Tip: 看不到本站引用 Flickr 的图片? 下载 Firefox Access Flickr 插件 | AD: 订阅 DBA notes --

2018-08-02 Thu

21:13 Extended Histograms – 2 (1 Bytes) » Oracle Scratchpad
08:56 Oracle 18.3 和 12.2 版本初始化参数缺省值的改变列表 (10056 Bytes) » Oracle Life

作者:eygle 发布在 eygle.com

从 Oracle Database 12.2 到 Oracle 18.3 ,有很多初始化参数的缺省值发生了变化,我整理了一个列表供大家参考。

这其中有些参数和硬件配置有关,但是绝大多数就是因为版本变化做出的改变。尤其是参数值的 True 和 False 的改变。

比如 _use_single_log_writer 的缺省值变成了 TRUE,意味着缺省将禁用自适应 LGWR 的特性。

------------------------------ ------------------------------ ------------------------------ ------------------------------------------------------------
_spin_count		       1			      2000			     Amount to spin waiting for a latch
_super_shared_conversion_threshold 300			      1280			     conversion threshold limit for supershared mode

_num_longop_child_latches      1			      8 			     number of child latches for long op array
_kill_session_dump	       FALSE			      TRUE			     Process dump on kill session immediate
_session_allocation_latches    1			      8 			     one latch per group of sessions
_disable_highres_ticks	       TRUE			      FALSE			     disable high-res tick counter
_enqueue_sync_retry_attempts   15			      3 			     max number of times the bg process to retry synchronous enqu
											     eue open if it failed because master could not allocate memo

_enable_shared_pool_durations  TRUE			      FALSE			     temporary to disable/enable kgh policy
_kse_snap_ring_size	       0			      100			     ring buffer to debug internal error 17090
_ksi_clientlocks_enabled       TRUE			      FALSE			     if TRUE, DLM-clients can provide the lock memory
_instant_file_create	       FALSE			      TRUE			     enable instant file creation on sparse media
_ksipc_service_mask	       1			      0 			     KSIPC Service Mask
_proc_grp_enabled	       3			      FALSE			     proc-group enabled
_ksrma_enabled		       off			      OFF			     turn ksrma off, make it automatic, or turn on CMI or KSMSQ p
											     rovider. Default is off

_hang_resolution_promote_proce TRUE			      FALSE			     Hang Management hang resolution promote process termination

_lm_singleton_pdb_opt	       TRUE			      FALSE			     RAC PDB singleton optimization
_lm_rm_object_bypass	       TRUE			      FALSE			     enable read-mostly object bypass for HARIM
_reconfiguration_opt_level     13			      0 			     reconfiguration optimization level
_dlm_stats_collect_mode        4			      0 			     DLM statistics collection mode
_lm_comm_rcv_msg_history_slots 50			      0 			     GES communication receive message history slots
_reliable_block_sends	       TRUE			      FALSE			     if TRUE, no side channel on reliable interconnect
_memory_broker_shrink_heaps    15			      0 			     memory broker allow policy to shrink shared pool
_dbwr_nowrite_assert_interval  7200			      0 			     dbwriter assert interval after no write seconds
_db_block_vlm_check	       FALSE			      TRUE			     check for mapping leaks (debugging only)
_db_cache_miss_check_les       FALSE			      TRUE			     check LEs after cache miss
_db_cache_mman_latch_check     FALSE			      TRUE			     check for wait latch get under MMAN ops in kcb
_db_block_chunkify_ncmbr       FALSE			      TRUE			     chunkify noncontig multi block reads
_db_cache_wait_debug	       0			      1 			     trace new kslwaits
_db_block_iterations_for_rm    2000			      20			     number of blocks to reduce every iteration for RM
_db_dump_from_disk_and_efc     0			      2 			     dump contents from disk and efc
_assert_encrypted_tablespace_blocks TRUE			      FALSE	             if TRUE, assert encrypted tablespace blocks must be encrypted

_high_intrinsic_scn_growth_alert 1440			      60			     high intrinsic SCN growth rate alert time in minutes

compatible		       18.0.0		     Database will be completely compatible with this software ve

_num_rlslaves		       2			      16			     number of rl slaves to clear orls
_use_single_log_writer	       TRUE			      ADAPTIVE			     Use a single process for redo log writing
_max_pending_scn_bcasts        8			      9 			     maximum number of pending SCN broadcasts
_gc_trace_freelist_empty       FALSE			      TRUE			     if TRUE, dump a trace when we run out of lock elements
_auto_export_preplugin_backup  TRUE			      FALSE			     auto export pre-plugin backup
_time_based_rcv_ckpt_target    180			      0 			     time-based incremental recovery checkpoint target in sec
_time_based_rcv_hdr_update_int 180			      0 			     time-based incremental recovery file header update interval
erval											     in sec

_nologging_sendbuf_ratio       90			      99			     Nologging standby: outstanding send buffer ratio
_nologging_sdcl_append_wait    200			      100			     Nologging standby append sdcl wait time
_shadow_lost_write_found       118			      22			     Specify shadow lost write error handling

