question of AWR report

Similar Questions

• questions about the awr report & sqlt

  Hi guys,.

  DB version: 10.2.0.5

  Need your advice.

  For my AWR report, I put the 30-day retention period. From my understanding, the information in awr report are taken from several views of dba_hist *.

  Points of view takes only those albums n intenstive SQL Statistics on resources?

  I ask this because I have a doubt. Let's say that on 18/03/2014, I detected a long SQL with sql id xxx.

  Using SQLT (such method), I can only see the historical performance of Plan (delta) for this particular SQLs 16/03, 17/03, 18/03.

  In this case, we can say that the SQLs shot only 3 days for the last 30 days? Or it can also mean that the particular SQL was not classfified at the top of the page resource n intensive 18/02 to 15/03 SQLs (probably use a better plan - for example: plan a different value)?

  Thank you

  That is right.

  Concerning

  Jonathan Lews

• begin and end in awr report

  Hi all

  Hope all are well. I need your help, I doubt reading the awr report.

  Summary of the report

  Cache sizes

  	Start	End
  Cache buffers:	572 M	572 M	Std block size:	8K
  Size of the shared pool:	408 M	408 M	Log buffer:	13 604 K

  My question is what is 'START' and 'END' in the figure above and why to cache buffer and shared pool sizes are the same for BEGIN and END.your help appreciated, thanks in advance.

  Best regards.

  size of the buffer at the start of the snapshot cache

  size of the buffer at the end of the snapshot cache

  Shared at the beginning of the snapshot pool

  Shared pool, at the end of the snapshot

  So that means: either you did not Oracle or Oracle to change these settings dynamically does not need to change.

  Always send four-digit version of your database and platform info. This forum is not keeping track of it.

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

  Sybrand Bakker

  Senior Oracle DBA

• Bug? : cannot execute predefined "last awr" report with the latest version of the SQL Developer

  Using sql developer before 4.x... report works fine

  Using SQL developer 4.0.0.13... report fails with ORA-01422

  using the actual query to generate awr against the same range of component snap works perfectly

  I opened a SR on the question

  Thank you

  It's always a good idea to open a SR when you are able.  I assume that you have already received a response from support.  For others I wondered about this, the question is likely related to running the report against a CCR environment, as indicated in the following bug:

  Bug 18132564 - LAST AWR REPORT FAILS WITH ORA-01422 IN CCR ENVIRONMENTS
  

  The call of the AWR report requires parameters for a DB and an ID of Instance, so that the report should run, as in earlier versions, for only the current instance, or be modified to run all instances in the cluster.

  Kind regards

  Gary

  SQL development team

  P.S.: And, in fact, a discussion prior forum - ai2 4.0-AWR and ASH giving ora-01427

  lead following bugs are connected and supposed to be fixed for SQL Developer 4.0...

  
  

  Bug 17481944 - PERFORMANCE PANELS NEED AWR/ASH/ADDM FOR RAC ENVIRONMENTS SUPPORT

  
  

  ... with the developer commenting

  I've added a number instance selection control to the various reports: ASH,.
  AWR, difference and SQL. The control is supposed to be visible when the
  connection is a RAC system, tests should a RAC.

  So if you are not on a RAC system, or are on CARS but do not see the selection number instance control, please let us know.

  Post edited by: Gary Graham

• How to take awr report

  A little confused in the generation of awr report using awrrpt.sql

  Here are the pictures that we have in the repository.

  	8257 November 13, 2013 02:00	1
  	8258 November 13, 2013 02:30	1
  	8259 November 13, 2013 03:00	1
  	8260 November 13, 2013 03:30	1
  	8261 November 13, 2013 04:00	1
  	8262 November 13, 2013 04:30	1

  And the test run we had was between 02:15 at 03:45.  What shots should I choose as the ID of the snap-snap-in begin end snap ID.

