_parallel_adaptive_max_users | Integer | 4 | yes | SYSTEM imm. | N | Y | | maximum number of users running with default DOP |
_parallel_blackbox_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | parallel execution blackbox enabled |
_parallel_blackbox_sga | Boolean | TRUE | yes | SYSTEM imm. | N | Y | | true if blackbox will be allocated in SGA, false if PGA |
_parallel_blackbox_size | Integer | 16384 | yes | SESSION | Y | Y | | parallel execution blackbox bucket size |
_parallel_broadcast_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | enable broadcasting of small inputs to hash and sort merge joins |
_parallel_cluster_cache_pct | Integer | 80 | yes | SESSION | Y | Y | | max percentage of the global buffer cache to use for affinity |
_parallel_cluster_cache_policy | String | ADAPTIVE | yes | SESSION | Y | Y | | policy used for parallel execution on cluster(ADAPTIVE/CACHED) |
_parallel_conservative_queuing | Boolean | FALSE | yes | SYSTEM imm. | N | Y | | conservative parallel statement queuing |
_parallel_ctas_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | enable/disable parallel CTAS operation |
_parallel_default_max_instances | Integer | 1 | yes | STATIC | N | N | | default maximum number of instances for parallel query |
_parallel_execution_message_align | Boolean | FALSE | yes | STATIC | N | N | | Alignment of PX buffers to OS page boundary |
_parallel_fake_class_pct | Integer | 0 | yes | SESSION | Y | Y | | fake db-scheduler percent used for testing |
_parallel_fault_tolerance_enabled | Boolean | FALSE | yes | STATIC | N | N | | enables or disables fault-tolerance for parallel statement |
_parallel_fault_tolerance_threshold | Integer | 3 | yes | SESSION | Y | Y | | total number of faults fault-tolerance will handle |
_parallel_fixwrite_bucket | Integer | 1000 | yes | STATIC | N | N | | Number of buckets for each round of fix write |
_parallel_frz_step_reconfig | Integer | 0 | yes | SYSTEM imm. | N | Y | | parallel rs work for freeze step for instance reconfig. |
_parallel_heartbeat_snapshot_interval | Integer | 2 | yes | SYSTEM imm. | N | Y | | interval of snapshot to track px msging between instances |
_parallel_heartbeat_snapshot_max | Integer | 128 | yes | SYSTEM imm. | N | Y | | maximum number of historical snapshots archived |
_parallel_inmemory_min_time_threshold | String | AUTO | yes | SESSION | Y | Y | | threshold above which a plan is a candidate for parallelization for in-memory tables (in seconds) |
_parallel_inmemory_time_unit | Integer | 1 | yes | SESSION | Y | Y | | unit of work used to derive the degree of parallelism for in-memory tables (in seconds) |
_parallel_lmd_reconfig | Integer | 1 | yes | SYSTEM imm. | N | Y | | parallel lmd work in reconfiguration for enqueues. |
_parallel_load_bal_unit | Integer | 0 | yes | SESSION | Y | Y | | number of threads to allocate per instance |
_parallel_load_balancing | Boolean | TRUE | yes | SESSION | Y | Y | | parallel execution load balanced slave allocation |
_parallel_load_publish_threshold | Integer | 10 | yes | SESSION | Y | Y | | diffrence in percentage controlling px load propagation |
_parallel_log_write_cleanup | Boolean | TRUE | yes | STATIC | N | N | | Perform (null) log write cleanup in parallel |
_parallel_min_message_pool | String | 491520 | yes | STATIC | N | N | | minimum size of shared pool memory to reserve for pq servers |
_parallel_optimization_phase_for_local | Boolean | FALSE | yes | SESSION | Y | Y | | parallel optimization phase when all slaves are local |
_parallel_queuing_max_waitingtime | Integer | | yes | SYSTEM imm. | N | Y | | parallel statement queuing: max waiting time in queue |
_parallel_recovery_stopat | Integer | 32767 | yes | SYSTEM imm. | N | Y | | stop at -position- to step through SMON |
_parallel_replay_msg_limit | Integer | 4000 | yes | STATIC | N | N | | Number of messages for each round of parallel replay |
_parallel_scalability | Integer | 50 | yes | SESSION | Y | Y | | Parallel scalability criterion for parallel execution |
_parallel_server_idle_time | Integer | 30000 | yes | SYSTEM imm. | N | Y | | idle time before parallel query server dies (in 1/100 sec) |
_parallel_server_memory_threshold | Big integer | 0 | yes | SYSTEM imm. | N | Y | | max memory threshold for idle PX server to exit |