_imtxn_table_enable TRUE FALSE is In-Memory Txn Table cache enabled
_compression_compatibility 18.0.0 Compression compatibility _dbg_scan 0 2048 generic scan debug _inmemory_subcusize 512 0 In-memory subCU size _inmemory_fsdw_schedlrtm 1 10 in-memory faststart defer write scheduler cycle (sec) _inmemory_repopulate_disable 2 FALSE disable In-memory repopulate _inmemory_repopulate_flags 2 0 In-memory repopulate decision flags _inmemory_dynamic_scans AUTO DISABLE Enable IM Dynamic Scans _imado_optimize_period 0 20 IM-store optimize period for AIM (in minutes) _kql_clientlocks_enabled 15 1 clients allocating DLM memory _pdb_service_on_root_listener FALSE TRUE pdb services on CDB ROOT listeners _cursor_db_buffers_pinned 76 2 additional number of buffers a cursor can pin at once
_qesma_mvob_lru_sz 25 0 size of MVOB LRU list for QESMA session cache _qesma_bo_lru_sz 25 0 size of base object LRU list for QESMA session cache optimizer_features_enable 18.1.0 optimizer plan compatibility parameter audit_trail DB NONE enable system auditing _optimizer_undo_cost_change 18.1.0 optimizer undo cost change parallel_threads_per_cpu 1 2 number of parallel execution threads per CPU _parallel_adaptive_max_users 4 2 maximum number of users running with default DOP _qa_lrg_type 0 1 Oracle internal parameter to specify QA lrg type _bloom_folding_min 0 128 bloom filter folding size lower bound (in KB) _smm_isort_cap 102400 0 maximum work area for insertion sort(v1) _sqlmon_max_plan 20 160 Maximum number of plans entry that can be monitored. Default s to 20 per CPU parallel_servers_target 20 64 instance target in terms of number of parallel servers _px_chunklist_count_ratio 32 8 ratio of the number of chunk lists to the default DOP per in stance _emon_outbound_connect_timeout 7200000 30000 timeout for completing connection set up to clients _emon_send_timeout 7200000 10000 send timeout after which the client is unregistered _aq_pt_processes 10 15 Partition background processes _asm_disable_proact_client_cle FALSE TRUE disable proactive client cleanup anup _asm_disable_patch_compat FALSE TRUE disable DG patch compatibility checks _asm_allow_prepare_split TRUE FALSE Allow database prepare and split in non-appliance mode target_pdbs 1 2 Parameter is a hint to adjust certain attributes of the CDB max_pdbs 254 4098 max number of pdbs allowed in CDB or Application ROOT _cdb_disable_pdb_limit FALSE TRUE CDB Compatible _cdb_special_old_xplan TRUE FALSE display old-style plan for CDB special fixed table _upgrade_optim TRUE FALSE Upgrade optimization enabled _upgrade_capture_noops TRUE FALSE do not skip capturing noops during upgrade _enable_containers_subquery TRUE FALSE enable transformation to containers() sub query


相关文章|Related Articles

2018-08-01 Wed

19:20 Oracle 18.3 版本参数比较12.2 看新特性优化诊断增强 (34129 Bytes) » Oracle Life

作者:eygle 发布在 eygle.com

在 Oracle Database 18.3 版本发布之后,一系列新的参数被加入到数据库中,这些新的参数就代表了数据库中新增的特性。

以下列表是对比 Oracle Database 12.2 版本,18.3 中新增的参数。通过 x$ksppi 可以看到 12.2 版本有 4846 个参数,而 18.3 有 5159 个参数.
从参数数量上来说,增加了 313 个,这是一个不小的数字,无论是跟踪诊断,性能优化,还是 异常恢复,这些参数都是核心所在。

-------------------------------------------------- ------------------------- --------------------------------------------------------------------------------
_EnableDDLTtriggerTracing			   FALSE		     enable ddl trigger tracing
_EnableShadowTypes				   FALSE		     enable shadow types
_REMOVE_INACTIVE_STANDBY_TDE_MASTER_KEY 	   FALSE		     Remove Inactive Standby TDE Master Key
_REMOVE_STDBY_OLD_KEY_AFTER_CHECKPOINT_SCN	   TRUE 		     Remove standby old key after checkpoint SCN
_STFTranslateDynamicSQL 			   FALSE		     if TRUE translation profile will translate dynamic SQL statements
__memoptimize_xmem_pool_size_metadata		   0			     Size of extended cache metadata of MEMOPTIMIZE buffer pool for standard block si
									     ze buffers

__sga_current_size				   0			     Current size for PDB SGA
_adg_auto_close_pdb				   TRUE 		     ADG recovery auto close PDB upon PDB drop/unplug/rename marker
_adg_objectlock_attempts			   2			     Maximum attemps for objectlock get on ADG
_adg_objectlock_maxnum				   1000 		     The maximum limit of the objectlock number on ADG
_adg_objectlock_timeout 			   0			     timeout for objectlock get on ADG in centiseconds
_adg_redirect_flags				   0			     Flags for controlling ADG's Statement Redirection behavior
_allocate_flashback_buffer			   FALSE		     Allocate flashback buffer at mount time even if flashback is off
_allow_cross_endian_dictionary			   FALSE		     Allow cross-endian data dictionary
_app_container_debug				   0			     Enable Debug tracing in Application container
_apppdb_multi_slave_sync			   TRUE 		     Multiple slaves used during Application sync
_aq_dqblocks_in_cache				   0			     deq blocks in cache
_aq_free_list_pools				   10			     state object and transaction memory pools
_aq_lb_garbage_col_interval			   600			     AQ LB garbage collect interval
_aq_lb_subht_bkt_ltch				   32			     AQ LB subscriber ht bucket latches
_aq_lb_subht_elm_ltch				   128			     AQ LB subscriber ht element latches
_aq_max_pdb_close_msg				   1			     Max Number of AQ Recovery Messages when pdb is closed
_aq_scrambled_deqlog				   1			     scrambled deqlog
_asm_allow_dgname_special_chars 		   FALSE		     Enable the use of special characters in diskgroup names
_asm_async_scrub_reap_wait			   100000		     Async scrubbing reaping IO wait in microseconds
_asm_cancel_delta				   75000		     Delta timeout for blocking big writes in milliseconds
_asm_enable_batch_scrub 			   FALSE		     Allow scrubbing verification to run in batch mode
_asm_enable_kfios				   FALSE		     Enable KFIOS module
_asm_oda_type								     ODA Type
_asm_proxy_online_restart			   0			     Allow patching while ADVM volumes are online
_asm_pstonpartners				   TRUE 		     Select ideal PST disks following partnership
_asm_reloc_cic					   FALSE		     Allow relocation during rolling migration
_asm_relocation_scheme				   alloc_p2 alloc_s3 reb_p2  ASM allocation / relocation scheme
						   reb_s1 bal_p2 bal_s3 prep
						   _p2 prep_s3