  I got this question when I saw how the shots were taken as snap ID 8258

  SNAP_ID BEGIN_INTERVAL_TIME

  END_INTERVAL_TIME

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

  8258 13 NOVEMBER 13 02.00.19.314 AM

  13 NOVEMBER 13 02.30.23.284 AM

  where I was going to select 8257 as the ID of the snap-snap-in begin.

  Which snap IDs makes more sense for the time window of 02:15 at 03:45.

  Thank you

  Siva.

  Yes, after 02:15 activity is recorded in wink 8258, because at the time the 8257 snap was taken (02:00) the workload has not watched yet. In order something to be registered, it must extend the instant collection interval. Instant is the current state of the instance, and if you set the snapshot collection at intervals of more than an hour, you might be missing information. Also there is no need to set the very short interval (provable 15 min.). An hour is an average value, and that's by default. On the first response I wrote snap 8527 because under 'time window' 8527 is correct snap_id, but in the context of statistical collection snap_id 8528 is correct.

  Post edited by: IvicaArsov

• What is the measure of the value "Waiting" in AWR report?

  Event	Expected	% Time - out	Wait total (s) times	AVG wait (ms)	Waiting for /txn
  SQL * Net more data from dblink	406 747		466	1	184.13
  reading of scattered files DB	28 659		330	12	12.97
  SQL * Net message from dblink the	30 115		254	8	13.63
  log file parallel write	21 021		177	8	9.52

  Hi all

  As the example below, I question what is the measure of the value "Waiting" in the section "event wait" AWR report?

  Because "Expected" the value is high, but "Queue time Total (s)" is low. example "SQL * Net message from dblink the" a 30 115 "expected' but 'Queue time Total (s)' lower 'DB file scattered read' w 28 659 'wait '.

  Could you please explain more about the value "waiting"?

  Thank you

  Hiko

  taohiko wrote:

  Event	Waits	%Time -outs	Wait total (s) times	Avg wait (ms)	Waits /txn
  SQL * Net more data from dblink	406 747		466	1	184.13
  reading of scattered files DB	28 659		330	12	12.97
  SQL * Net message from dblink the	30 115		254	8	13.63
  log file parallel write	21 021		177	8	9.52

  Hi all

  As the example below, I question what is the measure of the value "Waiting" in the section "event wait" AWR report?

  It is simply a... count the number of times where the system had to "WAIT" on the specified event.

  Because "Expected" the value is high, but "Queue time Total (s)" is low. example "SQL * Net message from dblink the" a 30 115 "expected' but 'Queue time Total (s)' lower 'DB file scattered read' w 28 659 'wait '.

  Watch "AVG time-out.    In one case, there are fewer occurrences of the event, but the average wait by event time was higher.

  If I buy 3 burgers and pay $1000 each

  or I buy 500 burgers and pay $1.00 each.

  Why 3 hamburgers cost more than 500 burgers?

  Could you please explain more about the value "waiting"?

  "Waiting" is the number of times that the system had to wait for the named event, during the sampling period.

  "Total wait time" is the total time (measured in seconds) spent waiting on all occurrences of the specified event, during the sampling period.

  "Avg" wait"is the time (expressed in milliseconds) means spent on each occurrence of the specified event, during the sampling period.

  It seems pretty obvious.  You count how many times it happens something (wait), you measure it, time spent doing this something fracture (total wait time) and tells you the total time spent (total wait time) by the number of times it happened (waits) to find out the average time spent on each occurrence of that something.

  Thank you

  Hiko

• AWR report cannot locate the bad SQLs

  DB version: 11.2.0.2
  Platform: Solaris 10
  
  In our data base RAC 2 nodes, we have patterns of 20 applications. When an application is faced with performance problems, we try to generate AWR for the period in question. But the stats (like bad SQLs) for this scheme to be buried because there is another application SQLs performing even worse.
  
  I know that we cannot create AWR report for specific patterns. But I wonder how his stats are collected in a multi-schema RAC environments that will differentiate between applications.

  But the stats (like bad SQLs) for this scheme to be buried because there is another application SQLs performing even worse.

