ORACLE DATABASE 10GR2:

AN ENLIGHTENED REVISIT
(BEFORE WE GIVE UP AND MOVE TO
11G!)

John Kanagaraj, DB Soft Inc

Expert Session #319 (2 hour)
Speaker Qualifications
• John is a Principal Consultant @ DB Soft Inc.
• Executive Editor for IOUG’s SELECT Journal
• Co-author of “Oracle Database 10g Insider Solutions”
• Technical Editor for various books
• Frequent presenter – IOUG/OAUG/OOW/NoCOUG
• Published in SELECT, OAUG Insight, SQL Server
Magazine and other publications
• Recognized by Oracle Corp as an “Oracle ACE”
SELECT: Call for Articles/Reviewers
– The SELECT Journal is IOUG’s Technical Quarterly
– Distributed to all IOUG members worldwide
– Yearly Best Practices and Tips booklet
• Submit an article or Review one!
• Contact ‘[email protected]’
What this presentation is about
• Oracle Database 10g now in “full production” mode
• R1 released in 2003 (?), R2 in 2005 (?)
• A little history lesson – book, papers
• Lead time between books/initial articles and
“consolidation” / “better understanding”
• Need to revisit and update the “first revelation”
• Reconsider “new features” – fresh look
• Hidden surprises and little bonuses
• A peek into the future – 11g titbits
Audience survey
• Using Oracle Database 11g in production?
• Still using Oracle 7 or 8i?
• Migrating to Oracle Database 10g?
• Read the “New features” manual for every release?
• Used Advisories in 9i?
• Used Active Session History for troubleshooting?
• Read AWR/ADDM/ASH reports?
• Used SQL Profiles to fix SQL issues?
Overview of new perf features
• Strictly related to performance related features
• Need to “unlearn the old and embrace the new”
• AWR – The Performance Warehouse
• ASH – What happened to the sessions?!?
• ADDM – Your inbuilt (and unpaid!) expert
• Tuning advisors – Bonus freebies
• In-Memory metrics and Server generated alerts
• Where to look for more details
Philosophy behind Oracle DB 10g
• Automation
– Incremental steps in 9i (Advisors, Time)
– Most significant change in Performance
management
– “Out of the box” setups
– GUI “hides” the complexity (and details!)
• Consistency
– Common/Unified interface
– Stats storage and presentation
– Interpretation
Reality check before we proceed
• Licensing changes in Oracle Database 10g
• A tale of two packs:
– Oracle Diagnostic Pack
• Automatic Workload Repository (AWR)
• Automatic Database Diagnostic Monitor (ADDM)
• Monitoring and Event notifications and history
– Oracle Tuning Pack
• SQL Access Advisor
• SQL Tuning Advisor and STS
• Object Advisor
• Required even to access Views directly!!!
AWR – Performance Warehouse
• Performance Data Warehouse for 10g
• Basis for most of the problem detection and reporting
• AWR collects, stores performance data
– Direct memory access (MMNL/MMON)
– In-memory component (V$/Metric views)
– “Persisted” in WR tables (SYSAUX)
– 162 tables – WRI$, WRH$, WRM$
– Exposed via DBA_HIST_* Views
• Self managing “out of the box”
• Set retention, frequency, baseline
AWR Contents
• Active Session History (ASH)
• High-load SQL statements
• Time model statistics (both System/Session)
• Object usage - access counts for segments
• Snapshots of V$ and some Metrics
• Host CPU and Memory statistics (V$OSSTAT)
• Baseline information
• “Management” type information
– Snapshot details
– Advisor and other parameters
AWR – “Statspack on Steroids”
• Similar to STATSPACK “snapshots”
• Reportable – AWRRPT.SQL
• AWR snapshot automatically analyzed
• Accessible via GUI and API/SQL (*)
• High-impact SQL captured differently
• Stores session level info as well