_asm_write_badfdata_in_contentcheck		   TRUE 		     Write BADFDA7A in content check
_asm_zero_power_limit							     allow setting zero power limit
_async_scn_sync 				   OFF			     Asynchronous SCN Sync Setting
_auto_dismount_on_pdb_close			   FALSE		     Enable implicit PDB dismount on PDB close
_auto_rcv_pdb_open				   1			     enable automatic recovery during pdb open
_awr_incremental_flush_enabled			   TRUE 		     Enable/Disable AWR automatic incremental flushing
_backup_block0					   default		     backup block0
_bequeath_via_broker				   FALSE		     Bequeath connections via broker
_bloom_extent_size				   0			     bloom filter extent size in bytes
_bloom_filter_setops_enabled			   TRUE 		     Allow bloom filter to be pushed through set operations
_bloom_max_wait_time				   50			     bloom filter wait time upper bound (in ms)
_bloom_pruning_setops_enabled			   TRUE 		     Allow bloom pruning to be pushed through set operations
_bloom_use_shared_pool				   FALSE		     use shared pool for bloom filter
_bloom_wait_on_rac				   FALSE		     enables bloom filter (with hash hash distribution) wait on RAC
_capability_simulate							     Simulate the capabilities for testing
_cdb_fleet_sync_timeout 			   10			     Time in minutes to wait for sync of stub entry in CDB Fleet
_cdb_port					   0			     Port number for CDB
_cdb_view_no_skip_restricted			   FALSE		     do not skip RESTRICTED mode PDBs from results of CONTAINERS()
_cell_offload_hybrid_processing 		   TRUE 		     enable hybrid SQL processing offload to cells
_cell_offload_vector_groupby_force		   FALSE		     force offload of vector group by
_cgs_publish_netinfo_collect_event_chm		   TRUE 		     enable cgs publish collect_netinfo event to event stream for CHM
_cgs_publish_netinfo_collect_event_haip 	   FALSE		     enable cgs publish collect_netinfo event to event stream for HAIP
_clear_preserved_buffers			   TRUE 		     Clear preserved buffers before DB reopen after switchover
_client_result_cache_ramthreshold					     client_result_cache_ramthreshold
_cloud_service_sim				   0			     simulate cloud services in Database
_cloud_service_type							     cloud service type
_con_map_sql_enforcement			   TRUE 		     disable container map SQL enforcement
_connection_broker_handout_accept		   FALSE		     connection broker accepts prior to handout
_cross_con_collection				   FALSE		     enable cross container collection
_datapump_gather_stats_on_load			   FALSE		     Gather table statistics during Data Pump load rather thanimporting statistics fr
									     om the dump file. This should be set to TRUE in the lockdown profile in a DWCS e

