL/SQL at each
commit |
_column_elimination_off |
FALSE |
turn off predicate-only column elimination |
_compatible_no_recovery |
0.0.0 |
Database will be compatible unless crash
or media recovery is needed |
_complex_view_merging |
FALSE |
enable complex view merging |
_controlfile_enqueue_timeout |
900 |
control file enqueue timeout in seconds |
_corrupt_blocks_on_stuck_recovery |
0 |
number of times to corrupt a block when
media recovery stuck |
_corrupted_rollback_segments |
NULL |
corrupted undo segment list |
_cpu_to_io |
NULL |
multiplier for converting CPU cost to I/O
cost |
_cr_server |
NULL |
if TRUE, enable CR server TRUE (DFS) |
_cursor_db_buffers_pinned |
78 |
additional number of buffers a cursor can
pin at once |
_db_aging_cool_count |
1 |
Touch count set when buffer cooled |
_db_aging_freeze_cr |
2 |
Make CR buffers always be FALSE too cold
to keep in cache |
_db_aging_hot_criteria |
2 |
Touch count which sends a buffer to head
of replacement list |
_db_aging_stay_count |
99 |
Touch count set when buffer moved to head
of replacement list |
_db_aging_touch_time |
3 |
Touch count which sends a buffer to head
of replacement list |
_db_block_cache_clone |
FALSE |
Always clone data blocks on get (for
debugging) |
_db_block_cache_map |
0 |
Map / unmap and track reference counts on
blocks (for debugging) |
_db_block_cache_protect |
FALSE |
protect database blocks (true only when
debugging) |
_db_block_hash_buckets |
2000 |
Number of database block hash buckets |
_db_block_hi_priority_batch_size |
0 |
Fraction of writes for high priority
reasons |
_db_block_max_cr_dba |
10 |
Maximum Allowed Number of CR buffers per
dba |
_db_block_max_scan_cnt |
0 |
Maximum number of buffers to inspect when
looking for free |
_db_block_med_priority_batch_size |
0 |
Fraction of writes for medium priority
reasons |
_db_handles |
700 |
System-wide simultaneous buffer operations |
_db_handles_cached |
5 |
Buffer handles cached each process |
_db_large_dirty_queue |
0 |
Number of buffers which force dirty queue
to be written |
_db_no_mount_lock |
FALSE |
do not get a mount lock |
_db_percent_hot_default |
50 |
Percent of default buffer pool considered
hot |
_db_percent_hot_keep |
0 |
Percent of keep buffer pool considered hot |
_db_percent_hot_recycle |
0 |
Percent of recycle buffer pool considered
hot |
_db_writer_chunk_writes |
0 |
Number of writes DBWR should wait for |
_db_writer_histogram_statistics |
FALSE |
maintain dbwr histogram statistics in
x$kcbbhs |
_db_writer_max_writes |
0 |
Max number of outstanding DB Writer Ios |
_db_writer_scan_depth |
0 |
Number of LRU buffers for dbwr to scan
when looking for dirty blocks |
_db_writer_scan_depth_decrement |
0 |
Subtract from dbwr scan depth when dbwr is
working too hard |
_db_writer_scan_depth_increment |
0 |
Add to dbwr scan depth when dbwr is behind |
_dbwr_async_io |
TRUE |
Enable dbwriter asynchronous writes |
_dbwr_tracing |
0 |
Enable dbwriter tracing |
_defer_multiple_waiters |
TRUE |
if TRUE, defer down converts when there
were waiters (DFS) |
_disable_incremental_checkpoints |
FALSE |
Disable incremental checkpoints for thread
recovery |
_disable_latch_free_SCN_writes_via_32cas |
FALSE |
disable latch-free SCN writes using 32-bit
compare & swap |
_disable_latch_free_SCN_writes_via_64cas |
TRUE |
disable latch-free SCN writes using 64-bit
compare & swap |
_disable_logging |
FALSE |
Disable logging |
_disable_ntlog_events |
FALSE |
Disable logging to NT event log |
_discrete_transactions_enabled |
FALSE |
enable OLTP mode |
_distributed_lock_timeout |
60 |
number of seconds a distributed
transaction waits for a lock |
_distributed_recovery_connection_hold_time |
200 |
number of seconds RECO holds outbound
connections open |
_domain_index_batch_size |
2000 |
maximum number of rows from one call to
domain index fetch routine |
_domain_index_dml_batch_size |
200 |
maximum number of rows for one call to
domain index dml routines |
_dss_cache_flush |
TRUE |
enable full cache flush for parallel
execution |
_dynamic_stats_threshold |
30 |
Delay threshold (in seconds) between
sending statistics messages |
_enable_NUMA_optimization |
TRUE |
Enable NUMA specific optimizations |
_enable_block_level_transaction_recovery |
true |
enable block level recovery |
_enable_cscn_caching |
TRUE |
enable commit SCN caching for all
transactions |
_enable_default_affinity |
0 |
to enable default implementation of
affinity osds |
_enable_list_io |
TRUE |
Enable List I/O |
_enable_multitable_sampling |
FALSE |
enable multitable sampling |
_enqueue_debug_multi_instance |
FALSE |
Debug enqueue multi-instance |
_enqueue_hash |
265 |
enqueue hash table length |
_enqueue_hash_chain_latches |
2 |
enqueue hash chain latches |
_enqueue_locks |
1479 |
Locks for managed enqueues |
_fairness_threshold |
4 |
number of times to CR serve before
downgrading lock (DFS) |
_fast_full_scan_enabled |
FALSE |
enable fast full scan |
_foreground_locks |
TRUE |
if TRUE, foregrounds acquire PCM locks (DFS) |
_full_pwise_join_enabled |
TRUE |
enable full partition-wise join when TRUE |
_gc_class_locks |
0 |
set locks for the minor classes (DFS) |
_gc_latches |
4 |
number of latches per lock process (DFS) |
_gc_lck_procs |
1 |
number of background parallel server lock
processes to start |
_groupby_nopushdown_cut_ratio |
3 |
groupby nopushdown cut ratio |
_high_server_threshold |
0 |
high server thresholds |
_idl_conventional_index_maintenance |
TRUE |
enable conventional index maintenance for
insert direct load |