* Read the fine print : License required, even for SQL access!
AWR Baselining and comparison
• Enables performance “baselining”
• Collection of two or more snapshots
• Stored in “_BL” tables; data not purged
• View using WRM$_BASELINE/DBA_HIST_BASELINE
• Reports diff via AWRDDRPT.SQL
• Execution statistics for specific SQL statement using
AWRSQRPT.SQL
ASH – What’s up with sessions
• Historical view of active sessions
• Active sessions sampled every second
• Stored in circular memory buffer
• Every 10th sample persisted in AWR
• V$ACTIVE_SESSION_HISTORY : “In-memory”
• WRH$_ACTIVE_SESSION_HISTORY : “Persisted”
• Enables “after-the-fact” analysis!!!
• Reported via ASHRPT (Not available in 10gR1)
• “Slice-and-dice” analysis can reveal a lot of info
ASH – Session states exposed!
• “On-the-spot” analysis
• Retroactive analysis
– From memory buffer (V$ACTIVE_SESSION_HISTORY)
– From persisted AWR data
(WRH$_ACTIVE_SESSION_HISTORY connected via
SNAP_ID)
– Supports manual drill down from AWR/ADDM
• Tracks High load SQL execution behavior
• Determine Blocking sessions and “hot” segments
• SESSION_STATE : “ON CPU” or “WAITING”
ADDM – Your unpaid Tuning Expert!
• Starting point for most investigations
• Runs after every AWR snapshot
• Determines and records performance issue
• Recommends corrective action
• Generates probable benefit
• Suggest use of other advisors
• Common currency - “DB Time” (qualitative!)
• Oracle DB 11g – new “Instance ADDM”
ADDM – Partial check list
• CPU bottlenecks
• Excessive parsing
• Lock contention
• Concurrency
• I/O capacity
• Incorrect sizing of Oracle memory and file structures
• High-load SQL, Java and PL/SQL statements
• Poor connection management
• Hot objects
• RAC-specific issues
ADDM – Findings/Recommendations
• Qualitative rather than Quantitative analysis
• Hardware changes
• Database-configuration changes
• Schema-level changes
• Application changes
• Using other advisors (for example)
– SQL Tuning Advisor / SQL Access Advisor
– Segment Advisor
• Don’t stare at the screen – Use SQL to summarize
– Details in my 2007 paper
Advisors – A step beyond
• Builds on 9i advisors
– Buffer cache advisor
– Shared pool advisor
– MTTR (Mean Time To Recover) advisor
– Summary (MVIEW) advisor
– PGA Target Advisor
• New in 10g
– SQL Tuning Advisor
– SQL Access Advisor
– Segment Advisor
SQL Tuning Advisor
• Frontend to Automatic Tuning Optimizer
• Extension (reuse) of Optimizer (CBO)
• Performs “what-if” analysis
• Not restricted by “time to optimize”
(_optimizer_max_permutations = 2000)
• The following advice is provided
– Gather missing or stale statistics
– Create new indexes
– Restructure SQL statement
– SQL profiles
SQL Tuning Advisor
• SQL Profile
– Collects additional information via sampling/partial
execution techniques
– Verifies and adjusts CBO’s estimates at runtime
– Similar in function to Outlines
– Enabled by category : “test-and-set”
– Access/manipulate – DBMS_SQLTUNE
– Precedence given to Stored Outlines
– Runs against individual SQL or SQL Tuning Sets
(STS)
Other Advisors
• SQL Access Advisor
– Works alongside SQL Tuning Advisor
– Advice on MV, Indexes, MV logs
– Considers space usage vs performance
– Inputs: STS, User-defined, Hypothetical
– Advanced: Workload type (RO), Drop unused indexes,
Filters (Top N, Module)
• Segment Advisor
• Undo Advisor
• Memory Advisor
• Metrics and Server Generated Alerts
Avoiding Advisor Pitfalls

• Out-of-the-box thinking (redesign; rethink approach)

• False positives (check validity for all situations – e.g.
Index non-usage)
• Changing workload or environment (additional load,
new code, H/W or S/W changes)
Features revisited
• STATSPACK – Not dead yet!
• New and changed views – tons of goodies!
• Event Breakout – Greater visibility
• Mutexes – Boon or Bane?
• Operating System Statistics – all in one place
• Time and Wait Model – aids better understanding
• Hidden Surprises – watch out!
• Useful AWR sections – Segstats and much more
• Extracting Data from AWR – Drill down example
• DBMS_XPLAN – in-depth view
STATSPACK – Not dead yet!
• STATSPACK is still available in Oracle DB 10g
• No license required to track performance stats
• Setup/maintenance/reporting : same as before
– Default is Level 5, set at Level 7 to get Segment stats
• Number of new sections available
– OS Statistics
– DB Time reporting
– Response time Histograms (Event, File and Temp)
– Additional SQL sections in Report
• Comparison to AWR
• Check Metalink Note: 394937.1 (incomplete)
STATSPACK – Host Statistics
Host CPU (CPUs: 32)
~~~~~~~~ Load Average
Begin End User System Idle WIO WCPU
------- ------- ------- ------- ------- ------- --------
0.04 0.03 3.38 1.06 95.56 3.13 #######

Instance CPU
~~~~~~~~~~~~
% of total CPU for Instance: 3.64
% of busy CPU for Instance: 81.90
%DB time waiting for CPU - Resource Mgr:

Memory Statistics Begin End

~~~~~~~~~~~~~~~~~ ------------ ------------
Host Mem (MB): 65,274.7 65,274.7
SGA use (MB): 6,144.0 6,144.0
PGA use (MB): 2,038.8 2,233.6
% Host Mem used for SGA+PGA: 12.5 12.8
-------------------------------------------------------------
STATSPACK – Host Statistics
• Calculation of Total and Busy CPU (spcpkg.sql)
% of total CPU for Instance:' ch45n, 100*
((:dbcpu+:bgcpu)/1000000) / (:ttics) pctval
% of busy CPU for Instance:' ch45n, 100*
((:dbcpu+:bgcpu)/1000000) / ((:btic)/100) pctval
• Available in Time Model (V$SYS_TIME_MODEL)
• Discrepancy between estimated and actual CPU
– (Busy_Time + Idle_Time) vs. (No.of.secs * No. of CPUs)
– Usually present when load is high (queue depth/wait for
CPU)
• Note memory statistics – tracked in V$OSSTAT
• This section not reported in AWR Report!!!
STATSPACK – Response Histograms
Wait Event Histogram DB/Inst: MYRAC/MYRAC1 Snaps: begin_snap-end_snap
-> Total Waits - units: K is 1000, M is 1000000, G is 1000000000
-> % of Waits - column heading: <=1s is truly <1024ms, >1s is truly >=1024ms
-> % of Waits - value: .0 indicates value was &lt.05%, null is truly 0
-> Ordered by Event (idle events last)

Total ----------------- % of Waits ------------------

Event Waits <1ms <2ms <4ms <8ms <16ms <32ms <=1s >1s
-------------------------- ----- ----- ----- ----- ----- ----- ----- ----- -----
enq: TX - row lock content 7561 4.1 .0 .0 .1 .1 95.8
enq: WF - contention 7 57.1 42.9
enq: WL - contention 4 75.0 25.0
gc buffer busy 1893 74.2 4.5 3.4 11.5 3.4 1.2 1.7
gc cr block busy 33K 21.9 8.4 3.5 5.0 5.0 27.3 28.7
gc cr block congested 117 3.4 23.1 53.8 18.8 .9
gc cr disk read 1626 99.6 .4 .1
gc cr failure 1 100.0
<snip>
SQL*Net break/reset to cli 2276 99.3 .5 .1 .2
SQL*Net message from dblin 683 43.6 25.0 14.5 4.2 6.4 2.0 4.1
SQL*Net message to dblink 683 100.0
<snip>
db file scattered read 2070K 91.4 7.3 .7 .5 .1 .0 .0
db file sequential read 1413K 96.4 1.2 .6 1.5 .3 .0 .0
STATSPACK – Response Histograms
• Response Time Histograms
• Snapped for Event, File and Temp
– V$EVENT_HISTOGRAM
– V$FILE_HISTOGRAM
– V$TEMP_HISTOGRAM
• Helpful in seeing spread, bumps in wait events
– Troubleshooting network response (example later)
• SQL*Net break/reset to client - Client traffic
• SQL*Net message to client (and “to dblink”) – Inter server
• gc cr block congested (Internode traffic) – Intra instance
• Not present in AWR (neither snapped or reported)
STATSPACK – New SQL sections
• New SQL reporting sections in Oracle DB 10g SP
– SQL ordered by CPU : From CPU_TIME in V$SQL
– SQL ordered by Elapsed: From ELAPSED_TIME in V$SQL
– SQL ordered by Cluster Wait Time: From new
CLUSTER_WAIT_TIME in 10g V$SQL
• Report compares total time for SQL statements captured to total
time reported in the DB
SQL ordered by CPU DB/Inst: MYRAC/MYRAC1 Snaps: begin_snap-end_snap
-> Resources reported for PL/SQL code includes the resources used by all SQL
statements called by the code.
-> Total DB CPU (s): 1,826
-> Captured SQL accounts for 202.4% of Total DB CPU
-> SQL reported below exceeded 1.0% of Total DB CPU

CPU CPU per Elapsd

Time (s) Executions Exec (s) %Total Time (s) Buffer Gets
---------- ------------ ---------- ------ ---------- ---------------
947.13 437 2.17 51.9 3503.86 29,619,674
select q_name, state, delay, expiration, rowid, msgid, dequeue
_msgid, chain_no, local_order_no, enq_time, enq_tid, step_no,
New and changed views
• Large number of new as well as changed views
• (Sometimes) Inadequate documentation, but
powerful uses!
• Classified by
– Services related
– In-memory metrics
– Response Histograms
– Extensions to existing views
– Interesting views
– Common columns
Service related views