_datapump_inherit_svcname			   FALSE		     Inherit and propagate service name throughout job
_db_flash_cache_slow_io_adjustment_interval	   3600 		     Decrement interval
_db_imoltp_hashidx_force_nonctg 		   0			     kcbw imoltp hash index force non contig granules
_db_shadow_lost_write_protect			   NOT_SET		     alter shadow lost write detection for PDB
_disable_actualization_for_grant		   FALSE		     disable actualization of editioned objects on grant
_disable_asm_audit_feat 			   0			     disable ASM audit features
_disable_dblink_optim				   TRUE 		     disable intra CDB dblink optimizations
_disable_system_tablespaces_enc_enforcement	   FALSE		     if TRUE, disable system tablespaces encryption enforcement
_dmm_pga_load_threshold 			   3			     Model size less than threshold are loaded into PGA
_drain_on_ping_database 			   TRUE 		     Should session drain be performed when user issues OPING
_dskm_single_instance				   FALSE		     DSKM and Diskmon operating in Single Instance mode
_dss_cache_flush_threshold			   1			     threshold when thread ckpt considered over obj ckpt
_duplicated_table_complete_refresh_percent	   50			     percent threshold for duplicated table to do complete refresh
_dupt_noupdate					   0			     disable duplicated table updates
_ena_storage_lmt				   DEFAULT		     Enable storage clause for LMT
_enable_auto_upgrade				   FALSE		     Enable automatic PDB upgrade
_enable_dbwr_auto_tracing			   0			     enable dbwriter auto-tracing
_enable_imc_mira				   FALSE		     enable IMC with  multi instance redo apply
_enable_ios_spm 				   FALSE		     Allow Split Mirror operations via IOServer
_enable_multiple_fgprepares			   FALSE		     Allow concurrent PREPAREs on the same database
_enable_parallel_dml				   FALSE		     enables or disables parallel dml
_enable_pmo_outside_begin_end			   TRUE 		     Enable PMO outside begin-end block
_enable_proxy_adg_redirect			   FALSE		     Enable Statement Redirection Support for ADG using Proxy PDB
_enable_single_dgprepare			   FALSE		     Disable concurrent PREPAREs in same disk group
_error_row_predicate_evaluation 		   AUTO 		     skip predicate evaluation for error rows
_exadata_feature_on				   FALSE		     Exadata Feature On
_fg_fast_sync_sleep_usecs			   50			     DAX log file sync sleep time
_fg_fast_sync_spin_usecs			   300			     DAX log file sync spins time
_force_active_temp_detection			   FALSE		     Force active temp detection during PDB close
_freeze_kgh_timestamp				   FALSE		     prevent heap manager timestamp from advancing
_gc_lease_time					   500			     lease time for rdma reads
_gc_max_reg_sz					   68719476736		     maximum length for memory registration
_gc_partial_cleanout				   FALSE		     if TRUE, partial cleanout is enabled
_gc_undo_rdma_read				   FALSE		     if TRUE, rdma read of undo blocks is enabled
_gcs_disable_imc_preallocation			   FALSE		     disable preallocation for imc memory requirement in RAC
_gcs_enable_private_iterator			   TRUE 		     Enable private iterator for GCS locks
_gcs_recoverable_asserts			   1			     if non-zero, enable recoverable assert resolution
_gdr_control_flags				   0			     gdr control flags
_gds_allow_nullkey				   0			     allow the use of nullable shard key columns
_gds_max_chunk_num				   0			     max chunk_num used in sharding environment
_ges_mseq_demo					   0			     demo mseq wrap scenarios (dflt is 0)
_globaldict_analyzer_pct			   100			     Maximum percentage of unique values in analyzer phase for GD
_globaldict_chain_limit 			   32			     Chain limit for global dictionary
_globaldict_check				   0			     Dictionary checking
_globaldict_chunk_minalloc			   FALSE		     Force minimum chunk allocation size
_globaldict_dbg 				   0			     Global dictionary debug modes
_globaldict_enable				   2			     Enable segment dictionary
_globaldict_max_gdsize				   1073741824		     Maximum number of entries in a Global dictionary
_globaldict_reprobe_limit			   1			     Reprobe limit for global dictionary
_globaldict_use_ndv				   TRUE 		     Use NDV for sizing global dictionary, if available
_gwm_db_unique_name							     gsm db_unique_name name
_hang_appl_issue_session_threshold		   0			     Hang Management application issue session threshold
_hang_asm_hang_resolution_enabled		   FALSE		     Hang Management ASM hang resolution enabled
_hang_deadlock_resolution_enabled		   TRUE 		     Hang Management deadlock resolution enabled
_hang_hang_blocked_session_delta_percent_threshold 20			     Hang Manager hang's blocked session delta percent threshold
_hang_ignore_hngmtrc_interval			   150			     Hang Management ignore hang metric dependent hang interval
_hang_intersecting_chains_scanning_enabled	   TRUE 		     Hang Management intersecting chains scanning is allowed
_hang_log_important_hangs_to_alert		   TRUE 		     Hang Management log important hangs to alert log
_hang_promote_process_termination_interval	   70			     Hang Management promote process termination interval in seconds
_hang_root_ha_phase_trigger_time		   300			     Hang Management root HA phase trigger time
_hcs_expose_with_expr				   FALSE		     expose internal with_expression
_hcs_gen_aggr_opt_estimate			   FALSE		     apply OPT_ESTIMATE hint to aggregation queries
_hcs_ignore_latest_compat_check 		   FALSE		     skip compatibility check for latest compatible version
_hcs_in_mem_cdt_hint				   FALSE		     add hint opt_param('_in_memory_cdt', 'off')
_hcs_mdx_cache_name_col 			   FALSE		     add column for mdx cache name
_hcs_mdx_cache_name_no_sid			   FALSE		     don't use session id in mdx cache hint names
_hcs_mdx_sleep_after_pin			   FALSE		     sleep after pinning MDX caches
_hcs_no_calc_dtm_to_out_opt			   FALSE		     apply move calc determined hier to output hier optimization
_hcs_no_cell_qry_atr_prune_opt			   FALSE		     apply attribute prune optimization to cell query
_hcs_no_cell_qry_lvl_prune_opt			   FALSE		     apply level prune optimization to cell query
_hcs_no_cell_qry_meas_prune_opt 		   FALSE		     apply measure prune optimization to cell query
_hcs_no_cell_qry_mv_cache_opt			   FALSE		     apply mv cache optimization to cell query
_hcs_no_cell_qry_no_calc_nav_opt		   FALSE		     apply no calc navigation optimization to cell query
_hcs_no_cell_qry_no_out_data_opt		   FALSE		     apply no output data optimization to cell query
_hcs_no_cell_qry_tmpls				   FALSE		     no cell query templates for optimization
_hcs_no_col_prune_optz				   FALSE		     skip column pruning optimization on generated SQL
_hcs_no_fltr_fact_opt				   FALSE		     apply filtered fact optimization
_hcs_no_fltr_hier_star_opt			   FALSE		     apply filtered hierarchy star optimization
_hcs_no_inline_tmpl_opt 			   FALSE		     apply inline single ref top blocks optimization to cell query
_hcs_no_jback_opt_for_hord_in_oby		   FALSE		     optimize analytic view joinback for HIER_ORDER
_hcs_no_mdx_cache_hint				   FALSE		     generate hcs query mdx cache hints
_hcs_no_mv_rewrite_check			   FALSE		     skip MV rewrite check when generating for ANALYTIC VIEW cache
_hcs_no_opt_cell_qry				   FALSE		     optimize cell query
_hcs_no_rm_like_withs_optz			   FALSE		     skip like WITH element removal optimization on generated SQL
_hcs_no_rm_unused_withs_optz			   FALSE		     skip unused WITH element removal optimization on generated SQL
_hcs_no_rmv_unref_top_opt			   FALSE		     apply remove unref top blocks optimization to cell query
_hcs_no_tgt_depths_opt				   FALSE		     apply target depths optimization
_hcs_opt_av_pred_push				   TRUE 		     enable optimizer AV predicate pushing via reparse
_hcs_parse_dynamic_all_cache			   FALSE		     allow parsing of dynamic across all levels cache
