Oracle Scanner 1.0.6 by patrik@cqure.net
————————————————–
[-] Checking host 120.23.45.104
[-] Checking sid (t432) for common passwords
[x] Got IO Exception
[-] Checking sid (t433) for common passwords
[-] Account DBSNMP/DBSNMP found
[-] Enumerating system accounts for SID (t433)
[-] Succesfully enumerated 22 accounts
[-] Account MDSYS/MDSYS is locked
[-] Account ORDPLUGINS/ORDPLUGINS is locked
[-] Account ORDSYS/ORDSYS is locked
[-] Account OUTLN/OUTLN found
[-] Account PERFSTAT/PERFSTAT found
[-] Checking sid (t434) for common passwords
[-] Account DBSNMP/DBSNMP found
[-] Enumerating system accounts for SID (t434)
[-] Succesfully enumerated 122 accounts
[-] Account MDSYS/MDSYS is locked
[-] Account ORDPLUGINS/ORDPLUGINS is locked
[-] Account ORDSYS/ORDSYS is locked
[-] Account OUTLN/OUTLN found
[-] Account PERFSTAT/PERFSTAT found
…
(This output is not from a real database, its faked

OScanner can be downloaded here. Installation is easy: unzip it. Since it is written in java it can be run on a variety of platforms. Using OScanner is as easy as installing it:

C:\Programme\oscanner> scanner.bat -s 192.168.2.3
Oracle Scanner 1.0.6 by patrik@cqure.net
————————————————–
[-] Checking host 192.168.2.3
…

OScanner saves its reports in xml files. The xml files can be displayed as a tree by OScanner Report Viewer:

C:\Programme\oscanner> reportviewer.exe oscanner_192_168_2_3_report.xml

I don’t like the viewer. I use OScanner like this:

C:\Programme\oscanner> scanner.bat -s 192.168.2.3 > out.txt

Oscanner is a great little tool, but documentation is poor, actually there is no documentation at all.

Installing and configuring RDA takes a few easy steps. Installation and configuration is described in Metalink note 314422.1. RDA can be downloaded for Windows and different flavours of Unix from Metalink.

Have a look at the output of Remote Diagnostic Agent: RDA_Report.zip

Kanton Luzern uses LiveLink for web content management. LiveLink websites are stored in an Oracle database. Kanton Luzern hosts 144 websites.
There is a staging server where content is being added/edited and a live server where websites are accessed from the internet. New content is replicated from stage to live.

Whenever a content manager wishes to publish new or edited content to the live webserver he may trigger replication just by clicking on a button inside the content management tool. This is not a standard feature of LiveLink Content Manager. I wrote an extension for Content Manager, and a stored procedure to control replication.

Materialized View Replication has been introduced with Oracle 8i. Back then it was called Snapshot Replication. For a quick overview of materialized views read Duncan Davies’ post.
To set up materialized view replication is rather simple. But it takes a few steps.

hpfuchs@hplnbk:~> sqlplus /nolog

SQL*Plus: Release 11.1.0.6.0 - Production on Sun Jan 13 10:18:18 2008

Copyright (c) 1982, 2007, Oracle. All rights reserved.
SQL> connect hpfuchs/*****
Connected.

SQL> CREATE TABLE tab1 AS SELECT * FROM dict;

Table created.

Step 1 - Materialized View Log at Master Site
All changes in a base table are written to its materialized view log.

SQL> CREATE MATERIALIZED VIEW LOG ON tab1;
create materialized view log on tab1
*
ERROR at line 1:
ORA-12014: table ‘TAB1′ does not contain a primary key constraint

Ups! Materialized view log can’t be created on tables with no primary key. Let’s create primary key for tab1.

SQL> desc tab1;
Name Null? Type
—————————————– ——– —————————-
TABLE_NAME VARCHAR2(30)
COMMENTS VARCHAR2(4000)

SQL> ALTER TABLE tab1 ADD CONSTRAINT pk_tab1 PRIMARY KEY (table_name);

Table altered.

SQL> CREATE MATERIALIZED VIEW LOG ON tab1;

Materialized view log created.

Step 2 - Set up Materialized View Site
Most likely the materialized view site is another Oracle database. I have only one database on my notebook. Therefore I just create another user who is the owner of the replication site.