• “Service” concept created in Oracle 8.0
• Provides ability to connection to an abstract entity
• SERVICE_NAME column in various views
• New views
– V$SERVICES : Services description (failover and load
balancing information included)
– V$SERVICE_STATS : Workload statistics
– V$SERVICE_EVENT : Events by service
Service related views
SQL> select name, goal, dtp, aq_ha_notification, clb_goal
2* from v$services;

NAME GOAL D AQ_HA_NOTIFICATION CLB_GOAL

-------------------- ---------- - ------------------ ----------
APPLSYS.WF_CONTROL NONE N NO LONG
MYRAC1.MYDOM.COM NONE N NO LONG
SYS$BACKGROUND NONE N NO SHORT
SYS$USERS NONE N NO SHORT

SQL> select service_name, stat_name, value

2 from v$service_stats
3 where stat_name in
4 ('cluster wait time','db block changes'); -- among 28 different stats

SERVICE_NAME STAT_NAME VALUE

-------------------- ------------------------------ ----------------
SYS$USERS db block changes 34,618,836
MYRAC1.MYDOM.COM db block changes 471,456,178
APPLSYS.WF_CONTROL db block changes 0
SYS$BACKGROUND db block changes 2,383,337
SYS$USERS cluster wait time 52,810,072,281
MYRAC1.MYDOM.COM cluster wait time 798,080,328,466
APPLSYS.WF_CONTROL cluster wait time 0
SYS$BACKGROUND cluster wait time 4,405,806,181
Service related views
SQL> select service_name, event, total_waits,
2 total_timeouts, time_waited
3 from v$service_event
4 where event in (
5 'latch: library cache', 'read by other session'); -- from 114 events

SERVICE_NAME EVENT TOTAL_WAITS TIMEOUTS TIME_WAITED

---------------- --------------------- ----------- -------- -----------
MYRAC1.MYDOM.COM read by other session 25107023 23985 10848655
MYRAC1.MYDOM.COM latch: library cache 420975 0 4352231
SYS$BACKGROUND read by other session 29 0 14
SYS$BACKGROUND latch: library cache 6742 0 104454
SYS$USERS read by other session 27368974 435 3657701
SYS$USERS latch: library cache 144347 0 734451
• Ability to segment and measure workload
• Needs to be setup (becomes complex in RAC environments)
In-memory metrics
• V$METRIC/V$SYSMETRIC (V$SYSSTAT) ->
V$SYSMETRIC_HISTORY
• V$EVENTMETRIC (V$SYSTEM_EVENT) ->
V$EVENT_METRIC_HISTORY
• V$SYSMETRIC_SUMMARY (Avg/Min/StDev)
• WRH$_SYSMETRIC_SUMMARY
• V$METRICNAME / V$METRICGROUP
• Metrics -> Rate of change of statistics counters
• Alerts on rate -> Server Generated Alerts
(changeable via EM)
• Short-lived (3 mins of 15 secs, 1 hr of 60 secs)
In-memory metrics
SQL> select event, wait_time_milli, wait_count
2 from v$event_histogram
3* where event like 'SQL*Net %to client'
EVENT WAIT_TIME_MILLI WAIT_COUNT
------------------------------ --------------- ----------
SQL*Net message to client 1 137034612
SQL*Net message to client 2 19183
SQL*Net message to client 4 9922
SQL*Net message to client 8 5446
SQL*Net message to client 16 2931
SQL*Net message to client 32 1519
SQL*Net message to client 64 752
SQL*Net message to client 128 310
SQL*Net message to client 256 111
SQL*Net message to client 512 29
SQL*Net message to client 1024 1
SQL*Net message to client 2048 1
SQL*Net more data to client 1 8582781
SQL*Net more data to client 2 34556
SQL*Net more data to client 4 22114
SQL*Net more data to client 8 20108
Existing/New views
• Number of existing views expanded
– V$SQL, V$SESSION
• Information merged from various views
– V$SESSION (V$SESSION_WAIT and V$LOCK)
• New visibility
– V$SQL_BIND_CAPTURE
– V$SQL_SHARED_CURSOR
• Common Columns
– WAIT_CLASS and related
• WRH Tables exposed via DBA_HIST_% views
• V$PROCESS_MEMORY – session wise memory
• V$SQLSTATS – Low overhead access
V$SESSION view
• Wait information merged from V$SESSION_WAIT
– EVENT, P1 – P3, SEQ#, WAIT_CLASS
• Lock information from V$LOCK
