Lowyat.NET: Latest topics by a_sim

JSP Charset Encoding For Vietnamese Not Displaying

a_sim — Wed, 08 Jun 2011 17:56:14 +0800

Hi all,

We already have our web application displaying properly in Simplified chinese (GB18030) and english (Cp1252).

Right now, I am trying to make some changes to have the web application display in Vietnamese. I have tried 'UTF-8' but to no avail.

Below is how I am currently doing it

CODE

<%@ page contentType="text/html;charset=UTF-8"%>

The characters in the database are displayed correctly though when I use Oracle SQLDeveloper to do a

CODE

SELECT * FROM TABLENAME

.

We are currently using
1. Java 1.4
2. Oracle DB 10g
3. Oracle Internet Appliaction Server 10g (OC4J)

Oracle Locks

a_sim — Tue, 10 Jun 2008 18:22:38 +0800

I am having a hard time understanding how and what Oracle locks when a DML (INSERT, UPDATE, DELETE) happens. The reason I need to understand this is because our application recently had 10 deadlocks at the same time. But before going into the deadlocks, I would like to understand Oracle's locking mechanism (if I can call it that)

Here's a scenario:

CODE

SQL> create table item (
2 sku varchar2(13) primary key,
3 ddate date
4 );

Let's insert a record

CODE

SQL> insert into item values ('00000017', to_date('10-06-2008', 'dd-MM-yyyy'));

1 row created.

SQL> commit;

Commit complete.

Then I do an update on this record

CODE

SQL> update item set ddate = to_date('11-06-2008', 'dd-MM-yyyy');

1 row updated.

Next I view the records in v$lock for this session (SID: 150) and below is what I see

CODE

SQL> select type, id1, id2, lmode, request, block
2 from v$lock
3 where sid = '150';

TY ID1 ID2 LMODE REQUEST BLOCK
-- ---------- ---------- ---------- ---------- ----------
TX 262182 223257 6 0 0
TM 103897 0 3 0 0

Here's what I understand from the v$lock result above
1. ID1 where the value is 103897 is the object_id for item (select object_name from dba_objects)

Here's what I am confuse.
1. TM is a lock. Is TM a row, table or row lock?
2. My previous understanding of TX is it is a row lock. But after reading articles on locking in asktom.oracle.com, I found out that TX is just showing a transaction has happened.

Can someone be kind enough to shed some light on this for me?

Many thanks

Problem starting up OC4J

a_sim — Fri, 09 May 2008 11:46:47 +0800

Our production server is currently running on Red Hat Linux with Oracle Application Server 9.0.4.3.0. All has been fine until last month when I added 3 JVMs to one of the OC4Js.

Since then, every time when we stop and start it, we will see the below error message:

QUOTE

08/04/21 23:14:13 Start process
--------
08/04/21 23:14:17 Error instantiating application at file:/oracle/App/j2ee/abc/applications/abc.ear: Unable to find/read assembly info for /oracle/App/j2ee/abc/applications/abc (META-INF/application.xml)
08/04/21 23:14:20 Error initializing site OracleAS Java Web Site: No application named 'abc' found in the server
08/04/21 23:14:20 Oracle Application Server Containers for J2EE 10g (9.0.4.3.0) initialized

And there were a couple of times when we hit the below error message:

QUOTE

08/05/05 07:51:53 Start process
--------
08/05/05 07:51:57 /oracle/App/j2ee/abc/applications/abc/abc/META-INF/MANIFEST.MF could not be deleted ...ArchiveContext.close() returns false Read: false Writ
e: false
08/05/05 07:51:57 /oracle/App/j2ee/abc/applications/abc/abc/META-INF could not be deleted ...ArchiveContext.close() returns false Read: false Write: false
...

Temporary what we did to bring it up is to stop it again and then start it. Then only it can start up normally.

And today, a new error appeared stating that it couldn't find web.xml inside WEB-INF. We had to redeploy the WAR file again.

I have been searching google and oracle's metalink but couldn't find anything.

Thanks.

Linux Server Hang

a_sim — Wed, 02 Apr 2008 18:36:31 +0800

I have a production server which has Red Hat Linux 3.4 (Kernel 2.4.21 - 32).

This morning at 4.04 am, this server hanged and could not be accessed via telnet but there was a response when we pinged it.

Our customer rebooted the server via kvm because the server is actually in another location so he was not able to see if there's a kernel panic message on the server screen.

