Performance Tuning :: Analyzing Non-sequential Snapshots In AWR
Mar 5, 2013
How can check database performance in a non-sequential parts of a day using AWR tool. For example, suppose we divide a day to 2 parts: low-traffic time and high-traffic time with the following time interval:
low-traffic time : 7:00am - 10:00am and 7:00pm - 11:00pm
high-traffic time : 10:00am - 01:00pm and 4:00pm - 06:00pm
I want to examine performance using snapshots gathered by AWR. If I get 2 AWR reports for low-traffic time of day (one for 7:00am - 10:00am, another for 7:00pm - 11:00pm ) and compute average of values in 2 reports, could I called it database performance in low-traffic part of day or not?
We are experiencing Network waits on one of our 2-node clustered databases...In every 1 hour of clock time we are finding 700-900 seconds of Network waits
From the AWR data I find that "ARCH wait on SENDREQ" is one of the main constituent for these Network waits and as such I suspect Network between Production database and its corresponding database might be slow
Question 1) Does this understanding look correct?
Question 2) Apart from the above what could be the other causes of the Network waits. Can we point out any particular area from following AWR extract...Seeing some gc* waits initially I thought it might be due to slow interconnect between the cluster nodes but some google search denotes it is not the case...So what could be other causes? I mean which network link I would check?
Snap Id Snap Time Sessions Curs/Sess --------- ------------------- -------- --------- Begin Snap: 22631 22-May-13 10:00:11 976 7.9 End Snap: 22632 22-May-13 11:00:28 978 8.1 Elapsed: 60.29 (mins) DB Time: 795.66 (mins) [code].....
In a 3-node RAC setup; one node is showing high CPU utilization around 40~50%. The CPU utilization was less than 20% 10 days back but from 9th oldest day it jumped and consistently shows the double figure. I ran AWR reports on all three nodes and found one node with high CPU utilization and shows below tops events-
EVENT WAITS TIME(S) AVG WAIT(MS) %TOTAL CALL TIME WAIT CLASS CPU time 5,802 34.9
RFS ping 15 5,118 33,671 30.8 Other
Log file sequential read 234,831 5,036 21 30.3 System I/O
Sql*Net more data from client 24,1711,08745 6.5 Network
Db file sequential read130,939 4533 2.7 User I/O
Findings:- On AWR report(file attached) for node= sipd207; we can see that "RFS PING" wait event takes 30% of the waits and "log file sequential read" wait event takes 30% of the waits that occurs in database.
1)Are these symptoms of undersized log buffer? 2)I feel Network wait can be reduced by tweaking SDU & TDU values based on MDU.
Following query is hanging either with 'Sequential access read' or 'Latch Free' wait event Important thing is the table which is self joined in subquery here does not have any index at all While it was hanged I tried to get trace of it and terminated twice. As such haven't got 'row source generataion' The table has only 120000 records and it shall update 34000 records
UPDATE invoice_header inv SET inv.modified_due_date = (SELECT inv1.btn_due_date FROM invoice_header inv1 WHERE inv.dct_code = inv1.dct_code AND inv1.release = 'A5')
[code]...
During 'sequential read' using p1,p2 values tried to get what the session is reading and found that it is using the table itself.
During lath free I found following SELECT name, 'Child '||child#, gets, misses, sleeps FROM v$latch_children WHERE addr= (select p1raw from v$session_wait where sid=18) UNION
[code]...
However instead of self join when I creaed global temporary table as
create global temporary table t as select * from invoice_header where release='A5'
And used it in the update as
UPDATE invoice_header inv SET inv.modified_due_date = (SELECT t.btn_due_date FROM t WHERE inv.dct_code = t.dct_code AND t.release = 'A5') WHERE inv.release = 'A5' AND inv.btn_due_date >= TRUNC (SYSDATE)
It updated the records in a second!!
Questions are 1) why it is producing 'sequential read' wait event when there is no index access or else why it is doing single block access when FTS is required? 2) Why is the 'latch free' wait event here and what it indicates here with 'cache buffer handles'? Is it because we are reading and updating the same segment?
know in case DDL of table is required. It has all nullable columns and no index at all. Since it is 9i I am unable to use MERGE effectively in this case
Looking to understand the difference between instance tuning and database tuning.