– BLOCKING_SESSION, BLOCKING_SESSION_STATUS,
BLOCKING_INSTANCE
• Tons of new columns
– SQL_ID (Current and Previous)
– SQL_CHILD_NUMBER (Current and Previous)
– PLSQL related information (Entry, Subprogram, Object, etc)
– SQL_TRACE columns (binds and waits as well)
– CLIENT_IDENTIFIER (DBMS_MONITOR)
• No need to join to V$SESSION_WAIT, V$LOCK now
V$SQL view
• Complete text in SQL_FULLTEXT
• Parallel execution (PX_SERVERS_EXECUTION)
• Different types of wait times
– APPLICATION_WAIT_TIME
– CONCURRENCY_WAIT_TIME
– CLUSTER_WAIT_TIME
– USER_IO_WAIT_TIME
– PLSQL_EXEC_TIME
– JAVA_EXEC_TIME
• Additional details:
– PARSING_SCHEMA_NAME
– Optimizer details
– BIND_DATA
– LAST_LOAD_TIME (Docs: LAST_ACTIVE_TIME??)
V$SQL related views
• V$SQL_OPTIMIZER_VIEW
– Optimizer parameters for SQL_ID and particular child cursor
• V$SQL_PLAN
– TIMESTAMP, OTHER_XML, QBLOCK_NAME, etc.
• DBMS_XPLAN exposes details*
• V$SQL_SHARED_CURSOR
– Why was a cursor not shared?
– USER_BIND_PEEK_MISMATCH column
– Groundwork for 11g bind-awareness/bind-sensitiveness
• V$SQL_BIND_CAPTURE
– Use with ASH to get some bind values

* Undocumented : “ADVANCED” parameter to get OTHER_XML details

Event Breakout
• Event names for “blocking” type events changed
• “latch free” now has 29 types (10.2.0.3)
– latch: cache buffers chains
– latch: library cache
– latch: library cache pin
• “enqueue” has 205 types (10.2.0.3)
– enq: TX - row lock contention (Waiting for Row lock)
– enq: UL – contention (Waiting for User generated lock)
– enq: XR - quiesce database (Waiting for DB Quiesce)
– enq: ST – contention (Space transaction – rare now with LMT)
• “read by other session” (BBW with reason code 130)
• Break out at all levels (session/system- compare R1)
• Major change in monitoring / reporting code!!!
Mutexes – Boon or Bane?
• Low overhead mechanism replacing some latches
• “Simulated” on certain platforms
– HP-UX uses a CAS (“compare-and-swap”) simulator
– This itself is a latch: quite an overhead
• Track using V$MUTEX_SLEEP and
V$MUTEX_SLEEP_HISTORY
• “_kks_use_mutex_pin = FALSE” turns this off
• Snapshotted in STATSPACK, not in AWR 
Operating System Statistics
• Set SQL*Plus format to get right value! (issue in AWR)

SQL> select stat_name, value

2 from v$osstat;
STAT_NAME VALUE
---------------------------------------------------------------- ----------
NUM_CPUS 24
IDLE_TIME 802033123
BUSY_TIME 878091185
USER_TIME 489231709
SYS_TIME 388859476
IOWAIT_TIME 187539358
AVG_IDLE_TIME 33398667
AVG_BUSY_TIME 36564711
AVG_USER_TIME 20362160
AVG_SYS_TIME 16180042
AVG_IOWAIT_TIME 7794964
OS_CPU_WAIT_TIME 2.6867E+13
RSRC_MGR_CPU_WAIT_TIME 0
LOAD 2.359375
NUM_CPU_SOCKETS 24
PHYSICAL_MEMORY_BYTES 1.0287E+11
VM_IN_BYTES 7868502024
VM_OUT_BYTES 0
Time Model
• CPU breakup; updated every 3 seconds!!!
SQL> select stat_name, value from v$sys_time_model
where value > 0;
STAT_NAME VALUE
--------------------------------------------- --------------
DB time 893170091346
DB CPU 176244910473
sequence load elapsed time 10215471781
parse time elapsed 4524012412
hard parse elapsed time 3657262901
failed parse elapsed time 103540062
hard parse (sharing criteria) elapsed time 365217641
hard parse (bind mismatch) elapsed time 5923514
repeated bind elapsed time 14768010
connection management call elapsed time 328536127
PL/SQL execution elapsed time 5554924592
PL/SQL compilation elapsed time 333815896
background elapsed time 13782131027
background cpu time 4572399582
Wait Model
SQL> select wait_class, time_waited, total_waits
from v$system_wait_class;
WAIT_CLASS TIME_WAITED TOTAL_WAITS
-------------------- ---------------- ----------------
Other 1370371162 10777690366
Application 11031616 77468