I am trying to find if a kernel panic did happened or not but I can't find any so far. I have gone through all the log files in /var/log/ and there was no log entries after 4.04 am until the server rebooted at 8.52am. I have also gone through Oracle's log files (this server is running Oracle Application Server) and no log messages that says why it hanged.

Any other way I can find this kernel panic message (if there is a kernel panic) or somewhere I can look at that can tell me what has caused it to hang?

Thanks.

Project Management

a_sim — Wed, 26 Mar 2008 18:36:44 +0800

What tools does your company use to manage its projects? I am interested in something which also allows collaboration from other stakeholders such as project managers, users, system analyst and architects, developers, testers ,etc.

Our company is currently looking for something of sort to implement into our projects.

Thanks.

Weird Behaviour From Oracle Database

a_sim — Tue, 05 Feb 2008 16:43:50 +0800

Hi, it's me again with another problem and this time is the Production database.

Let me tell you about the setup first.

The database is an Oracle 10g Database Standard Edition release 10.1.0.2.0 running on AIX version 5.3 ML 03 with 4 CPUs and 4GB of RAM.

I have created a schema and tablespaces. Below is a list of the tablespaces created:
a) MASTERSPACE
- MASTERSPACE01.DBF [300MB]
- MASTERSPACE02.DBF [300MB]
b) TRANSSPACE
- TRANSSPACE01.DBF [300MB]
- TRANSSPACE02.DBF [300MB]
c) HISTSPACE
- HISTSPACE01.DBF [300MB]
d) WORKSPACE
- WORKSPACE01.DBF [300MB]
e) INDEXMASTERSPACE
- INDEXMASTERSPACE01.DBF [300MB]
f) INDEXTRANSSPACE
- INDEXTRANSSPACE01.DBF [300MB]
- INDEXTRANSSPACE02.DBF [300MB]
g) INDEXHISTSPACE
- INDEXHISTSPACE01.DBF [300MB]
h) INDEXWORKSPACE
- INDEXWORKSPACE02.DBF [300MB]

We have a day job running daily at 1am and this day job consists of a list of programs.

But unfortunately, just one of the program suffered tremendously in this day job. The timing for that program increased from around 2 minutes to 2 hours.

I am curious as to what could have affected the performance of this program.

What I have done to check this problem"
a) Check program.
It has not been changed recently and it has the same md5sum as the other instances which are not having this problem.
b) Ran Oracle trace within the application itself
I changed the program a bit by implement Oracle trace and from the tkprof output, I saw that the program ran for 45 seconds and not 2 hours.
c) Simulated in UAT DB server
With the same tablespaces setup, I ran the day end job again and this program took 2 minutes to finish.
d) Simulated in Live DB Server
I created a test schema and also the same setup of test tablespaces in the live db server and I am able to simulate the problem. Next, I created 3 new tablespaces with 1 database file [500MB] for each tablesoace. Used ALTER TABLE to move the tables from MASTERSPACE and TRANSSPACE to MASTERSPACENEW and TRANSSPACENEW which has only 1 database file each. Ran the day job again and this time 2 minutes.
e) 2nd Simulation in Live DB Server
I created a different setup whereby each tablespaces has total of 300MB of database file but broken into 100MB each e.g. MATERSPACE01.DBF [100MB], MASTERSPACE02.DBF [100MB] and MASTERSPACE03.DBF [100MB]. Ran the day end job again and now the program is taking 2 hours.

The only difference which I could see between the uat db server and the live db server is the aix version. The uat server is aix 5.2.

Why is the performance affected if oracle tablespaces has more than one database file? Is this normal? Our company's DBA consultant can't figure out why.

Thanks.

How Much To Allocate For Swap Space

a_sim — Tue, 22 Jan 2008 12:20:00 +0800

In the office, we are having different opinions on whether the swap space size is sufficient on our customer's Linux server.

The server is a dell server with 2 CPUs and 8GB of RAM. It is running Red Hat Linux ES 3.4 2.4.21-32.ELsmp. This server is our Java Web application server.

Currently the swap space is 2GB. My colleagues are arguing that it is not enough and should be 16GB (2 X 8GB). I am saying it is enough because I do not believe in setting bigger swap space for a production server. 1GB or 2GB is sufficient.

I would like to get some of your opinions on this issue.

Thanks.