_parallel_server_resiliency | Integer | 0 | yes | SESSION | Y | Y | | enable resilient execution of parallel query |
_parallel_server_sleep_time | Integer | 1 | yes | SYSTEM imm. | N | Y | | sleep time between dequeue timeouts (in 1/100ths) |
_parallel_slave_acquisition_wait | Integer | 1 | yes | SESSION | Y | Y | | time(in seconds) to wait before retrying slave acquisition |
_parallel_statement_queuing | Boolean | FALSE | yes | SESSION | Y | Y | | parallel statement queuing enabled |
_parallel_syspls_obey_force | Boolean | TRUE | yes | SESSION | Y | Y | | TRUE to obey force parallel query/dml/ddl under System PL/SQL |
_parallel_time_unit | Integer | 10 | yes | SESSION | Y | Y | | unit of work used to derive the degree of parallelism (in seconds) |
_parallel_txn_global | Boolean | FALSE | yes | SESSION | N | N | | enable parallel_txn hint with updates and deletes |
_parallelism_cost_fudge_factor | Integer | 350 | yes | SESSION | Y | Y | | set the parallelism cost fudge factor |
_px_adaptive_dist_bypass_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | enable the bypass table queue for adaptive distribution methods |
_px_adaptive_dist_bypass_optimization | String | 1 | yes | SESSION | Y | Y | | sets optimizations for bypass table queue for ADM |
_px_adaptive_dist_method | String | CHOOSE | yes | SESSION | Y | Y | | determines the behavior of adaptive distribution methods |
_px_adaptive_dist_method_threshold | Integer | 0 | yes | SESSION | Y | Y | | Buffering / decision threshold for adaptive distribution methods |
_px_adaptive_dist_nij_enabled | String | ON | yes | SESSION | Y | Y | | enable adaptive distribution methods for left non-inner joins |
_px_adaptive_offload_percentage | Integer | 30 | yes | SESSION | Y | Y | | percentage for PQ adaptive offloading of granules |
_px_adaptive_offload_threshold | Integer | 10 | yes | SESSION | Y | Y | | threshold (GB/s) for PQ adaptive offloading of granules |
_px_autodop_pq_overhead | Boolean | TRUE | yes | SESSION | Y | Y | | adjust auto dop calculation using pq overhead |
_px_back_to_parallel | String | OFF | yes | SESSION | Y | Y | | allow going back to parallel after a serial operation |
_px_bind_peek_sharing | Boolean | TRUE | yes | SESSION | Y | Y | | enables sharing of px cursors that were built using bind peeking |
_px_broadcast_fudge_factor | Integer | 100 | yes | SESSION | Y | Y | | set the tq broadcasting fudge factor percentage |
_px_buf_subheap_size | Integer | 4096 | yes | STATIC | N | N | | buffer subheap size |
_px_buffer_ttl | Integer | 30 | yes | SESSION | Y | Y | | ttl for px mesg buffers in seconds |
_px_cdb_view_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | parallel cdb view evaluation enabled |
_px_cdb_view_join_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | disable parallelism cap on CDB view |
_px_chunk_size | Integer | 1000000 | yes | SESSION | Y | Y | | default size of a chunk for row id granules (in KB) |
_px_chunklist_count_ratio | Integer | 32 | yes | STATIC | N | N | | ratio of the number of chunk lists to the default DOP per instance |
_px_compilation_debug | Integer | 0 | yes | SESSION | Y | Y | | debug level for parallel compilation |
_px_compilation_trace | Integer | 0 | yes | SESSION | Y | N | | tracing level for parallel compilation |
_px_concurrent | Boolean | TRUE | yes | SESSION | Y | Y | | enables pq with concurrent execution of serial inputs |
_px_cpu_autodop_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | enables or disables auto dop cpu computation |
_px_cpu_operator_bandwidth | String | | yes | SESSION | Y | Y | | CPU operator bandwidth in MB/sec for DOP computation |
_px_cpu_process_bandwidth | Integer | 200 | yes | SESSION | Y | Y | | CPU process bandwidth in MB/sec for DOP computation |
_px_dist_agg_partial_rollup_pushdown | String | ADAPTIVE | yes | SESSION | Y | Y | | perform distinct agg partial rollup pushdown for px execution |
_px_dp_array_size | Integer | 32767 | yes | STATIC | N | N | | Max number of pq processes supported |
_px_dump_12805_source | Boolean | TRUE | yes | SESSION | Y | Y | | enables or disables tracing of 12805 signal source |
_px_dup_proc_check | Boolean | FALSE | yes | SYSTEM imm. | N | Y | | enable duplicate parallel process check with name service |