Configuration 433075 3972043
Administrative 986 10
Concurrency 5069513044 4023651698
Commit 4187059 4606985
Idle 26816217176 1179884266
Network 1964510 152943745
User I/O 845512795 3197266600
System I/O 21813081 47778673
Wait Class
• Types of wait: classified by WAIT_CLASS
SQL> select wait_class, count(*) from v$event_name
2 group by wait_class;
WAIT_CLASS COUNT(*)
----------------- -----
User I/O 17
Application 12
Network 27
Concurrency 25
Administrative 46
Configuration 23
Scheduler 2
Cluster 47
Other 592
Idle 62
System I/O 24
Hidden surprises
• GROUP BY does not ensure ORDER BY
– New code base avoids sort for GROUP BY operation
– Order was never guaranteed
– Large amount of legacy/custom code needs change
– Investigate “_gby_hash_aggregation_enabled” parameter
• System Statistics
– “CPU Cost” optional in 9i
– Fully functional (by default) in 10g
– Sample collected when DB installed/upgraded
– Check in SYS.AUX_STATS$ view
– See Metalink Notes: 457228.1, 153761.1, 149560.1
– Collecting System Stats does NOT invalidate shared pool
Hidden surprises
• Dynamic sampling
– 9i default was 1
– 10g default is 2
– all unanalyzed tables are sampled
– Number of blocks sampled: twice default number (2x32=64)
– Impact on Global Temp tables
– Impact on parse (not affected by plan permutation limits)
• Statistics gathering and Table monitoring
– Monitoring ON by default
– Used by Automatic Stats gathering (Dictionary stats included)
– Trigger fixed at 10%, changeable in 11g
– Stats collection for Fixed table (new but optional)
Hidden surprises
• DBMS_STATS
– 9i and below: Defaults were fixed
– 10g and above: Defaults can be changed (boon or bane?)
– Many parameters default to “decided by Oracle”
– Examples:
• NO_INVALIDATE: DBMS_STATS.AUTO_INVALIDATE
• CASCADE: DBMS_STATS.AUTO_CASCADE
• ESTIMATE_PERCENT: DBMS_STATS.AUTO_SAMPLE_SIZE
• METHOD_OPT: FOR ALL COLUMNS SIZE AUTO
• GRANULARITY: AUTO (For partitioned objects only)
– Automatically stores past 31 days DBA_TAB_STATS_HISTORY
– DIFF_TABLE_STATS_* function in 10.2.0.4
Useful sections: AWR Report
• Segment statistics in AWR report
– Same as “Top 5 Segments” in Level 7 STATSPACK reports (9i+)
– Segments by Global Cache Buffer Busy (RAC only, new section in 10g)
– Segments by CR Blocks Received (RAC only)
– Segments by Current Blocks Received (RAC only)
• New GC Load profile and GC Efficiency percentages
Global Cache Load Profile
~~~~~~~~~~~~~~~~~~~~~~~~~ Per Second Per Transaction
--------------- ---------------
Global Cache blocks received: 1,375.70 2.90
Global Cache blocks served: 164.77 0.35
GCS/GES messages received: 5,165.31 10.88
GCS/GES messages sent: 8,684.54 18.30
DBWR Fusion writes: 0.63 0.00
Estd Interconnect traffic (KB) 15,028.80
Global Cache Efficiency Percentages (Target local+remote 100%)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Buffer access - local cache %: 97.41
Buffer access - remote cache %: 0.43
Buffer access - disk %: 2.16
Extracting data from the AWR
• Complete source in paper
• Adaptable for different types of extracts
/* This cursor fetches details of the current snapshot plus the
next one using the LEAD function. We will use this to
make sure that there was no DB restart in-between */
cursor snapshot is
select snap_id, lead(snap_id, 1, 0) OVER (ORDER BY snap_id),
startup_time, lead(startup_time, 1) OVER (ORDER BY snap_id),
to_char(begin_interval_time) begin_interval_time,
to_char(end_interval_time) end_interval_time
from sys.dba_hist_snapshot
where instance_number = v_instance_number -- For RAC instances
and dbid = (select dbid from v$database);
Extracting data from the AWR
-- We don't subtract for certain types such as NUM_CPUS, etc.
cursor osstat is