  Since all applications run in a database and query of the other application run bad then they will be the first candidate of tunning since these are taking more resources and impact on other application requests.
  You can also generate report of ASH for module specific or customer with assistance from under the option set in report of ash.

  set target_module_name = ";
  set target_action_name = ";
  set target_client_id = ";

• ASH versus reports AWR report

  My DB is oracle10g.
  
  In enterprise manager, we can generate the report of ASH. Also, we can generate AWR report in the server through files awrrpt.sql.
  
  ASH report provides the information below.
  
  Main events
  Load profile
  Top of page SQL
  Albums of PL/SQL
  Top of page sessions
  Top of page objects/files/locks
  Activity over time
  
  Provide information below AWR reports.
  
  Summary of the report
  Wait events statistics
  SQL Statistics
  Statistics from the instance activity
  IO statistics
  Buffer Pool statistics
  Advisory statistics
  Wait statistics
  Cancel statistics
  Latch statistics
  Segment statistics
  Cache dictionary statistics
  Library Cache statistics
  Memory statistics
  Flow statistics
  Statistics on resources limit
  init.ora parameters
  
  
  My question is, under what circumstances, report of ASH can be seen and what circumstances AWR report is good to see? Any help is appreciated. Thank you

  in what circumstances, report of ASH can be seen and what circumstances AWR report is good to see? Any help is appreciated. Thank you

  Ash can help you when there is felt a sudden degradation in performance of the database.
  AWR - past historical snapshot intervals.

  AWR, stores the performance statistics of session for analysis later.
  ASH - storage is not persistent and with time, the older entries are removed to accommodate the new. They can be viewed using V$ ACTIVE_SESSION_HISTORY

  HTH
  -André

• Question about AWR

  Hello
  
  I have a few question about AWR:
  
  (1) which means the metric % CPU Non-Parse?
  (2) the wait time event CPU means: the total amount of time CPU used during snapshot, is that correct?
  (3) indicating the Parse CPU to analyze Elapsd % ratio of 200.00?
  
  Will be grateful for the help.
  Thank you
  Orausern
  
  Published by: orausern on May 18, 2009 06:16

  orausern wrote:
  Can you suggest what the argument of the UC Parse to parse Elapsd % ratio of 200,00 specifies? In one of my AWR report, I get this value of 200%. -Given that the target is 100%, is 200% of the abnormal value, I have to do something?

  If you can not make sense a report, check the underlying data. In this case, the Instance activity statistics: "parse time cpu ' and 'analysis elapsed time '. I think you'll find that time cpu is greater than the time - suggesting that the see you a measurement error, especially if times are both very small anyway. Having done this, you can compare the "analysis time cpu" with time CPU anyway to see if there is any point in worrying about the time cpu spent in analysis.

  Concerning
  Jonathan Lewis
  http://jonathanlewis.WordPress.com
  http://www.jlcomp.demon.co.UK

  "Science is more than a body of knowledge; It's a way of thinking. "
  Carl Sagan

• A way to find my previous questions and problem reports?

  Is it possible on this forum to find my previous questions and problem reports? Since viewing a question earlier today, I found the greater PART of the answer and would like to add that I found the question, without waiting for someone else to answer. I can't remember the title I gave to the previous question, or I would use this title in the search for it.

  Information troubleshooting below is most relevant to a previous problem that this question.

  Contributions my link on main page of the forum should display a list of all your job offers.

  See the search in this forum for your username under "posed by".

• This output on AWR report mean?

  Hello.

  I'm having problems of performance with an Oracle Database EE 12.1.0.2.0 in a particular query.

  I ran an AWR report and in the ADDM section get me the following for the troubled query:

  Raison d'etre

  High level calls to execute the SELECT statement with SQL_ID

  '6a9fmtbam27fp' are responsible for 99% of the time spent on database

  the SELECT statement with SQL_ID '8uqsfw32bytvh '.

  
  

  The query that I'm having this problem is the SQL_ID '8uqsfw32bytvh '.

  I couldn't find a relationship between the SQL_ID '8uqsfw32bytvh' and '6a9fmtbam27fp '.

