NAME | VALUE |
---|---|
_b_tree_bitmap_plans | FALSE |
_fast_full_scan_enabled | FALSE |
_like_with_bind_as_equality | TRUE |
_optimizer_autostats_job | FALSE |
_sort_elimination_cost_ratio | 5 |
_sqlexec_progression_cost | 2147483647 |
_system_trig_enabled | TRUE |
_use_single_log_writer | TRUE |
aq_tm_processes | 4 |
audit_sys_operations | TRUE |
audit_trail | DB (default is NONE) |
compatible | 12.1.0 |
control_file_record_keep_time | 30 |
control_files | /path/to1/ctrl01.dbf, /path/to2/ctrl02.dbf, /path/to3/ctrl03.dbf |
cursor_sharing | EXACT |
db_block_checking | FALSE |
db_block_checksum | TRUE |
db_block_size | 8192 |
db_cache_size | 5368709120 (or, 5G) |
db_files | 1000 |
db_name | EBSPROD |
diagnostic_dest | /path/to/db_ebsprod/admin/EBSPROD_ebsdb1p |
dml_locks | 10000 |
filesystemio_options | setall |
java_jit_enabled | FALSE |
job_queue_processes | 20 |
local_listener | (DESCRIPTION=(ADDRESS=(PROTOCOL=IPC)(KEY=EBSPROD_ipc))) |
log_archive_dest | /path/to/EBSPROD/arch/arch.log |
log_buffer | 10485760 (or, 10M) |
log_checkpoint_interval | 100000 |
log_checkpoint_timeout | 1200 |
log_checkpoints_to_alert | TRUE |
max_dump_file_size | 10000 |
memory_max_target | 25769803776 (or, 24G) |
memory_target | 25769803776 |
nls_comp | binary |
nls_date_format | DD-MON-RR |
nls_language | AMERICAN |
nls_length_semantics | BYTE |
nls_numeric_characters | ., |
nls_sort | binary |
nls_territory | AMERICA |
olap_page_pool_size | 4194304 (or, 4M) |
open_cursors | 4096 |
optimizer_adaptive_features | FALSE |
optimizer_secure_view_merging | FALSE |
os_authent_prefix | ext$ |
parallel_max_servers | 64 |
parallel_min_servers | 0 |
parallel_servers_target | 64 |
parallel_threads_per_cpu | 2 |
pga_aggregate_limit | 0 |
pga_aggregate_target | 0 (this is the default. Could be set to 1G) |
plsql_code_type | INTERPRETED |
plsql_optimize_level | 0 |
processes | 3000 |
query_rewrite_enabled | true |
recovery_parallelism | 1 |
recyclebin | ON |
remote_login_passwordfile | NONE |
remote_os_authent | FALSE |
sec_case_sensitive_logon | FALSE |
sec_protocol_error_further_action | drop,3 |
session_cached_cursors | 500 |
session_max_open_files | 100 |
sessions | 4536 |
sga_max_size | 25769803776 (or, 24G) |
shared_pool_reserved_size | 214748364 |
shared_pool_size | 4294967296 (or, 4G) |
sql92_security | TRUE |
temp_undo_enabled | TRUE |
timed_statistics | TRUE |
undo_management | AUTO |
undo_retention | 43200 |
undo_tablespace | APPS_UNDO_TBS |
utl_file_dir | /path/to/EBSPROD/utl_dir, /usr/tmp |
workarea_size_policy | AUTO |
The database server is an IBM LPAR configuration. 4 Oracle databases on it share its resources of 6 cores CPUs and 92 GB memory, and it is all virtual.
parallel_max_servers = Maximum value should be 2 X #of CPUs.
This parameter depends on the number of cores. It uses logical CPUs and 5 cores (4-way SMP). The SMP threads don’t really replicate a full core.
optimizer_adaptive_features = FALSE
If it is set to true, queries may run OK on a first run, but poorly on subsequent runs. Oracle will try to adapt the plans using cardinality feedback and other features that can be good, but can also cause issues.
parallel_threads_per_cpu = 2
Because the core CPU are 4-way SMP, it shows 4x more CPUs than it really has. If it set to a higher value (say, 16), Oracle will keep overallocating the CPU resources for parallel queries and spinning up too many parallel threads for the memory and CPU that it has.
No comments:
Post a Comment