select e.instance_number, e.stat_name,
case
when e.stat_name in
('NUM_CPU_SOCKETS','NUM_CPUS','NUM_CPU_CORES','LOAD')
then e.value
else e.value - b.value
end stat_value
from sys.dba_hist_osstat b, sys.dba_hist_osstat e
where b.stat_id = e.stat_id
and b.snap_id = v_begin_snap_id and e.snap_id = v_end_snap_id
and b.instance_number = e.instance_number
and b.instance_number = v_instance_number;
Extracting data from the AWR
v_instance_number := &1; -- Send this from command line
open snapshot;
LOOP
fetch snapshot into v_begin_snap_id, v_end_snap_id,
v_begin_startup_time, v_end_startup_time,
v_begin_interval_time, v_end_interval_time;
exit when snapshot%NOTFOUND;
-- Run through only if the startup times for both snaps are same!
-- also, avoid the last line (lead will return 0 for end_id)
if ( v_begin_startup_time = v_end_startup_time ) and ( v_end_snap_id != 0 ) then
open osstat;
loop
fetch osstat into v_instance_number, v_stat_name, v_value;
exit when osstat%NOTFOUND;
dbms_output.put_line(v_end_snap_id || ','
|| to_char(v_end_interval_time, 'DD-MON-YY HH24:MI')
|| ',' || v_stat_name || ',' || v_value);
end loop;
close osstat;
end if;
END LOOP;
close snapshot;
end;
Extracting data from the AWR
Sample output: (portion of output)
SQL> @awr_osstat 1
old 38: v_instance_number := &1; -- Send this from command line
new 38: v_instance_number := 1; -- Send this from command line
5363,18-FEB-08 00:00,NUM_CPUS,24
5363,18-FEB-08 00:00,IDLE_TIME,5975106
5363,18-FEB-08 00:00,BUSY_TIME,2664727
5363,18-FEB-08 00:00,USER_TIME,970933
5363,18-FEB-08 00:00,SYS_TIME,1693794
5363,18-FEB-08 00:00,IOWAIT_TIME,531265
5363,18-FEB-08 00:00,LOAD,.3505859375
5363,18-FEB-08 00:00,NUM_CPU_SOCKETS,24
5363,18-FEB-08 00:00,PHYSICAL_MEMORY_BYTES,0
5363,18-FEB-08 00:00,VM_IN_BYTES,23502864
5363,18-FEB-08 00:00,VM_OUT_BYTES,0
--
5364,18-FEB-08 01:00,NUM_CPUS,24
5364,18-FEB-08 01:00,IDLE_TIME,6018339
5364,18-FEB-08 01:00,BUSY_TIME,2684043
5364,18-FEB-08 01:00,USER_TIME,967711
5364,18-FEB-08 01:00,SYS_TIME,1716332
5364,18-FEB-08 01:00,IOWAIT_TIME,600067
<snip>
New features in DBMS_XPLAN
• Excellent alternative for EXPLAIN_PLAN
• Introduced in 9i, expanded in 10g
– DISPLAY_AWR - Plan stored in the AWR
– DISPLAY_CURSOR – V$SQL cursor cache
– DISPLAY_SQLSET - Plan of a given statement
stored in a SQL tuning set (STS)
– Simple to use:
• “select * from table(dbms_xplan.display_awr(‘<SQL_ID>’,
options));”
– Displays all child cursors if present
New features in DBMS_XPLAN
SQL> select * from table(dbms_xplan.display_cursor('g6jzbgnku8024',NULL,'ADVANCED'));
PLAN_TABLE_OUTPUT
-------------------------------------------------------------------------------------------
SQL_ID g6jzbgnku8024, child number 0
-------------------------------------
SELECT R.RESPONSIBILITY_NAME FROM FND_RESPONSIBILITY_VL R
WHERE R.RESPONSIBILITY_ID = :1 AND R.APPLICATION_ID = :1
Plan hash value: 3221072286
---------------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
---------------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | | | 2 (100)| |
| 1 | NESTED LOOPS | | 1 | 51 | 2 (0)| 00:00:01 |
|* 2 | INDEX UNIQUE SCAN | FND_RESPONSIBILITY_U1 | 1 | 10 | 1 (0)| 00:00:01 |
| 3 | TABLE ACCESS BY INDEX ROWID| FND_RESPONSIBILITY_TL | 1 | 41 | 1 (0)| 00:00:01 |
|* 4 | INDEX UNIQUE SCAN | FND_RESPONSIBILITY_TL_U1 | 1 | | 0 (0)| |
---------------------------------------------------------------------------------------------------------
Query Block Name / Object Alias (identified by operation id):
-------------------------------------------------------------
1 - SEL$F5BB74E1
2 - SEL$F5BB74E1 / B@SEL$2
3 - SEL$F5BB74E1 / T@SEL$2
4 - SEL$F5BB74E1 / T@SEL$2
Outline Data
-------------