Oracle Application Server On Linux Problem

a_sim — Mon, 14 Jan 2008 17:08:38 +0800

Our client's application server has been live for 6 months now and currently there is an issue.

This server is a Dell 6850 running Red Hat Linux ES 3.4 (Taroon Update 5) and kernel 2.4.21-32 with 2 CPUs and 8Gb of ram. Our J2EE web application runs in Oracle Application Server and currently 5 OC4Js are up and running.

The issue we are facing is the server has frequently gone to a hang state. When this happened, we won't be able to telnet into the server but there are ping responses. So the only thing to do then would be to reboot the server.

Now I have setup nmon for linux to run 24 hours in the server taking snap shots of the server's performance every 10 minutes. From the nmon output, I could see free memory is always around 20Mb - 50 Mb and sometimes would shoot up to 1Gb and then in 3 hours back down to 20Mb - 50Mb. Is this a normal pattern for a linux server?

From the Oracle console under memory usage, I saw that around 1.5Gb is being used by the Application Server, 36Mb free and 6.5Gb used by other processes. Now, I am curious as what this other processes are and why are they using 6.5Gb.

I am not sure if this is caused by a faulty hardware, OS/kernel bug, or our web application and I hope you guys in Codemaster could help me.

What I have done:
i) monitor /var/log/messages but no suspicious error being logged not even before or during the server hang.
ii) monitor Oracle Application Sever log files but no suspicious errors logged.
iii) monitor nmon output daily
iv) searched for similar cases using Google and Oracle's Metalink but found nothing.

Thanks.

Managing files between UAT and Live environment

a_sim — Thu, 13 Dec 2007 19:23:04 +0800

Our development team is currently having a serious issue in our deployment process. Our users are complaining that after the deployment, the new functions are not taken effect or the bugs are still there. And when we check the file size it is different from the one in UAT and the last modified date is months ago i.e. the file was not deployed.

Our deployment process is as below:
1) run shell script to append file to a jar file
2) at the end of the day or early in the morning, the jar file is uploaded to the server and extracted out
3) the OC4J instance is stopped and started
4) run few shell scripts which compares file sizes for all the files in each OC4J instance in the UAT and Live environment.

I know that this is mostly due to human error but I am looking for a solution to avoid this problem.

Anyone who was or is in this kind of situation? Or I would greatly appreciate it if you could share with me your deployment process.

Executing java programs outside of OC4J

a_sim — Fri, 14 Sep 2007 19:35:23 +0800

We have two servers - App and DB Server. In the App Server, we have OC4J (Oracle Container for Java) installed and our Java Classes are sitting inside it.

My question is, can I directly port some of our Java classes to the DB Server without installing OC4J on it and execute the Java classes directly via shell script?

I have tried but I am getting a NoClassDefFoundError (below).

CODE

Exception in thread "main" java.lang.NoClassDefFoundError: oracle/dms/console/DMSConsole
at oracle.jdbc.driver.DMSFactory.<clinit>(DMSFactory.java:27)
at oracle.jdbc.driver.OracleConnection.createDMSSensors(OracleConnection.java:522)
at oracle.jdbc.driver.OracleConnection.<init>(OracleConnection.java:345)
at oracle.jdbc.driver.OracleDriver.getConnectionInstance(OracleDriver.java:515)
at oracle.jdbc.driver.OracleDriver.connect(OracleDriver.java:345)
at java.sql.DriverManager.getConnection(DriverManager.java:512)
at java.sql.DriverManager.getConnection(DriverManager.java:171)
at com.util.ejb.connection.DataSource.getJDBCConnection(DataSource.java:98)
at com.util.ejb.connection.DataSource.getLocalConnection(DataSource.java:76)
at com.util.ejb.connection.DataSource.getLocalConnection(DataSource.java:43)
at com.system.logging.DBLogHandler.jbInit(DBLogHandler.java:42)
at com.system.logging.DBLogHandler.<init>(DBLogHandler.java:20)
at com.system.logging.LoggerFactory.setRootLogger(LoggerFactory.java:101)
at com.system.logging.LoggerFactory.getLogger(LoggerFactory.java:39)
at com.batch.EOD.EODActivator.<clinit>(EODActivator.java:27)
command : + /oracle/DB/jdk/bin/java -cp /oracle/App/j2ee/home/oc4j.jar:/oracle/App/j2ee/Pro/applib:/oracle/App/j2ee/Pro/applib/itext-1.3.1.jar:/oracle/App/j2ee/Pro/applib/iTextAsian.jar:/oracle/App/j2ee/Pro/applib/LinearBarCode.jar com.batch.EOD.EODActivator COM TESTSLSUPL

And below is the command in the shell script which I used to execute it.