_hcs_stats_max_card				   2000000		     maximum value for stats cardinality
_hcs_use_dynamic_all_cache			   FALSE		     use dynamic across all levels cache
_hcs_use_multi_parent_gen			   FALSE		     generate hcs query using multi-parent aggregation
_hugetlbfs_mount_point_for_sga						     HugeTLBFS mount point to be used
_imado_diagtasks_log_period			   5			     IM-ADO diagnostic tasks logging period (in seconds)
_imado_diagtasks_purge_period			   30			     IM diagnostic tasks purge older than X days
_imado_evict_sf 				   2			     AIM evict safety factor
_imado_sysaux_usage_limit			   90			     SYSAUX usage percent limit for storing AIM diagnostics
_imtxnrma_table_enable				   FALSE		     is In-Memory RMA Txn Table cache enabled
_index_load_analysis_frequency			   4			     index load compression analysis frequency
_inmemory_adg_batched_flush			   TRUE 		     If true, batched flush is performed
_inmemory_adg_journal_quota			   FALSE		     If true, throttled mining is performed under space pressure
_inmemory_adg_parallel_flush			   TRUE 		     If true, parallel flush is performed
_inmemory_adg_periodic_sort			   FALSE		     If true, periodic sort is performed
_inmemory_delta_population			   0			     Control Delta IMCU population
_inmemory_dynamic_scans_analyze_batch_size	   5			     Inmemory dynamic scan analyze batch size
_inmemory_external_table			   TRUE 		     Enable inmemory extern table
_inmemory_grpcolinv_buffer_size 		   131072		     In-memory grpcolinv buffer size
_inmemory_grpcolinv_granularity 		   1			     In-memory grpcolinv blks per colmap
_inmemory_hwm_expand_percent			   20			     max blks percent can exceed in hwmimcu
_inmemory_hwm_expansion 			   0			     If 0, the highwatermark CU is dropped when expanded
_inmemory_max_delta				   5			     Inmemory max number of deltas per CU
_inmemory_min_delta_blocks			   50			     Inmemory minimum blocks per delta
_inmemory_min_delta_rows			   255			     Inmemory minimum rows per delta
_inmemory_patch_background_blocks		   200			     In-memory SMU patch threshold blocks to start background tasks
_inmemory_patch_commit_path			   FALSE		     If true, enable the SMU patching from DML path
_inmemory_patch_threshold_blocks		   25			     In-memory SMU patch threshold blocks
_inmemory_prefix_encode_dsbs			   FALSE		     Prefix-encode the DSB-dictionary
_inmemory_scan_invalid_percent			   30			     Inmemory scan invalid percentage
_inmemory_smu_patch_options			   7			     Inmemory SMU patching options
_inmemory_vector_encode_override		   FALSE		     Populate and use DSBs for Vector Encode columns
_instance_recovery_bloom_filter_fprate		   0			     Allowable false positive percentage [0-100]
_instance_recovery_bloom_filter_size		   83886080		     Bloom filter size (in num of bits) used during claim phase
_intra_cdb_dblink				   FALSE		     enable intra CDB implicit dblink
_kcfis_pmem_enabled				   FALSE		     RDMA Persistent Memory support enabled
_kdfip_buf_nclatch				   64			     memopt w buffer child latches
_kdfip_bufl_nbkt				   64			     memopt w buffers list buckets
_kdfip_cmap_nbkt				   16			     memopt w chunk map buckets
_kdfip_debug					   0			     memopt w debug
_kdfip_drain_sleeps				   60			     memopt w drain coord sleep count
_kdfip_elem_nclatch				   64			     memopt w chunkmap elem child latches
_kdfip_flush_nrows				   10			     memopt w flush num rows
_kdfip_flush_rowsz				   1024 		     memopt w flush row size
_kdfip_flush_rowtm				   60			     memopt w flush time
_kdfip_iga_bufsz				   65536		     memopt w write buffer size
_kdfip_iga_maxsz				   52428800		     memopt w max global area size
_kdfip_par_flush				   TRUE 		     memopt w parallel flush
_kdfip_trace					   FALSE		     memopt w tracing on-off
_kdkv_fg_drop_memopt				   TRUE 		     hashindex foreground drop
_kdkv_fg_no_memopt				   FALSE		     hashindex foreground cleanup
_kdkv_fg_populate				   FALSE		     hashindex foreground populate
_kdkv_fg_repopulate				   FALSE		     hashindex foreground repopulate
_kdkv_force_fastpath				   FALSE		     kdkv fast path on-off
_kdkv_index_lossy				   TRUE 		     hashindex lossiness on-off
_kdkv_index_relocate				   FALSE		     hashindex relocation on-off
_kdkv_indexinvalid				   FALSE		     objd and rdba based index invalidation
_kdkv_trace					   FALSE		     kdkv tracing on-off
_kdlxp_no_dedup_on_insert			   FALSE		     disable deduplication for new inserts of deduplicated lobs
_key_vector_double_enabled			   TRUE 		     enables or disables key vector support for binary_double
_key_vector_shared_dgk_ht			   TRUE 		     allows shared DGK hash table
_key_vector_timestamp_enabled			   TRUE 		     enables or disables key vector support for timestamp
_kra_cfile_compaction				   TRUE 		     controlfile record compation
_kse_alt_stack_sig_syms 			   10			     The number of top symbols used for the alternate stack signature
_ksipc_trace_bucket				   PRIVATE		     memory tracing: use ksipc-private or rdbms-shared bucket
_ksipc_trace_bucket_size			   IPC0:1048576-REST:8192    KSIPC trace bucket size in bytes (format: "IPC0:-REST:")
_ksipc_window_size							     Configure IPC Windowing Value for clients on a per transport basis
_ksipcsnsrv								     Configure Shared Nothing Server Name
_ksm_sp_rcr_hits				   10			     hits to make object recurrent
_ksmd_trace					   0			     ksmd tracing
_ksmsq_hintmaxinst				   1024 		     KSMSQ Hint Max Instances
_ksmsq_hintmaxproc				   300			     KSMSQ Hint Max Processes
_ksws_java_patching				   0			     java patching mode
_kswsas_drain_kill_batch_size			   5			     Batch size for killing non-drained sessions
_kswsas_num_jp_slaves							     Number of slaves for java patching
_ksxp_validate_cnh_life_cycle			   0			     enable validation of ksxp connection life cycle
_ktst_tscleanup_timeout 			   1800 		     Time in seconds to wait for active temp clients
_ldap_config_force_sync_up			   FALSE		     LDAP configure force sync up
_ldap_config_ssl_for_sasl_md5			   TRUE 		     LDAP configure SSL for SASL-DIGEST-MD5
_ldap_reset_user_account_flc			   TRUE 		     LDAP reset user account lockout counter
_lm_chk_inv_domenq_ops				   TRUE 		     enable checks for invalid enqueue operations on domains
_lm_comm_slow_op_loop_threshold 		   15			     GES communication slow operation loop threshold in ms
_lm_enq_iso_enabled				   TRUE 		     if TRUE enables enqueue isolation
_lm_idle_connection_action			   kill 		     GES idle connection action
_lm_idle_connection_load_check			   TRUE 		     GES idle connection load check
_lm_nonisolated_restype 			   TOTTUL		     string of resource types banned from enqueue isolation
_lm_pdb_wait_all_gone				   FALSE		     pdb domain attach wait until all locks are gone
_lm_recovery_set				   1			     enable recovery set checking for accessing invalid domains
_lm_share_lock_pdbisolation			   TRUE 		     if TRUE enables share lock optimization with pdb isolation
_lm_use_us_timer				   FALSE		     Use microsecond timer for LM hist
_lm_wait_for_hub_rcv_timeout			   300000		     read-only insts wait for hub avaliable to recover in millis
_lmhb_procstate_dump_cputime_limit		   60			     hb diagnostic cputime limit for process state dump in secs
_lmhb_procstate_dump_runtime_limit		   60			     hb diagnostic runtime limit for process state dump in secs
_lob_use_locator_varying_width			   FALSE		     use varying width flag in lob locator
_max_lwt_cpu_ratio				   2			     ratio to determine the maximum CPUs for LWTs
_memoptimize_xmem_pool_size			   0			     Size of MEMOPTIMIZE buffer pool for standard block size buffers on extended memo

