Here’s a little detail I’ve just noticed in 19.11.0.0. There are a lot of parameters with names like “\_bug%”, some having proper descriptions, some simply echoing their previous pattern of things like: “tenth spare parameter – integer”.
Parameter System Value Description --------------------------------------------------------- -------------------- ---------------------------------------------------------------------- _bug27355984_xt_preproc_timeout 100 external table preprocessor timeout _bug28322973_asm_ownerid_trace_timeout 600 ownerid tracing threshold in seconds _bug28482048_fb_log_max_rereads 100 Maximum number of rereads after reading a corrupt flashback block _bug29274428_modsvc_call_out_enabled FALSE one-hundred-and-forty-third spare parameter - boolean _bug29290173_securefiles_dealloc_cfs 0 securefiles deallocate directly to CFS _bug29302220_tcpinfo_statistics_save_atexit FALSE TCP Info Statistics Save At Exit _bug29386835_enable_per_container_acl FALSE Enable Per Container ACL _bug29394014_allow_triggers_on_vpd_table FALSE Allow triggers on VPD protected table in DM _bug29504103_runtime_index_key_length_check 0 allow runtime checking of index key length _bug29558926_ctas_numfrags_hint 0 Hint for number of partition fragments to be loaded in a CTAS _bug29796916_audit_sys_privilege TRUE one-hundred-and-seventy-ninth spare parameter - boolean _bug29825525_bct_public_dba_buffer_dynresize_delay 3600 sixth spare parameter - integer _bug29903454_ksws_enable_alb TRUE enable ALB metrics processing _bug30159581_cputime_limit_parallel_fptr twenty-seventh spare parameter - integer _bug30159581_runtime_limit_parallel_fptr twenty-eighth spare parameter - integer _bug30165506_unified_pga_max_pool_size twenty-first spare parameter - integer _bug30186319_lm_hb_lgio_timeout 70 LGWR io timeout in seconds _bug30196629_dbopen_breakpoint none ALTER DATABASE OPEN codepath breakpoint _bug30224950_kjac_direct_path_enabled TRUE one-hundred-and-fifty-fifth spare parameter - boolean _bug30316897_allow_fallback_to_dbkey FALSE one-hundred-and-sixty-third spare parameter - boolean _bug30346330_hang_sess_enq_wait_blocked_session_threshold 5 tenth spare parameter - integer _bug30346330_hang_sess_enq_wait_resltn_trig_time 0 seventeenth spare parameter - integer _bug30352623_asm_enable_vfreloc_on_rm 0 Enable voting file relocation on rolling migration _bug30382982_keep_relocated_source_pdb FALSE one-hundred-and-sixty-second spare parameter - boolean _bug30570083_mv_master_lock_timeout 3 timeout that mv create/drop waits to lock the master table _bug30624792_hang_px_resolution_on_asm_enabled TRUE Hang Management px resolution on ASM enabled _bug31261467_disable_block_corruption_track FALSE one-hundred-and-seventy-eighth spare parameter - boolean _bug31304945_utlfile_dirpath FALSE return Utl_File Accessible _bug31417007_lm_mp_additional_reg_buffers 20 forty-seventh spare parameter - integer _bug31417007_lm_mp_share_lms_pools TRUE one-hundred-and-eighty-sixth spare parameter - boolean _bug31417007_lm_mp_small_lms_pools TRUE one-hundred-and-eighty-seventh spare parameter - boolean _bug31701239_hang_high_session_count_update_hb_threshold Hang Manager high session count update heartbeat threshold _bug31720839_lm_ges_rht_alloc_size 4096 ges rht allocation chunk size _bug31720839_lm_recovery_pdbw_reduction 4 recovery member pdbw reduction _bug31734583_hang_analyze_elapsed_time_limit 1800 Hang Analyze elapsed time limit in seconds _bug31747989_cputime_limit_parallel_fptr 60 CPU-time limit for for parallel first-pass recovery _bug31747989_runtime_limit_parallel_fptr 60 run-time limit for for parallel first-pass recovery _bug31823051_asm_slow_client_timeout_percentage 50 Percentage of the write cancel timeout to detect slow client. _bug32093558_j0_deq_maxwait_time 0 forty-first spare parameter - integer _bug32093558_srvntfn_max_job_time 180 fortieth spare parameter - integer _bug32095097_defer_rcv_pdb_attach TRUE Defer PDB domain attach in instance recovery _bug32171622_asm_reloc_restart_timeout 900 csecs before signaling ORA-15424 when waiting for orphans 42 rows selected.
It looks as if Oracle may be implementing a new strategy to replace the prevalence of fix_control settings to allow for improved visibility of the state of point patches. Possibly the aim is to cater for systems that have had to tweak multiple fix controls and have problems with the maximum length of the parameter value.
It’s interesting to note that 21.3.0.0 has only 11 such parameters – some new and a few matching thise in 19.11.0.0. I haven’t checked yet whether all the parameters that have disappeared from 19.11.0.0 are still in 21.3.0. with the bug number prefix removed, but I’ve spotted a couple with that appearance.
Many years ago Coskan Gundogar wrote a note about checking v$system_fix_control to see if it gave any clues about unexpected (optimizer) behaviour: the method is still appropriate, but now you have another place to check.
Although one of these parameters is now the record holder for “longest parameter name”, the longest description belongs to a hidden parameter that isn’t a “bug” parameter.
Footnote
The script to generate this list is available in this note about trace files, with most of the columns commented out. I’ve run the script from SQL*Plus using the command
start params "\_bug"
Oracle Scratchpad 
[…] are a lot of “_bug%” parameters in the latest updates of 19c (the first one appeared in 19.3.0.0), possibly taking over […]
Pingback by 12c trivia | Oracle Scratchpad — January 12, 2022 @ 11:10 am GMT Jan 12,2022 |