CODE

/oracle/DB/jdk/bin/java -cp /oracle/App/j2ee/home/oc4j.jar:/oracle/App/j2ee/Pro/applib:/oracle/App/j2ee/Pro/applib/itext-1.3.1.jar:/oracle/App/j2ee/Pro/applib/iTextAsian.jar:/oracle/App/j2ee/Pro/applib/LinearBarCode.jar com.batch.EOD.EODActivator COM TESTSLSUPL

Thanks.

Java PrintWriter not displaying output

a_sim — Fri, 17 Aug 2007 17:08:25 +0800

I have a Servlet which reads a text file and writes it out using PrintWriter. I have no problems when it is in a windows or an AIX environment, but it doesn't work properly in a Linux environment.

To be more specific, in a Linux environment, if the text file has more than 3 lines, it has no problem writing it out. If it has 3 lines or less, then the output will only be out after 1 hour.

I have searched everywhere for a work around but it just didn't work.

This function is suppose to loop continuesly so that the users will be able to see new lines which are created inside the text file.

Below is the method in the Servlet which does the write out:

CODE

public void display(String filename, HttpServletRequest request, HttpServletResponse response)
throws Exception
{
HTTPSessionAttributeWrapper jbSession = (HTTPSessionAttributeWrapper)request.getSession().getAttribute("jbSession");
if (jbSession == null)
{
jbSession = new HTTPSessionAttributeWrapper();
request.getSession().setAttribute("jbSession", jbSession);
}
// set a flag to session to indicate the view status process is running
jbSession.setAttribute("VIEW_STATUS", "Y");
jbSession = null;

this.pwrt = response.getWriter();
BufferedReader IN = new BufferedReader(new FileReader(filename));
String line;
int in_sleep_mode_before = 0;
int count = 0;

flushWinOs();

while(EXECUTE)
{
//System.out.println(new java.util.Date().toString());

line = IN.readLine();

if(line != null)
{
if (in_sleep_mode_before > 0)
{
println("."); // new line
in_sleep_mode_before = 0;
}

// after printed out lines in buffer (arr1 and arr2)
// now we can print all the new comming lines
if (print_all_lines == true)
{
println(line);
response.flushBuffer(); // this also does not work when it is in IBM Server
}
else
{
if (count++ < start_at_last_lines)
{
// keep the last few lines in buffer arr2
arr2.add(line);
}
else
{
// arr2 is full, keep it in arr1
arr1 = arr2;
// init arr2
arr2 = new ArrayList();
count = 0; // reset counter
}
}
}
else
{
if (print_all_lines == false)
printLastLines();

jbSession = (HTTPSessionAttributeWrapper)request.getSession().getAttribute("jbSession");
// if it is null, means user has change to other jsp page,
// so we should stop the process
if (jbSession == null || jbSession.getAttribute("VIEW_STATUS") == null)
{
EXECUTE = false;
break; // exit the while loop
}

jbSession = null;

try{
if (in_sleep_mode_before++ % 60 == 0)
{
println(" "); // new line
flushAix();
}
else
{
print(".");
response.flushBuffer(); // this also does not work when it is in IBM Server
flushAix();
}
Thread.sleep(3800);
} catch(Throwable t){}
}
}

if (this.EXECUTE == false)
{
println("<<Stop reading file ... Exit.>>");
}

this.pwrt.close();
}

When I remark the entire while loop and replace it with the one below:

CODE

while((line = IN.readLine()) != null) {
println(line);
println(".");
}

pwrt.flush();
response.flushBuffer();

This one works. But this is not part of the requirement.

I hope you guys can give me solutions for this. Thanks.

[addedon]August 20, 2007, 4:55 pm[/addedon]i thought it was because the Response Buffer Size which I set was not enough. Initially it was set at 8186. now i am testing at 256 X 1024 and it is still not flushed out.

Anymore suggestions are greatly appreciated.

[addedon]August 20, 2007, 5:04 pm[/addedon]i thought it was because the Response Buffer Size which I set was not enough. Initially it was set at 8186. now i am testing at 256 X 1024 and it is still not flushed out.