_memory_adi_module_mask 			   0			     bit mask of modules to enable ADI versioning
_mga_large_page_path							     large page path
_next_pdbid					   3			     Hint for the next PDBID to use when creating a new PDB entry
_no_catalog								     options whose schemas should not be created
_no_snapshot_root_clone 			   FALSE		     No snapshot clone during Application Root Clone creation
_nologging_apply_stall_time			   1000 		     milli-seconds recovery waits after DTC full before changing RCVID
_nologging_fetch_initial_interval		   2			     seconds recovery waits between issuing fetches for old ranges
_nologging_mode_override			   0			     Override the database logging mode; 1=Force, 2=NoForce, 3=LP, 4=DA
_nologging_progress_keep_alive			   10			     Nologging standby progress keep alive time
_nologging_slots				   20			     Nologging standby: initial buffer count
_nologging_standby_cold_buffer_timeout		   500			     Centi-seconds recovery will wait for a buffer that is not being sent by a direct
									      load client in Nologging Standby Data Availability mode

_nologging_standby_dtc_expire			   600			     The number of seconds a Data Transfer Cache buffer may remain unclaimed
_nologging_standby_fetch_disable		   FALSE		     controls whether invalid block ranges are fecthed during recovery
_nologging_standby_hot_buffer_timeout		   500			     Centi-seconds recovery will wait for a buffer being sent by a direct load client
									      in Nologging Standby Data Availabilty mode

_nologging_standby_refetch_disable		   FALSE		     controls fetching of pre-existing invalid block ranges during standby recovery
_optimizer_allow_all_access_paths		   TRUE 		     allow all access paths
_optimizer_answering_query_using_stats		   FALSE		     enable statistics-based query transformation
_optimizer_gather_stats_on_load_all		   FALSE		     enable/disable online statistics gathering for nonempty segments
_optimizer_gather_stats_on_load_hist		   FALSE		     enable/disable online histogram gathering for loads
_optimizer_key_vector_payload			   TRUE 		     enables or disables dimension payloads in vector transform
_optimizer_vector_fact_payload_ratio		   20			     cost based vector transform payload dimension to fact ratio
_parallel_lmd_reconfig				   1			     parallel lmd work in reconfiguration for enqueues.
_partition_by_con_name				   FALSE		     enable partitioning by con$name
_pdb_ignore_table_clauses			   TRUE 		     Ignore subset of clauses in CREATE TABLE and ALTER TABLE
_pdb_isolation_class				   NONE 		     PDB isolation level: NONE, HIGH or MAX
_pdb_ldp_cascade				   0			     pluggable database lockdown profile cascade param
_pga_limit_per_process_minimum			   3145728		     pga_aggregate_limit per-process minimum
_pmon_exitnfy_enabled				   FALSE		     PMON Exit notification enabled
_ptf_enable_objects				   FALSE		     enable object types in PTF
_ptf_max_rows					   1024 		     number of rows per row-buffer
_ptt_max_num					   16			     Maximum number of PTTs per session
_px_granule_alignment				   1024 		     default size of a rowid range granule (in KB)
_px_join_skew_null_handling			   TRUE 		     enables null skew handling improvement for parallel outer joins
_px_join_skew_sampling_percent			   0			     Sets sampling percentage for skew handling sampling queries
_px_join_skew_sampling_time_limit		   50			     Sets execution time limit for skew handling sampling queries
_px_join_skew_use_histogram			   TRUE 		     Enables retrieval of skewed values from histogram when possible
_px_nlj_bcast_rr_threshold			   10			     threshold to use broadcast left random right distribution for NLJ
_px_partition_skew_threshold			   80			     percentage of table size considered as threshold for skew determination
_px_pwise_wif_enabled				   TRUE 		     enable parallel partition wise window function
_px_reuse_server_groups 			   MULTI		     enable/disable reusing of server groups that are already acquired
_px_shared_hash_join				   FALSE		     enable/disable shared hash join
_px_tq_memcpy_threshold 			   100			     threshold for PQ TQ to use memcpy for packing columns
_qksfgi_dynamic_partitions						     enable partition bind SQLIDs for
_qksfgi_feature_level				   0			     controls the feature level for fine-grained cursor invalidation
_rac_propagate_last_rba 			   TRUE 		     Propagate last written log redo address in RAC for log mining
_read_only_slave_timeout			   30			     Time to wait on read only node when hub not available
_read_optimized_table_lookup			   TRUE 		     enables read-optimized table lookup
_reader_farm_isolation_enable			   FALSE		     Reader Farm instances isolation enable
_reader_farm_isolation_time_threshold		   200			     reader farm isolation time threshold
_record_module_name				   TRUE 		     Record module name in BEGIN action
_redo_log_key_reuse_count			   0			     Determines the number of redo logs sharing the same redo log key
_redo_write_sync_single_io			   TRUE 		     enable sync I/O for single redo write
_ref_cons_sql_enforcement			   TRUE 		     disable extended/data ref constraint SQL enforcement
_reg_cache_status				   FALSE		     Enables or disabled caching
_remessage_dbwrs				   0			     DBWR wait and remessage time - in cs
_request_boundaries				   1			     request boundaries mode
_restore_block0 				   on			     restore block0
_result_cache_per_pdb				   TRUE 		     Is service result cache per pdb
_rm_exadata_partition_fc			   FALSE		     Partition flash cache for Exadata
_rm_exadata_pdb_cpu_cnt_mult			   2			     Multiplication factor for PDB cpu count
_rmv_dynamic_priority				   TRUE 		     enable rmv priority modification
_root_clone_state_from_root			   TRUE 		     Application Root Clone's state comes from Application Root
_sequence_scale_extend				   FALSE		     Force sequence creation with scale extend
_sequence_scale_noextend			   FALSE		     Force sequence creation with scale noextend
_shd_atomic_move				   0			     Use atomic move
_show_login_pdb_sessions			   FALSE		     return logon pdb container id from v$session
_simulate_dax_storage				   DISABLE		     Simulate log on DAX storage
_skip_pdb_recovery_if_keystore_not_open 	   FALSE		     Skip PDB recovery when the PDB's keystore is not open
_skip_unconverted_change_vector 		   FALSE		     Skip apply of a CV that has no endian conversion function
_slow_kill_on_pdb_close_immediate		   FALSE		     Kill sessions periodically on pdb close immediate
_sqlexec_cache_aware_hash_aggr_enabled		   TRUE 		     enable/disable cache aware hash aggregation
_sqlexec_pwiseops_with_binds_enabled		   TRUE 		     enable partition wise execution in the presence of bind variables in inlist equa
									     lity operator in where clause

_sqlexec_pwiseops_with_sqlfuncs_enabled 	   TRUE 		     enables partition wise operations even in the presence of functions over table p
									     artition keys