_px_dynamic_granules | Boolean | TRUE | yes | SESSION | Y | Y | | enable dynamic granule generation |
_px_dynamic_granules_adjust | Integer | 10 | yes | SESSION | Y | Y | | adjust dynamic granule regeneration |
_px_dynamic_opt | Boolean | TRUE | yes | SESSION | Y | Y | | turn off/on restartable qerpx dynamic optimization |
_px_dynamic_sample_size | Integer | 50 | yes | SESSION | Y | Y | | num of samples for restartable qerpx dynamic optimization |
_px_execution_services_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | enable service-based constraint of px slave allocation |
_px_extended_join_skew_handling | Boolean | TRUE | yes | SESSION | Y | Y | | enables extended skew handling for parallel joins |
_px_external_table_default_stats | Boolean | TRUE | yes | SESSION | Y | Y | | the external table default stats collection enable/disable |
_px_filter_parallelized | Boolean | TRUE | yes | SESSION | Y | Y | | enables or disables correlated filter parallelization |
_px_filter_skew_handling | Boolean | TRUE | yes | SESSION | Y | Y | | enable correlated filter parallelization to handle skew |
_px_flood_ctrl_time | Integer | 900 | yes | SESSION | Y | Y | | parallel flood control time in seconds |
_px_freelist_latch_divisor | Integer | 2 | yes | STATIC | N | N | | Divide the computed number of freelists by this power of 2 |
_px_gim_factor | Integer | 100 | yes | SESSION | Y | Y | | weighted autodop global index maintenance factor |
_px_granule_alignment | Integer | 1024 | yes | SESSION | Y | Y | | default size of a rowid range granule (in KB) |
_px_granule_batch_size | Integer | 26 | yes | SESSION | Y | Y | | maximum size of a batch of granules |
_px_granule_randomize | Boolean | TRUE | yes | SESSION | Y | Y | | enables or disables randomization of parallel scans rowid granules |
_px_granule_size | Integer | 1000000 | yes | SESSION | Y | Y | | default size of a rowid range granule (in KB) |
_px_groupby_pushdown | String | FORCE | yes | SESSION | Y | Y | | perform group-by pushdown for parallel query |
_px_hold_time | Integer | 0 | yes | SESSION | Y | Y | | hold px at execution time (unit: second) |
_px_hybrid_TSM_HWMB_load | Boolean | TRUE | yes | SESSION | Y | Y | | Enable Hybrid Temp Segment Merge/High Water Mark Brokered load method |
_px_hybrid_partition_execution_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | enable parallel hybrid partition execution |
_px_hybrid_partition_skew_threshold | Integer | 10 | yes | SESSION | Y | Y | | partitions bigger than threshold times average size are skewed |
_px_increase_join_frag_size | Boolean | TRUE | yes | SYSTEM imm. | N | Y | | increase join message fragment size |
_px_index_sampling_objsize | Boolean | TRUE | yes | SESSION | Y | Y | | parallel query sampling for index create based on object size |
_px_io_process_bandwidth | Integer | 200 | yes | SESSION | Y | Y | | IO process bandwidth in MB/sec for computing DOP |
_px_io_system_bandwidth | Integer | 0 | yes | SESSION | Y | Y | | total IO system bandwidth in MB/sec for computing DOP |
_px_join_skew_handling | Boolean | TRUE | yes | SESSION | Y | Y | | enables skew handling for parallel joins |
_px_join_skew_minfreq | Integer | 30 | yes | SESSION | Y | Y | | sets minimum frequency(%) for skewed value for parallel joins |
_px_join_skew_null_handling | Boolean | TRUE | yes | SESSION | Y | Y | | enables null skew handling improvement for parallel outer joins |
_px_join_skew_ratio | Integer | 10 | yes | SESSION | Y | Y | | sets skew ratio for parallel joins |
_px_join_skew_sampling_percent | Integer | 0 | yes | SESSION | Y | Y | | Sets sampling percentage for skew handling sampling queries |
_px_join_skew_sampling_time_limit | Integer | 50 | yes | SESSION | Y | Y | | Sets execution time limit for skew handling sampling queries |
_px_join_skew_use_histogram | Boolean | TRUE | yes | SESSION | Y | Y | | Enables retrieval of skewed values from histogram when possible |
_px_join_skewed_values_count | Integer | 0 | yes | SESSION | Y | Y | | force join skew handling with specified skewed values count |
_px_kxib_tracing | Integer | 0 | yes | SESSION | Y | Y | | turn on kxib tracing |
_px_load_balancing_policy | String | UNIFORM | yes | SESSION | Y | Y | | parallel load balancing policy |