  
  

  Do you know guys what the ADDM meant by "call for higher level? I couldn't find information online.

  
  

  Thank you.

  Jonas
  

  This comes from data ASH - V$ ACTIVE_SESSION_HISTORY / DBA_HIST_ACTIVE_SESS_HISTORY.

  For a SQL statement running in the titles the sql entry highlights so sql called from procs, sql calling functions that execute sql, etc.

  -HIGH LEVEL SQL ID (and other useful columns as PLSQL_ENTRY_OBJECT_ID, PLSQL_ENTRY_SUBPROGRAM_ID, PLSQL_OBJECT_ID, PLSQL_SUBPROGRAM_ID)

• Number of processors and cores in AWR report

  Hello

  I have a report AWR to my system and a couple of entries confuse me.

  Q1. It's showing 3 carrots, 2 Sockets and 4 processors. The CPU and the Sockets, I understand that 2 chips with 2 processors per chip so 4 CPU.

  What confuses me, is the fact that it shows 3 carrots. Is that that by CPU cores? Othererwise 3-4 makes no sense?

  Q2. What is the difference between DB, time CPU and CPU DB?

  My understanding is that DB time is the time used by the process, including the time processor and wait time. Because a system can have more than one processor, it is quite natural that DB time may exceed to elapse time - since you have multiple processors that contribute to this time (I guess that DB time cannot be really more time elapse * our processors?)

  In fact, you could say if DB time does not exceed the time to elapse, then you have a Nock idle system?

  My understanding of time CPU, is that it's time where each process actually gets treatment CPU (so it does not time-out). I think that the AWR report confusingly used the term time CPU and CPU DB, to mean the same thing?

  clearly greatly appreciated

  Jim

  
  

  Hi Jimbo,

  Q1. "Nuclei" in the CWA report represents the quantity of physical processor cores and the "processors" takes into account the simultaneous multithreading so in your case, one option could be a CPU Socket has a dual core (not type hyperthreading) and the other CPU Socket has one heart (hyperthreaded). If this is the case, you will have 3 physical 4 CPU cores, (because it takes into account the hyper-threaded architecture) and 2 taken. It's just an option, you will need to check it out.

  Q2. No-Idle wait events are events that are not classified by Oracle as a "Idle" of events to wait wait:

  • Example of Idle wait-event: SQL * Net client message (server process is waiting for the client process to do something)
  • Idle wait-event example: db file scattered read (event waiting user IO representing a multilock read as full scan of the table or full scan small index.
  • You can get a complete list of active waiting events by running:

  Select name from v$ event_name where wait_class <> "inactive".

  • You can get a full list of the events of inactive waiting by running:

  Select name from v$ event_name where wait_class = 'Idle '.

  
  

  DB time is therefore the sum of all events of active waiting and also the CPU time that is a statistic:

  • You can see the value of the amount of time CPU in 10s of miiliseconds (centisecondes) for an exeuting session:

  SELECT THE VALUE

  FROM v$ sesstat JOIN v$ statname USING (statistics #)

  WHERE name = 'CPU used by this session' AND SID =

  • You can see the value of the time total CPU in 10s of milliseconds (centisecondes) for the instance of exeuting:

  SELECT the value from v$ sysstat

  WHERE name = 'CPU used by this session.

• awrrpti script - awr report fails

  Friends...

  OS: Linux

  DB: 11 GR 2

  I'm trying to solve this riddle, supposed to be a simple fix but not able to finish it...

  We run awrrpti script to generate AWR report for the whole day. What I do is ask DBA_HIST_SNAPSHOT and get the snapshot min (first) and max (last) of the day. Then I fed that range from snap_id to awrrpti script to generate AWR report all day to single instance.

  It works perfectly well, but if instance gets restarted during the day then my approach above does not work and AWR report generation fails.