_sqlexec_reorder_wif_enabled			   TRUE 		     enable re-ordering of window functions
_sqlexec_use_kgghash3				   TRUE 		     use CRC-based hash function
_standby_newkey_keystore_location					     Location of keystore on standby having new key
_uniq_cons_sql_enforcement			   TRUE 		     disable extended data unique constraint SQL enforcement
_use_fallocate_for_mga				   FALSE		     MGA fallocate enabled
_use_hugetlbfs_for_SGA				   FALSE		     Enable HugeTLBFS usage for SGA
_use_large_pages_for_mga			   FALSE		     MGA largepage enabled
_verify_encrypted_tablespace_keys		   TRUE 		     if TRUE, verify encryption key signature for data files
_wait_outlier_dump_flags			   0			     Wait Outlier Dump Flags
_wait_outlier_lambda_x1000			   1500 		     Lambda (in thousands) to compute outliers
_wait_outlier_min_waits 			   10			     Minimum waits required to enable wait outliers detection
_wait_outlier_num_outliers			   600			     Max number of outliers tracked
_xt_def_compression_ratio			   4			     Default compression ratio for external table data files
_xt_http_wscl					   FALSE		     Use KGWSCL for HTTP requests
_xt_legacy_debug_flags				   0			     Debug flags for ORACLE_LOADER and ORACLE_DATAPUMP
adg_account_info_tracking			   LOCAL		     ADG user account info tracked in standby(LOCAL) or in Primary(GLOBAL)
awr_pdb_max_parallel_slaves			   10			     maximum concurrent AWR PDB MMON slaves per instance
forward_listener							     forward listener
inmemory_automatic_level			   OFF			     Enable Automatic In-Memory management
inmemory_optimized_arithmetic			   DISABLE		     Controls whether or not DSBs are stored in-memory
inmemory_prefer_xmem_memcompress					     Prefer to store tables with given memcompress levels in xmem
inmemory_prefer_xmem_priority						     Prefer to store tables with given priority levels in xmem
inmemory_xmem_size				   0			     size in bytes of in-memory xmem area
memoptimize_pool_size				   0			     Size of cache for imoltp buffers
multishard_query_data_consistency		   strong		     consistency setting for multishard queries
multishard_query_partial_results		   not allowed		     enable partial results for multishard queries
optimizer_ignore_hints				   FALSE		     enables the embedded hints to be ignored
optimizer_ignore_parallel_hints 		   FALSE		     enables embedded parallel hints to be ignored
parallel_min_degree				   1			     controls the minimum DOP computed by Auto DOP
pdb_template								     PDB template
private_temp_table_prefix			   ORA$PTT_		     Private temporary table prefix
standby_pdb_source_file_dblink						     database link to standby source files
standby_pdb_source_file_directory					     standby source file directory location
tde_configuration							     Per-PDB configuration for Transparent Data Encryption
unified_audit_systemlog 						     Syslog facility and level for Unified Audit
wallet_root								     wallet root instance initialization parameter

356 rows selected.

废弃的参数,在12.2 中存在,可是在 18.3 中已经被驱逐的废弃参数,数量是 43 个,参考如下列表:


43 rows selected.

一点一滴,Oracle 数据库在发生大变化。

相关文章|Related Articles

16:58 Shared Pool 中 ges resource dynamic 和 ges enqueues 超高诊断 (4086 Bytes) » Oracle Life

作者:eygle 发布在 eygle.com


在自动内存管理下,Shared Pool 超高怎么办?

超高是多高呢?我们看一下下面的截图,Buffer Cache 最后被调节为只有 64M,这是最小值了,而Shared Pool Size 竟然高达 23,552M,超乎想象:


分析的思路很简单,看一看这些共享内存被谁使用了,在以下细节中,ges resource dynamic 和 ges enqueues 超高,分别占用了 6G、9G,这显然是不正常的:


这个数据库的版本是:,MOS 上很容易就可以找到匹配的 BUG:

ORA-04031 Errors Occurring with High "ges resource dynamic" & "ges enqueues" Memory Usage In The Shared Pool (文档 ID 2063751.1)

ORA-04031 errors occurring in a RAC instance with the trace file showing high memory usage for "ges resource dynamic" and "ges enqueues" memory in the shared pool. This can cause LMD processes to become unresponsive leading to an instance termination.


This is due to unpublished bug:21373473 fixed in 12.2, and occurs due to DX and BB locks being cached but not cleared out. This fix removes the need to cache DX and BB locks and hence reduces overall shared pool memory usage in RAC instances.

ORA-04031 errors occurring with high memory utilization for "ges enqueues" & "ges resource dynamic" memory allocations.


Download and apply the one-off patch number 21373473 for your platform and version combination.

Please note that if using, then you should also apply the related patch number 21260431 which also impacts the shared pool memory allocations identified here.

This issue can also be worked around by setting _GES_DIRECT_FREE_RES_TYPE="CTARAHDXBB" in the instances.

解决方案是打补丁:21373473 。客户是 版本,需要 21260431 补丁。



相关文章|Related Articles

15:54 SMCO 进程:空间管理调度进程-Space Management Coordinator (6113 Bytes) » Oracle Life

作者:eygle 发布在 eygle.com

在『云和恩墨大讲堂』微信群中,有朋友提出了一个问题,SMCO 进程是做什么用的,SMON 启动 SMCO 失败会有什么影响?



SMCO 进程是 空间管理调度进程 - Space Management Coordinator,在Oracle 11g 版本中引入。显然这不是一个核心进程,其启动失败不影响数据库的关键运行工作,可以通过参数设置禁用和启用,来重启该进程。

MOS 如下文档详细说明了这个进程的作用:

SMCO (Space Management Coordinator) For Autoextend On Datafiles And How To Disable/Enable (文档 ID 743773.1)

What is the function of SMCO background process

SMCO coordinates the following space management tasks. It performs proactive space allocation and space reclamation. It dynamically spawns slave processes (Wnnn) to implement the task.

SMCO 协调以下空间管理任务。 它执行主动空间分配和空间回收。 它动态生成从属进程(Wnnn)来实现任务。

  • 表空间级别的空间预分配 - Tablespace-level space (Extent) pre-allocation.

    Pre-allocation here refers to datafile extention, Datafile extension happens when a space request (extent allocation) operation generally triggered via Insert / loading to a segment does not find contiguous space in the tablespace, the session will extend the file by the next increment set for the datafile and will continue with the space request or Extent allocation.