_px_load_factor | Integer | 300 | yes | SESSION | Y | Y | | weighted autodop load factor |
_px_load_monitor_threshold | Integer | 10000 | yes | SESSION | Y | Y | | threshold for pushing information to load slave workload monitor |
_px_load_publish_interval | Integer | 200 | yes | SESSION | Y | Y | | interval at which LMON will check whether to publish PX load |
_px_loc_msg_cost | Integer | 1000 | yes | SESSION | Y | Y | | CPU cost to send a PX message via shared memory |
_px_max_granules_per_slave | Integer | 100 | yes | SESSION | Y | Y | | maximum number of rowid range granules to generate per slave |
_px_max_map_val | Integer | 32 | yes | SYSTEM imm. | N | Y | | Maximum value of rehash mapping for PX |
_px_max_message_pool_pct | Integer | 40 | yes | STATIC | N | N | | percentage of shared pool for px msg buffers range [5,90] |
_px_message_compression | Boolean | TRUE | yes | SESSION | Y | Y | | enable compression of control messages for parallel query |
_px_min_granules_per_slave | Integer | 13 | yes | SESSION | Y | Y | | minimum number of rowid range granules to generate per slave |
_px_min_ndv_per_slave_perc_func | Integer | 2 | yes | SESSION | Y | Y | | mininum NDV of TQ keys needed per slave for parallel percentile |
_px_minus_intersect | Boolean | TRUE | yes | SESSION | Y | Y | | enables pq for minus/interect operators |
_px_net_msg_cost | Integer | 10000 | yes | SESSION | Y | Y | | CPU cost to send a PX message over the internconnect |
_px_nlj_bcast_rr_threshold | Integer | 10 | yes | SESSION | Y | Y | | threshold to use broadcast left random right distribution for NLJ |
_px_no_granule_sort | Boolean | FALSE | yes | SESSION | Y | Y | | prevent parallel partition granules to be sorted on size |
_px_no_stealing | Boolean | FALSE | yes | SESSION | Y | Y | | prevent parallel granule stealing in shared nothing environment |
_px_nss_planb | Boolean | TRUE | yes | SESSION | Y | Y | | enables or disables NSS Plan B reparse with outline |
_px_numa_stealing_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | enable/disable PQ granule stealing across NUMA nodes |
_px_numa_support_enabled | Boolean | FALSE | yes | SYSTEM imm. | N | Y | | enable/disable PQ NUMA support |
_px_object_sampling | Integer | 200 | yes | SESSION | Y | Y | | parallel query sampling for base objects (100000 = 100%) |
_px_object_sampling_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | use base object sampling when possible for range distribution |
_px_object_sampling_multiplier | Integer | 2 | yes | SESSION | Y | Y | | number of samples to send to QC per slave is DOP * prmixsm |
_px_onepass_slave_acquisition | Boolean | TRUE | yes | SESSION | Y | Y | | enable/disable one pass slave acquisition for parallel execution |
_px_oneslave_per_instance_parse | Boolean | TRUE | yes | SESSION | Y | Y | | send parse to one slave per instance |
_px_overhead_exec_ctrlmesg | Integer | 4 | yes | SESSION | Y | Y | | pq overhead on sending one control message during execution (in microseconds) |
_px_overhead_init_endpoints | Integer | 4 | yes | SESSION | Y | Y | | pq overhead on setting up one end point during initialization (in microseconds) |
_px_overhead_init_slavealloc | Integer | 1000 | yes | SESSION | Y | Y | | pq overhead on allocating per slave during initialization (in microseconds) |
_px_overhead_teardown | Integer | 600 | yes | SESSION | Y | Y | | pq overhead on setting up one end point during initialization (in microseconds) |
_px_parallel_spawn_min_count | Integer | 2 | yes | STATIC | N | N | | PQ parallel spawn min threshold count |
_px_parallelize_expression | Boolean | TRUE | yes | SESSION | Y | Y | | enables or disables expression evaluation parallelization |
_px_parallelize_non_native_datatype | Boolean | TRUE | yes | SESSION | Y | Y | | enable parallelization for non-native datatypes |
_px_partial_rollup_pushdown | String | ADAPTIVE | yes | SESSION | Y | Y | | perform partial rollup pushdown for parallel execution |
_px_partition_load_dist_threshold | Integer | 64 | yes | SESSION | Y | Y | | # of load slave to # of partition ratio to switch distribution |
_px_partition_load_skew_handling | String | OFF | yes | SESSION | Y | Y | | enable partition load skew handling |