SQL> connect system/*****
Connected.
SQL> CREATE USER otherdb IDENTIFIED BY ***** DEFAULT TABLESPACE users QUOTA UNLIMITED ON users;

User created.

SQL> GRANT connect, resource, create materialized view, create database link TO otherdb;

Grant succeeded.

Step 3 - Database Link
To access objects in another database you need a database link between the two databases. Because I have only one database on my notebook, I will create a loopback database link. Omit the @{connect identifier} - in my case: @orcl - to create a real database link.

SQL> connect otherdb/*****
Connected.

SQL> CREATE DATABASE LINK orcl.kt.lunet.ch@orcl CONNECT TO hpfuchs IDENTIFIED BY ***** USING ‘orcl’;

Database link created.

Step 4 - Materialized View

SQL> CREATE MATERIALIZED VIEW mv1 REFRESH FAST AS SELECT * FROM tab1@orcl.kt.lunet.ch@orcl;

Materialized view created.

Materialized views can be indexed.

SQL> CREATE INDEX ix_mv1_objecttype ON mv1 (table_name);

Index created.

Step 5 - Refresh Group
Materialized views can be grouped together into refresh groups. Refresh groups are replicated as a whole. My replication interval is set to one hour. Materialized view replication uses DBMS_JOBS to schedule its intervals. To switch off all scheduled replication activity set JOB_QUEUE_PROCESSES to 0 at the replication site. Make sure all non-replication jobs use DBMS_SCHEDULER. You better check V$JOBS before changing that parameter to 0…

SQL> BEGIN
dbms_refresh.make(
name => ‘mv1_refgroup’,
list => ‘MV1′,
next_date => sysdate,
interval => ’sysdate + 1′,
implicit_destroy => true,
lax => true);
END;
/

PL/SQL procedure successfully completed.

Step 6 - Refresh At Will
The following command refreshes group “mv1_refgroup”.

SQL> exec dbms_refresh.refresh(name => ‘mv1_refgroup’);

PL/SQL procedure successfully completed.

The tiniest of my databases stores only 17 MB of user data. For what waste more than 300 MB of memory for these 17 MB of actual user data? It seems that customers quite often order databases when they actually mean schemas. In most cases many small applications of a customer can be grouped together in one database. And when different departments or even customers use the same application it could be grouped together in one database too.

My goal is to have as few databases as necessary on as few servers as necessary. It’s somehow better to have one mid size database on high availlability hardware than 20 tiny tiny databases on three old standalone servers. It’s easier to manage and less expensive.

In my old job I was working with a wide range of products. Too wide to delve deeply into all of them. I want to narrow this range and focus on something. In the past I already put most of my learning efforts into Oracle technology. And I want to focus even more on Oracle. It’s what I like most and do best. Preferably my new job shouldn’t be in-house IT but at a managed services provider. The company I was looking for should not be a pure managed services provider but also do software development. Perhaps I could use my developer skills too at some point.

I didn’t have to look far. These days it’s easy to find a new job because there are lots of opportunities for database administrators. Starting on February 1st I will work as an Oracle DBA for Abraxas Informatik AG in Zurich. Abraxas’ activities is divided into three areas:

• Business Process Integration
• Software Development
• Managed Services

I count the days…

Of course, GUI tools like Enterprise Manager or SQL Developer offer the same information in a more user friendly manner. But all those GUI tools rely on static dictionary views and dynamic performance views. As a DBA you will likely get into situations where there is no GUI tools available. At times you will be alone with your SQL*Plus skills. Your mastery of data dictionary views will make all the difference!

HKEY_LOCAL_MACHINE/SYSTEM/CurrentControlSet/Services

Once you delete an entry its service is removed and will not show up in the services panel anymore after a system reboot.

We run 26 databases on a SQL Server 2000 machine for customers from every corner of the public administration. Out of those customers only one has already asked if we are going to migrate to SQL Server 2005. Most applications on that server would run as well on SQL Server 2005. Many software companies have not yet finished implementing SQL Server 2005 features into their applications. And many IT professionals are in the middle of developing their SQL Server 2005 skills. I think the market is not ready for SQL Server 2008 at the beginning of next year.