    这里的预分配是指数据文件扩展,当通常通过插入/加载到段时触发的空间请求(扩展区分配)操作在表空间中找不到连续空间时会发生数据文件扩展,会话将按照 increment 设置扩展文件,同时将继续使用空间请求或Extent分配。

    For SMCO to autoextend a datafile, the AUTOEXTEND should be set to ON for the datafile. SMCO decides to expand the tablespace based on history, extension is split evenly across all datafiles in the tablespace which have not reached their maxsize and are still limited to 10% of the full tablespace size in one hourly SMCO wakeup.
    (Full tablespace size = Sum of datafile sizes at any given instance of time.)
  • 要使SMCO自动扩展数据文件,应将数据文件的AUTOEXTEND设置为ON。 SMCO决定根据历史记录扩展表空间,扩展在表空间中的所有数据文件中均匀分配,这些数据文件尚未达到其maxsize,并且在每小时SMCO唤醒时,扩展空间限制为完整表空间大小的10%。(完整表空间大小=任何给定时刻的数据文件大小总和。)

Apart from the above mentioned task, the SMCO process is also responsible for performing the following tasks.

  • updating block and extents counts in SEG$ for locally managed tablespaces after adding an extent (from unpublished Bug 12940620)
  • Securefile lob segment pre-extension.
  • Securefile lob segment in-memory dispenser space pre-allocation.
  • Securefile lob segment space reclamation (moving free chunks from uncommitted free space area to committed free space area).
  • Temporary segment space reclamation.

The main advantage is that sessions need not wait for reactive space allocation/deallocation operations as this is proactively done, and hence a gain in performance.

How to disable and enable the SMCO background process

The feature provided by the SMCO process is called "Tablespace-level space (Extent) pre-allocation."

It can be turned off by setting "_enable_space_preallocation"=0 as below:


The feature can be turned on again any time by setting "_enable_space_preallocation"=3 which is the default value:


The parameter is dynamic and the scope is ALTER SYSTEM.

Please use the double quotes as indicated to prevent an ORA-911 because of the leading '_'

Following is other possible settings:

* 0 to turn off the tbs pre-extension feature.
* 1 To enable tablespace extension.
* 2 To enable segment growth.
* 4 To enable chunk allocation.

以上摘自 MOS 内容,仅供参考。

相关文章|Related Articles

06:05 Extended Histograms (1 Bytes) » Oracle Scratchpad

2018-07-31 Tue

11:09 内核分析:使用 oradebug 转储 library_cache 内存信息 (9032 Bytes) » Oracle Life

作者:eygle 发布在 eygle.com

Shared_Pool 是 Oracle SGA中最复杂的一部分,在分析很多 library cache 问题时,经常会用到转储命令。

oradebug dump library_cache 的常用级别包括 (注意,一定要测试之后再采用,要先看看 library cache 大小,如果库缓存非常大,这个转储的日志可能会是 Huge 的):



等级4:持有对象结构 Bucket 信息,可以看到一个对象的lock,pin,mutex信息。

等级8:Level 4 + 相关数据块

等级16:Level 8 + 每个对象的 Heap 信息

等级32:Level 16 + 完整的 Heap DUMP

使用 oradebug 可以很方便的转储 library_cache 信息,示范如下:

[oracle12c@enmotech ~]$ sqlplus / as sysdba

SQL*Plus: Release Production on Tue Jul 31 10:17:00 2018

Copyright (c) 1982, 2016, Oracle. All rights reserved.

Connected to:

Oracle Database 12c Enterprise Edition Release - 64bit Production

SQL> oradebug setmypid

Statement processed.

SQL> oradebug dump library_cache 1;

Statement processed.

SQL> select value from v$diag_info where name like 'Defa%';




SQL> !

[oracle12c@enmotech ~]$ cat /oracle12c/db/diag/rdbms/enmo12c/enmo12c/trace/enmo12c_ora_7277.trc

Trace file /oracle12c/db/diag/rdbms/enmo12c/enmo12c/trace/enmo12c_ora_7277.trc

Oracle Database 12c Enterprise Edition Release - 64bit Production

Build label: RDBMS_12.

ORACLE_HOME: /u01/oracle12c/db/product/12.2.0/dbhome_1

System name: Linux

Node name: enmotech

Release: 3.10.0-514.6.2.el7.x86_64

Version: #1 SMP Thu Feb 23 03:04:39 UTC 2017

Machine: x86_64

Instance name: enmo12c

Redo thread mounted by this instance: 1

Oracle process number: 8

Unix process pid: 7277, image: oracle@enmotech (TNS V1-V3)

*** 2018-07-31T10:17:07.023214+08:00 (CDB$ROOT(1))

*** SESSION ID:(982.48265) 2018-07-31T10:17:07.023262+08:00

*** CLIENT ID:() 2018-07-31T10:17:07.023270+08:00

*** SERVICE NAME:(SYS$USERS) 2018-07-31T10:17:07.023277+08:00

*** MODULE NAME:(sqlplus@enmotech (TNS V1-V3)) 2018-07-31T10:17:07.023285+08:00

*** ACTION NAME:() 2018-07-31T10:17:07.023292+08:00

*** CLIENT DRIVER:(SQL*PLUS) 2018-07-31T10:17:07.023298+08:00

*** CONTAINER ID:(1) 2018-07-31T10:17:07.023305+08:00

Processing Oradebug command 'setmypid'

*** 2018-07-31T10:17:07.023346+08:00 (CDB$ROOT(1))

Oradebug command 'setmypid' console output: <none>

*** 2018-07-31T10:17:13.956173+08:00 (CDB$ROOT(1))

Processing Oradebug command 'dump library_cache 1'



namespace gets hit ratio pins hit ratio reloads invalids

-------------- --------- --------- --------- --------- ---------- ----------

SQL AREA 1217773 0.891 21268193 0.985 55850 35229

TABLE/PROCEDURE 881146 0.961 2039858 0.909 76152 0

BODY 192474 0.993 289193 0.994 348 0

TRIGGER 61079 0.986 61077 0.986 2 1

INDEX 152995 0.979 139263 0.816 12489 0

CLUSTER 20622 0.992 20963 0.992 0 0



从Library Cache等待事件深入剖析SQL解析 http://www.enmotech.com/web/detail/1/288/2.html

相关文章|Related Articles