  Current approach:

  1. the DBA_HIST_SNAPSHOT query

  2. download MIN (snap_id), MAX (snap_id) for the day

  3. power #2 to awrrpti via bind variable

  4 generate the unique HTML report for the instance

  Query used to collect the id of component software snap-in min/max:

  Select min (snap_id), max (snap_id)

  Of

  (select snap_id, begin_interval_time, end_interval_time

  of dba_hist_snapshot

  where begin_interval_time > = trunc(sysdate-1)

  and being_intreval_time < = trunc(sysdate-1+23/24)

  );

  Tips to avoid failures of AWR report generation if instance gets restarted?

  How to go to the entry below in simple sql? It is possible to start «WITH max_startup...» "in the sql script?

  Yes, there is no problem. Is it just another way to write the query

  But you can use this one instead:

  SELECT MIN(dhs.snap_id) min_snap_id, MAX(dhs.snap_id) max_snap_id
  FROM DBA_HIST_SNAPSHOT dhs,
  (SELECT MAX(startup_time) max_startup_time FROM DBA_HIST_SNAPSHOT) mst
  where dhs.startup_time = mst.max_startup_time
  and begin_interval_time >= trunc(SYSDATE-1)
  and begin_interval_time <= trunc(sysdate-1+23/24)
  ;
  

  again will try, test and update this blog

  Forum! Please, I beg you.

  Concerning

  Juan M

• How to activate the keys to generate and view the AWR reports Developer SQL

  Version: 4.0.3.16

  Build: HAND - 16.84

  S/n > (sid) > Performance > AWR report viewer

  I am able to select the start and end ID, but the following buttons are disabled:

  • Entries Toggel
  • Generate the report
  • Save to file
  • Open in browser

  What needs to be done to enable these buttons?

  Hello

  After entering start Id and the Id of the end, the button "report" will be activated, and then after that the report generated (by clicking the report generate) rest buttons button is activated. As shown in the image below.

  

  Hope it will be useful.

  Kind regards

  Sanjaya

• The main events in AWR report

  Hi all

  11.2.0.1

  AIX 6.1

  I had this big event awr report.

  Can you help me how to interpret this please. Or what is the value of looking here?  Thank you very much

  The main events in AWR report

  I have

  n

  Snap              s       Snap                                                                                                    A

  Snap the t hard event time start Avgwt DB A

  Time event ID # (m) expected grade (s) (ms) % S waiting class

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

  13902 13/09/16 09:00 1 60.05 os thread start 1 140,00 11.08 79,11 56 0,0 Concurrency

  13902 13/09/16 09:00 60.05 1 CPU time 2 0.00 0.00 9,45 48 0.0 CPU

  13902 13/09/16 09:00 1 60.05 log file parallel write 3 1506.00 1.84 1.22 9 0.0 IO system

  13902 13/09/16 09:00 1 60.05 control file parallel write 4 1291.00 1.41 1.09 7 0.0 IO system

  13902 13/09/16 09:00 1 log file sync 60.05 5 721.00 1.03 1.43 commit 5 0.0

  13903 13/09/16 10:00 1 60.05 os thread start 1 138.00 10.36 75,10 47 0.0 Concurrency

  13903 13/09/16 CPU 60.05 1 10:00 hour 2 0.00 0.00 10,28 47 0.0 CPU

  13903 13/09/16 10:00 1 60.05 log file parallel write 3 1575.00 1.73 1.10 8 0.0 IO system

  13903 13/09/16 10:00 1 60.05 direct path read 3 486.00 1.73 3.56 8 0.0 IO user

  13903 13/09/16 10:00 1 60.05 control file parallel write 4 1285.00 1.51 1.17 7 0.0 IO system

  13903 13/09/16 10:00 1 log file sync 60.05 5 823.00 1.08 1.32 commit 5 0.0

  13904 13/09/16 11:00 60.05 1 CPU time 1 0.00 17,06 0.00 CPU 32 0.0

  13904 13/09/16 11:00 1 60.05 os thread start 2 145.00 10,65 73.46 20 0.0 Concurrency

  Thanks, I just tell the users, there is no problem in the comic book, and the stats is normal and no suspiscious activity and that the problem is in the minds of users only.