/*+
BEGIN_OUTLINE_DATA
IGNORE_OPTIM_EMBEDDED_HINTS
OPTIMIZER_FEATURES_ENABLE('10.2.0.3')
OPT_PARAM('_b_tree_bitmap_plans' 'false')
OPT_PARAM('_fast_full_scan_enabled' 'false')
*/
New features in DBMS_XPLAN
Peeked Binds (identified by position):
--------------------------------------

1 - :1 (NUMBER): 50357
2 - :1 (NUMBER, Primary=1)

Predicate Information (identified by operation id):

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

2 - access("B"."APPLICATION_ID"=:1 AND "B"."RESPONSIBILITY_ID"=:1)

4 - access("T"."APPLICATION_ID"=:1 AND "T"."RESPONSIBILITY_ID"=:1 AND
"T"."LANGUAGE"=USERENV('LANG'))

Column Projection Information (identified by operation id):

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

1 - "T"."RESPONSIBILITY_NAME"[VARCHAR2,100]
3 - "T"."RESPONSIBILITY_NAME"[VARCHAR2,100]
4 - "T".ROWID[ROWID,10]

SQL_ID g6jzbgnku8024, child number 1

-------------------------------------
SELECT R.RESPONSIBILITY_NAME FROM FND_RESPONSIBILITY_VL R WHERE R.RESPONSIBILITY_ID = :1 AND
R.APPLICATION_ID = :1

<snipped out further information about child cursor 1 and 2>

Items Learned in this Session

• So there you have it all!

• This was a summary of a few of the useful features
• Hopefully, this adds to the vast information out there
• Where do you go from here?
– Papers, OTN, Blogs, and….. the Oracle manuals! 
• What do you plan to use when you get back to the
office?
Questions and feedback

Questions?
Contact information:
[email protected]
Please fill up your evaluation form!
Session # 319
Oracle DB 10g: An enlightened revisit
Thank you!