_px_partition_load_skew_threshold | Integer | 3 | yes | SESSION | Y | Y | | partition loads bigger than threshold times average are skewed |
_px_partition_scan_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | enables or disables parallel partition-based scan |
_px_partition_scan_threshold | Integer | 64 | yes | SESSION | Y | Y | | least number of partitions per slave to start partition-based scan |
_px_partition_skew_threshold | Integer | 80 | yes | SESSION | Y | Y | | percentage of table size considered as threshold for skew determination |
_px_proactive_slave_alloc_threshold | Integer | 8 | yes | SESSION | Y | Y | | parallel proactive slave allocation threshold/unit |
_px_proactive_slave_release | Integer | 1 | yes | SESSION | Y | Y | | early release of producer slaves of final DFO |
_px_proc_constrain | Boolean | TRUE | yes | SYSTEM imm. | N | Y | | reduce parallel_max_servers if greater than (processes - fudge) |
_px_pwg_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | parallel partition wise group by enabled |
_px_pwise_wif_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | enable parallel partition wise window function |
_px_pwmr_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | parallel partition wise match recognize enabled |
_px_replication_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | enables or disables replication of small table scans |
_px_reuse_server_groups | String | MULTI | yes | SESSION | Y | Y | | enable/disable reusing of server groups that are already acquired |
_px_round_robin_rowcnt | Integer | 1000 | yes | SESSION | Y | Y | | round robin row count to enq to next slave |
_px_rownum_pd | Boolean | TRUE | yes | SESSION | Y | Y | | turn off/on parallel rownum pushdown optimization |
_px_scalable_gby_invdist | Boolean | TRUE | yes | SESSION | Y | Y | | scalable PX plan for GBY w/ inverse distribution functions |
_px_scalable_invdist | Boolean | TRUE | yes | SESSION | Y | Y | | enable/disable px scalable plan for inverse distribution functions |
_px_scalable_invdist_mcol | Boolean | TRUE | yes | SESSION | Y | Y | | enable/disable px plan for percentile functions on multiple columns |
_px_send_timeout | Integer | 300 | yes | SESSION | Y | Y | | IPC message send timeout value in seconds |
_px_shared_hash_join | Boolean | FALSE | yes | SESSION | Y | Y | | enable/disable shared hash join |
_px_single_server_enabled | Boolean | TRUE | yes | SESSION | Y | Y | | allow single-slave dfo in parallel query |
_px_slave_parse_first_with_outline_hints | Boolean | TRUE | yes | SESSION | Y | Y | | Enable/Disable trying outlined plan first |
_px_split_batches_per_slave | Integer | 7 | yes | SESSION | Y | Y | | number of split granules per slave for exadoop |
_px_split_multi_msg | Boolean | TRUE | yes | SESSION | Y | Y | | use multiple messages to send batch of split granules |
_px_split_use_single_list | Boolean | TRUE | yes | SESSION | Y | Y | | use single list for split granules |
_px_tq_memcpy_threshold | Integer | 100 | yes | SESSION | Y | Y | | threshold for PQ TQ to use memcpy for packing columns |
_px_tq_rowhvs | Boolean | TRUE | yes | SESSION | Y | Y | | turn on intra-row hash valueing sharing in TQ |
_px_trace | String | none | yes | SESSION | N | N | | px trace parameter |
_px_ual_serial_input | Boolean | TRUE | yes | SESSION | Y | Y | | enables new pq for UNION operators |
_px_wif_dfo_declumping | String | CHOOSE | yes | SESSION | Y | Y | | NDV-aware DFO clumping of multiple window sorts |
_px_wif_extend_distribution_keys | Boolean | TRUE | yes | SESSION | Y | Y | | extend TQ data redistribution keys for window functions |
_px_wif_min_ndv_per_slave | Integer | 2 | yes | SESSION | Y | Y | | mininum NDV of TQ keys needed per slave for scalable WiF PX |
_px_window_skew_handling | Boolean | TRUE | yes | SESSION | Y | Y | | enable window function skew handling |
_px_window_skew_sample_size | Integer | 1000 | yes | SESSION | Y | Y | | window function skew sample size |
_px_xt_granule_sampling_ratio | Integer | 200 | yes | SESSION | Y | Y | | range table queue object sampling ratio for external table |
_px_xtgranule_size | Integer | 10000 | yes | SESSION | Y | Y | | default size of a external table granule (in KB)
|
No comments:
Post a Comment