What is the difference between these two tuning exercises? I understand that an instance is memory based structures (logical) where as database consists of physical structures.
However, how does one tune a database the physical structure? Does it have to do with file placements/block sizes etc. Would you agree that a lot of that is taken care by ASM now in 11g? What tools are required/available (third party as well as oracle supplied) for these types of tuning scenarios?
I have two tables with 113M records in DWH_BILL_DET & 103M in prd_rerate_chg_que and Im running following merge query, which is running for 13 hrs to update records, which is quiet longer time.
SQL> explain plan for MERGE /*+ parallel (rq, 16) */ INTO DWH_BILL_DET rq USING (SELECT rated_que_rowid, detail_rerate_flag_code, rerate_sel_key,
How the length of column width effects index performance?
For example if i had IOT table emp_iot with columns: (id number, job varchar2(20), time date, plan number)
Table key consist of(id, job, time)
Column JOB has fixed list of distinct values ('ANALYST', 'NIGHT_WORKED', etc...).
What performance increase i could expect if in column "job" i would store not names but concrete numbers identifying job names. For e.g. i would store "1" instead 'ANALYST' and "2" instead 'NIGHT_WORKED'.
I have a question about database fragmentation.I know that fragmentation can reduce performance in query times. The blocks are distributed in many extents and scans process takes a long time. Oracle engine have to locate the address of the next extent..
I want to know if there is any system view in which you can check if your table or index has high fragmentation. If it's needed I will have to re-create, move or rebulid the table or index, but before I want to know if the degree of fragmentation is high.
Any useful script or query to do this, any interesting oracle system view?
There is a simple way to increase the performance of a query by reducing the row-size of the table it hits. I used it in the past by dividing the table into smaller parts and querying respective smaller table in each query.
what is this method called ? just forgot the method and can't recall it. what this type of row-reduction optimization is called ?
How many records could I have in a single table without performance degradation with Standard Edition without partitioning with cutting-edge server (8 or 12 cores, 72 GB RAM, FC 4 Gbit, etc...) and good storage?
300 Millions in only one table with 500K transactions / day is too much?
Testing our 9i to 11g upgrade, we've imported the entire DB into the new machine.We've found that certain procedures are really suffering performance problems. BUT, we've also found, that if we check out a production copy of the procedure from our source code control, and reinstall it, the performance issue goes away. Just alter the procedure and recompiling does NOT work.
The new machine where the 11g database exists is slightly different than the source, but it's not like we have this problem with every procedure. It's only a couple.
any possible reason that we'd have to re-install a procedure to correct a performance problem?
I need to check the package performance and need to improve the package performance.
1. how to check the package performance(each and every statement in the package)? 2. In the package using the delete statement to delete all records and observed that delete is taking long time to delete all the records in the table(Table records 7000000). This table is like staging table.Daily need to clean the data before inserting the data into it. what can I use instead of Delete.
Somewhere I read that we should not use hints in Oracle production environments, but we can use hints in the development environment and on achieving the desired execution plan we can adjust the 'statistics' to follow that plan without hints.
Q1. If it is true what statistics do we adjust for influencing the execution plan and how?
For example, I have the following simple query:
select e.empid, e.ename, d.dname from emp e, dept d where e.deptno=d.deptno;
emp.empid, emp.deptno and dep.deptno columns have indexes and the tables have the standard structure as found in the basic oracle examples.
If I look at the execution plan of the above query then I see that the driving table is empand the driven table is dept.Also the type of join that is taking place is 'Nested Loop'.
Questions: With respect to the above query, Q 2. If I want to make dept the driving table and emp the driven table then how can I adjust the statistics to achieve that? Q 3. If I want to use hash join instead of a nested loop join then then how can I adjust the statistics to achieve that?
I can put the ordered and the use_hash hint to effect this but again I have heard that altering statistics is a more robust way to control an execution plan as compared to hints.
When i exporting an user using expdp utility, the load the on the server is going up-to 5. The size of the database is 180GB. Below is the command that i use for export.
The following query gets input parameter from the Front End application, which User queries to get Reports.There are many drop down boxes like LOB, FAMILY, BRAND etc., The user may or may not select values from drop down boxes.
