All entries for Friday 27 January 2006

January 27, 2006

Valueable Skills

Writing about web page

Tom mentioned :

  • flexibility: or you can say smart. I met quite a few smart Java developers
  • willingness to change: this is the principal of Agile
  • No being afraid to say "..." : this is a bit hard

Like James Bond:
ready to learn, not afraid to change.


5 out of 5 stars

Great Book.It bring me the whole world of refactoring.

Eclipse IDE makes refactoring job a joyable task. I can seriously say: it is fun.


2 out of 5 stars

Not very good


10g Complete Reference

2 out of 5 stars

Not very good. Superficially cover too many contents.


Oracle BAckup & Recovery Handbook

4 out of 5 stars

A old version Oracle book, talking about Oracle 9i, 2 generations behide now.

It once helped me a lot and the pricipals still apply


4 out of 5 stars

A good book, but it is not very well organized.

Duplicate to a Remote Server P408

  • have to either copy backups to or NTF mount backups at the remote serve
  • copy archivelogs from target to auxiliary
  • On target server, configure tnsnames.ora to point auxilariy instnce; On auxiliary sitse, need to edit listener.ora
  • rman target / auxiliary system/*@auxi_server
  • sql 'alter system switch logfile';
  • duplicate target database to nofilenamecheck pfile=/pfile/init.ora logfile '/loca1' size 100m, '/loca2' size 100m
  • Other clauses include: 1) skip tablespace 'users','tools' ; 2) until sequcne =1 thread = 1

Duplicate to a Same Server P406

  • make sure db_file_name_convert = ('target','aux');

High Availability

4 out of 5 stars

Oracle Database 10g New Features

5 out of 5 stars

Pretty good book. Give a very through and accurate pictures of 10g exciting new features. Much better to read a book than read Oracle e-manual.

  • P97: Flashback drop You do not need to enable Flashback Database in order to use this feature. This is great. In any 10g db, you can flashback dropped table
  • Flashback types: flashback database, flashback versions query (versions between timestamp), flashback transaction query (flashback_transaction_query), flashback table (enable row movement)

Effective Oracle Databases 10g Security by Design

4 out of 5 stars

P50 encrypt network is easy:


SQLNET.CRYPTO_SEED = dfksjkjrkejrkejrkji33333o-__9i9

Oracle OTN has detailed documentation as well

David Knox claims “independent lab tests how little overhead”

Specify the following properties in JBOSS connection pool configuration file if JBOSS runs on same server as Oracle and you do not want the overhead of encryption on a local connection.

<connection-property name="">REJECTED</connection-property>
<connection-property name="">REJECTED</connection-property>

Wait Interface

5 out of 5 stars


  • MMON collects data into AWR
  • MMNL sample data into active session history.
  • ADDM use AWR

OWI Component:

  • V$SYSTEM_EVENT:Health check instance; order by TIME_WAITED; To know bottlenecks between certain time frame, sample periodically and compute the differences(delta), this is used in my healthCheck app.
  • V$SESSION_EVENT: Query this view for particular SID and order by TIME_WAITED to check the bottlenecks. A goold place to get initial idea of problems but not suitalbe for root cause analysis.
  • V$SESSION_WAIT: To investigate a currently slow db, query this view repeatedly in quick successions. Pay attention to STATE=WAITING and SECONDS_IN_WAIT. V$SESSION_WAIT_HISTORY provide history information
  • time is centisecond (cs, 1 hundredths of a second) by default. Millisecond (ms, m means milli, not milliion), Microsecond ( 1 millionth of a second)


  • db file scattered read: FTS or Index Fast Full Scan (P1: file number; P2: starting block no; P3: numbers of block to read)
  • db file sequential read: a sginel block read operation occure when reading from index, rollback, table access by rowid (P1, P2 same as above; P3 : 1 in most cases)
  • db file parallel write: average_wait should less than 100ms (query v$system_event). On one of our server, the value is 11cs, not very good.
  • log file paralle write: average_wait should less than 10ms (1cs). On our server, it is 2 for log write, 4 for log sync. Not good.
  • control file paralle write: usually is symptomatic of high log switches
    select /*+ ordered */ a.sid, 
    decode (a.type, 'BACKGROUND','BACKGROUND-' 
       || substr(a.program, instr(a.program, '(',1,1)),'FORGROUND') type, 
    b.time_waited, round (b.time_waited/b.total_waits, 4) average_wait,
    round((sysdate -a.logon_time)*24) hours_connected
    from v$session a, v$session_event b
    where a.sid=b.sid and b.event='control file parallel write'
    order by type,time_waited;

Trace Event

  • Process Trace Event:
    alter session set events '10046 trace name context forever, level 12';
  • CBO Trace Event:
    alter session set events '10053 trace name context forever, level 2'; 
    alter session set events '10053 trace name context off'


  • Latches: apply to data structures in SGA and two modes(wrlling-to-wait/no-wait)
  • Lock: apply to db object s like tables, data blocks; six modes (null,row share..)
  • Enqueue: the act of placing a lock request in a queue (verb) or a specific lock such a sTX enqueue (noun)

High–Performance SQL Tuning

Not rated

9i Performance Tunning Reading Note

4 out of 5 stars


  • Give an good example of using stored outline

P325 show

  • CBO will choose the driving table cleverly, regardless of table order in the query;
  • The large table is the driven table after getting PRODUCT_ID passwd to it by the join. Very surprise !! I would think at least thins will change in example 3


  • The added index only contains 3 columns, not including the two columns in the join conditions. Why ? I would add them into the index.


  • The minimum of Oracle statistics gathered should be file I/O (V$FILESTAT), system statistics (V$SYSSTAT), and SQL statistics (V$SQLAREA or V$SQL, V$SQLTEXT and V$SQL_PLAN)
  • Buffer gets (V$SQLAREA.BUFFER_GETS, for high CPU using statements)
  • Disk reads (V$SQLAREA.DISK_READS, for high I/O statements)
  • Sorts (V$SQLAREA.SORTS, for many sorts)
  • You should verify the CBO cost of the statement with the actual number of resources used (BUFFER_GETS, DISK_READS, CPU_TIME from V$SQL or V$SQLAREA).

Search this blog

Search Warwick Blogs



Most recent comments

  • Oracle 11g is the clear leader in this area. Oracle has invested heavily into self–tuning capabiliti… by shaane on this entry
  • All the features are very nice.I like Repair Advisors the most.Oracle 10G introduced some of the dat… by lucy on this entry
  • "logical standby failed to re–start, I am also getting the same error. I have search on internet and… by anemia on this entry
  • Thanks for the nice information because from last three days i am stuck on this problem…. by robot vacuum on this entry
  • So what happen City speed dating when you tried this coding once in your PC.. Please tell me so i am… by John Bergeron on this entry
Not signed in
Sign in

Powered by BlogBuilder