Anymore suggestions are greatly appreciated.

Problem starting standalone OC4J

a_sim — Tue, 14 Aug 2007 09:58:05 +0800

I am having problem trying to startup this standalone OC4J in my new laptop.

My laptop is a Dell Vostro with an Intel Core 2 Duo processor and 2Gb Memory and it's running Vista Business Edition.

My java version is as below:

QUOTE

java version "1.4.2_04"
Java™ 2 Runtime Environment, Standard Edition (build 1.4.2_04-b05)
Java HotSpot™ Client VM (build 1.4.2_04-b05, mixed mode)

The error I get when I try to start the standalone OC4J is as below:

QUOTE

C:\oc4j10grob\j2ee\home>java -jar oc4j.jar
07/08/14 01:46:27 Node started with id=11595221023676
07/08/14 01:46:28 Auto-deploying ../../home/applications/admin_ejb.jar (Class 'oracle.oc4j.admin.jmx.ejb.MBeanServerEjbRemote' had been updated)... Error in application default: Error loading package at file:/C:/oc4j10grob/j2ee/home/applications/admin_ejb.jar, javac.exe not found under C:\Program Files\Java\j2re1.4.2_04, please use a valid jdk or specify the location of your java compiler in server.xml using the <java-compiler .../> tag
Error initializing server: Error initializing ejb-module, Exception: javac.exe not found under C:\Program Files\Java\j2re1.4.2_04, please use a valid jdk or specify the location of your java compiler in server.xml using the <java-compiler .../> tag
Fatal error: Cannot instantiate ApplicationServerThread AND not in embedded mode
- forcing exit

My JAVA_HOME and CLASSPATH is as below:

QUOTE

C:\oc4j10g\j2ee\home>echo %JAVA_HOME%
C:\j2sdk1.4.2_04\bin

C:\oc4j10g\j2ee\home>echo %CLASSPATH%
%CLASSPATH%

Java ResultSet problem

a_sim — Mon, 30 Jul 2007 20:49:09 +0800

Hi guys.

I have a program which executes a Select statement to our Oracle DB and the values it returns are numbers. But when my program loops the resultSet reaches a particular record, it will throw a Nullpointer Exception.

I used SQL Tools to view that particular record, but i do not see any null values. The values are either positive, negative or 0.0.

Below is the data i retrieved from SQL Tools

CODE

-74.237435 0 30583.8593 4619.777 264.76250 0 0 0 2773.51716 0 0 29210

Here's a snippet of my code

CODE

double dblA = 0.0;

while (resultSet != null && resultSet.next()) {
dblA = resultSet.getDouble("columnA");
}

I have googled but found no solution. I am currently running java version 1.42.

Thanks.

Managing code

a_sim — Sat, 02 Jun 2007 09:50:22 +0800

I would like to know (if it is practised in your company/work place):

a) does your project manager, senior developer, etc do code reviews on the codes written by your programmers?

b) if code reviews are done, how is it done?

c) does your company have a standard way in writing codes as in a distinctive format on how codes should be written i.e use 4 spaces for tabs or angular brackets should always be on a new line or comments should written this way or all the open connections and statements and resultsets must be in a try-catch-finally block or etc.

d) is there training for the new programmer staff?

The reason I am asking this is recently, we found out that there were problems in our codes, such as:
a) lots of rubbish written in our codes

b) most of the connections and resultsets and statements were not closed after being used or opened

c) there were basically no comments and most of us have to figure our what that piece of code does

d) new objects are always being created in a while or for loop

e) long and ineffective sql statements

f) poor exception and error handling

g) System.out.println() are everywhere in the code

h) functions/methods are not reused but copy-and-paste-and-modify

i) a single class can have more than 3000 or 4000 lines. A report program I saw had a class with 5000 lines.

By the looks of all these, I am beginning to worry. More problems may surface and the program my be harder to debug.

So, what is practised in your company/work place?

Credit Card payment Module

a_sim — Mon, 03 Apr 2006 10:02:24 +0800

Has anyone any experience in doing a web application module which involves credit card payment?

I have this project which enables users to do online booking through the web and pay via credit card. I would like to know how do I go about in building this online booking module. Are there certain parties which I should contact to enable me to pass the user's credit card information?

All I know right now is that I would need to implement SSL into the module? I've also done a brief search and found that Verisign has a service which needs to be puchased to enable the use of credit card payments via web.

Thanks