If the user select any one or more values ( against each drop down box) it has to fetch all matching values from DB. If the user does'nt select any values it has to fetch all the records, in this case application will send a value 'DEFAULT' (which is not a value in DB ) so that the DB will fetch all the records.
For getting this I wrote a query like below using DECODE, which colleague suggested that will hamper performance.From the below query all the variables V_ are defined in procedure which gets the values selected by user as a comma separated string here V_SELLOB and LOB_DESC is column in DB.
DECODE (V_SELLOB, 'DEFAULT', V_SELLOB, LOB_DESC) IN OPEN v_refcursor FOR SELECT /*+ FULL(a) PARALLEL(a, 5) */ * FROM items a WHERE a.sku_status = 'A'
what the principal things to look at when we have for the same query different performance results are?I have 2 different bases: the plan and data are the same but performance results are very differents.
I have created materialized view which hold few million records.Should i have to analyze the view and compute the statistics after i create the materialized view?
Also, just in case i need further indexing, should i have to take the statistics for the table again?
are the most important performance keys we have to calculate or take in account to preserve or to increase the DB performance in terms of response times, and whatsoever according to performance ?
The sqlcache is getting over-written.I would like to capture information from AWR snapshots and feed those (as a workload) into the DBMS_ADVISOR.I can't see where it's possible (other than manually creating my workload from AWR information)
I have a table with around half a million rows. I would like to determine how many sequential events occur per device per unique group of events. The requirement is that the events continue to occur within 1 second of each other, every second, with no gaps. For example, in the sample below there is a sequence of 2 events for device 2737, and another sequence of 5 events for device 2737. If there was a 1 second interval with no event (Signal_time) then that would be the end of one group of events, this is also shown below.
Using Oracle 11.2.0.2.0
DEVICE_ID SIGNAL_TIME 4401 17-JUN-12 12.59.59 PM 2332 02-MAY-12 12.59.59 PM 1086 30-MAY-12 12.59.58 PM 2737 17-JUN-12 12.59.58 PM -- start of one group / sequence of events 2737 17-JUN-12 12.59.58 PM -- end of one group / sequence of events
[Code]...
-- start of one group / sequence of events
2737 16-JUN-12 11.59.59 PM 2737 16-JUN-12 11.59.59 PM 2737 17-JUN-12 12.00.00 AM 2737 17-JUN-12 12.00.01 AM
-- end of one group / sequence of events
2737 17-JUN-12 12.00.03 AM
-- this is excluded from the above group of 5 events, but this may be the start of another group
/*---*/ Select G.id_group, C.rownum id_column from (select distinct id_group from MyTable) G cross join (SELECT rownum FROM dual CONNECT BY LEVEL <= (select max(id_column) from MyTable)) C /*---*/
but it shows record 7, 8 and 9 as members of group 2.
I have a table MYTABLE in database mydb1 duplicated via materialized view and materialized view log and refresh_snapshot commands to a MYTABLE on mydb2 database.
I like to duplicate this table MYTABLE to a third database mydb2, using the same method (materialized view and refresh_snapshot command).
Is it possible ? What's hapend to the materialized view log where I launch a refresh_snapshots on mydb2 ? How is this materialized view log truncated ?
I have set of few hundred reports in Oracle Reports 6i and we have to distinguish them in 3 classes as simple, medium and complex reports. I was wondering what are all the various criterias that should be evaluated to do this .. for example the query behind the report or the layout or graphs complexity etc.
Simplifying the data structure that my problem concern, let's say there are two materialized views between whose data there's a one-to-many relationship [the relationship can be logical without a need of creating any foreing keys].
The data should be as actual as possible respecting the content of the master tables, let's say it shoul be refreshed every 5 minutes.
As far as I know, the jobs related to each snapshot, even if they have a START WITH and NEXT parameter set to the same value, work independently...So: what would be the best manner to synchronize the jobs so as to make sure the data of both snapshots are coherent?
If I create an AWR Report via OEM or awrrpt.sql for an interval between two snapshots, in the report there are statements which were not executed in this interval.
Are there any known bugs for that problem or have you got an explanation for this behavior?
We have a table (call it PROBLEM table) that varies between 60 and 100 million records from one data archival process to the next.The table has data from all 50 states and is range partitioned by each states unique code and has a primary key based on the following:
STATE the data is from BMRK YEAR the year the data was last benchmarked AREA location the data is from SERIES industry the data is located in YEAR the data was reported MONTH the data was reported DATA TYPE the type of data that was reported ESTIMATE TYPE the type of statistical estimate that was produced from the data REPORT ID report id unique to each state REPT SFC state the reporter resided in
We have developed a job that will process data from other tables and create some statistics resulting in approx 2.2 million inserts, .5 million updates and .5 million deletes on PROBLEM table during each job run. Once this table is loaded another process (call this PROBLEM_PROCESS)takes off that reads this updated information to produce some statistical data that is stored in another table. This data is produced by summing anywhere between 2 and 5000 records per query from the PROBLEM table.Here is the query (call it PROBLEM_QUERY)
select round(sum(cm_value*sample_weight),3),round(sum(pm_value*sample_weight),3) from matched_sample where state_fips_code = :1 and area_fips_code = :2 and series_type_code = 'B' and year = :3 and month = :4 [code]...
When we run this job it is very fast (20 minutes) until it gets to the PROBLEM_QUERY, at this point we get DB file scattered read waits, DB file sequential read waits and async IO waits and the job has run for as long as 2 days. When I look at grid, the PROBLEM query is doing a full table scan and ignoring the index.
Our systems people advised us to re-organize and re-index that data in the PROBLEM table after all of the DML and before the PROBLEM_PROCESS takes off. If we do this, it works and cuts our time down alot. However, the re-organize and re-index process adds eight hours to this process. Note: When we re-roganized the data we create a copy table on another set of table spaces and insert the data into the copy sorting the data on the primary key columns. After this we re-index the new table and drop the old table renaming the new table. When we do this later, we move the data back to another set of table spaces. So, we move the data back and for between table spaces A and B we can say.
We have Oracle installed on a M5000 server with Solaris as the OS the binaries and data files are stored on a NetApp Storage array(model 3160) of 500 1TB SATA 7200 RPM drives. However, there are 128 other databases on the NetApp filer as well.
IMO the array and the slow disks are the problem. I believe this because we are catering to the slow disks by re-organizing and re-indexing the PROBLEM table during each run. I don't believe this should be neccessary. We normally re-organize and re-index our data each week in our production system after many more transactions than this.
Our systems people state it is our application. Oracle Support tells us the statistics are out of date but have not answered us on why the statistics are out of date and the index is abandoned after 1 run.
I have some questions about Oracle + EMC shared storage. I have Oracle 11gR1 RAC (2nodes) + ASM environment with shared shorage EMC Clariion AX4.
The database is running no archivelog mode. I'd like to implement point-in-time recovery using Snap View snapshots.
Currently my AX4 platform has the following LUNS:
LUN1 - registry 1 LUN2 - registry 2 LUN3 - vote 1 LUN4 - vote 2 LUN5 - vote 3 LUN6 - ASM - DISKGROUP DATA DISK DATA01 (actual db datafiles) LUN7 - ASM - DISKGROUP DATA DISK DATA02 (actual db datafiles)
Using source LUNs in consistent session will take sync snapshots of all the LUNs working against my database. Once something happens with database, the LUNs can be returned to specific point in time when snapshot consistent session was taken and I'm expecting the database will up and continue to work.
Questions:
1. Is my approach correct at all? (The database is running noarchivelog mode, not dealing with hot backups. The recovery point in time implemented purely via EMC snapshot consistent sessions.) 2. The AX4 platform has a limit 8 source LUNs in session. If I understand it correctly, I can't place more than 8 LUNs in session. What will be the workaround if my database will occupy more than 8 LUNs, I'd like to take their consistent snapshot; for example:
LUN1 - registry 1 LUN2 - registry 2 LUN3 - vote 1 LUN4 - vote 2 LUN5 - vote 3 LUN6 - ASM - DISKGROUP DATA DISK DATA01 (actual db datafiles) LUN7 - ASM - DISKGROUP DATA DISK DATA02 (actual db datafiles) LUN8 - ASM - DISKGROUP DATA DISK DATA03 (actual db datafiles) LUN9 - ASM - DISKGROUP DATA DISK DATA04 (actual db datafiles)
