What I want to do?
What I want to do is this.
There is a process A.
I want to deploy this to the server with name X.
How I did it?
This is just very simple.
Just open up the bpel.xml file
Change the following tag
<BPELProcess id="A"
to
<BPELProcess id="X"
Why I want to do this?
No particular reason actually. One of my friends wanted to know this.
If he wanted to know there could be other too who want to know this.
So I put this up.
Friday, December 19, 2008
Do All ASync BPEL process require correlation?
Why I ask this?
Well it is natural to ask this question once you have read about correlation.
When to use correlation? And a even better question when not to use correlation?
When not to use correlations?
If you have 2 BPEL processes that behave in the following way you do not require correlation.
A calls B --> ASync(Invoke)
B send result back to A --> ASync(Invoke)
This scenario does not require you to use correlation if you are with in BPEL realm.
How does this work?
BPEL does this behind the scenes for you.
Well there is some thing called WS-Adressing that takes care of this.
When does this above rule break?
If you are using a Async callback in a flowN branch you break this rule as there will be more than one process waiting to receive the input. WS-Addressing is not enough.
The other scenario is when you have atleast 3 processes that work as stated below.
A calls B
B calls C
C calls A
Now how does C know which instance of A to call. This forces you to use correlation.
There are other complex scenarios that can make you use correlation that is a variance of the above.
Links
How to use correlation in FlowN
http://oraclebpelindepth.blogspot.com/search/label/FlowN
Well it is natural to ask this question once you have read about correlation.
When to use correlation? And a even better question when not to use correlation?
When not to use correlations?
If you have 2 BPEL processes that behave in the following way you do not require correlation.
A calls B --> ASync(Invoke)
B send result back to A --> ASync(Invoke)
This scenario does not require you to use correlation if you are with in BPEL realm.
How does this work?
BPEL does this behind the scenes for you.
Well there is some thing called WS-Adressing that takes care of this.
When does this above rule break?
If you are using a Async callback in a flowN branch you break this rule as there will be more than one process waiting to receive the input. WS-Addressing is not enough.
The other scenario is when you have atleast 3 processes that work as stated below.
A calls B
B calls C
C calls A
Now how does C know which instance of A to call. This forces you to use correlation.
There are other complex scenarios that can make you use correlation that is a variance of the above.
Links
How to use correlation in FlowN
http://oraclebpelindepth.blogspot.com/search/label/FlowN
Friday, December 12, 2008
getFaultAsString( ) in SOA version and wonderful JDev
Why I write this?
When one of my team members questioned that getFaultAsString( ) was shown as invalid in JDev that had my bells ringing. He also dig up some posts that were talking about some versions.
Well when I read a few posts after googling them out to say that getFaultAsString( ) had gone and has come back. I wanted to see that in my version of SOA suite.
What did I find out?
A few posts read this feature had disappeared in soa 10.1.3 and has come back in soa 10.1.3.3.
So what the hell is this, do it work or not work.
So does it work with soa 10.1.3.4 version.
Working or not in versions?
soa 10.1.3.3 -> Yes it works
soa 10.1.3.3 after MLR patches -> Yes it works
soa 10.1.3.4 -> Yes it works
Whatz it with JDev?
Jdev 10.1.3.3 -> complains in JDev while creating the assign
Function of name "getFaultAsString" not Found
Just Ignore this and proceed with deployment, it will work in server.
Jdev 10.1.3.4 -> worked!! :)
Links I found
http://forums.oracle.com/forums/message.jspa?messageID=2139312
When one of my team members questioned that getFaultAsString( ) was shown as invalid in JDev that had my bells ringing. He also dig up some posts that were talking about some versions.
Well when I read a few posts after googling them out to say that getFaultAsString( ) had gone and has come back. I wanted to see that in my version of SOA suite.
What did I find out?
A few posts read this feature had disappeared in soa 10.1.3 and has come back in soa 10.1.3.3.
So what the hell is this, do it work or not work.
So does it work with soa 10.1.3.4 version.
Working or not in versions?
soa 10.1.3.3 -> Yes it works
soa 10.1.3.3 after MLR patches -> Yes it works
soa 10.1.3.4 -> Yes it works
Whatz it with JDev?
Jdev 10.1.3.3 -> complains in JDev while creating the assign
Function of name "getFaultAsString" not Found
Just Ignore this and proceed with deployment, it will work in server.
Jdev 10.1.3.4 -> worked!! :)
Links I found
http://forums.oracle.com/forums/message.jspa?messageID=2139312
Monday, December 01, 2008
SOA suite 10.1.3.1.0 with oracle 10.2.0.2 joke
Why I write this
I just tried to install SOA suite 10.1.3.1.0 onto my PC.
So i went onto download the oracle DB. That when I was surprised.
What is the version of oracle you want???
The SOA suite advanced install database version support recommended by oracle is
What is this to do with the OS !! :(
I have got Windows XP professional running on my PC.
So naturally I want to download oracle DB version 10.2.0.2 or later for windows.
But I can not find it. Oracle only gives these to download.
Oracle Database 10g Release 2 (10.2.0.1.0) for Microsoft Windows - This I can't use
Oracle Database 10g Release 2 (10.2.0.3/10.2.0.4) for Microsoft Windows Vista and Windows 2008 - This I don't want as this is for Vista
Oracle Database 10g Release 2 (10.2.0.4) for Microsoft Windows Vista x64 and Microsoft Windows Server 2008 x64 - This I don't want as this is for Vista
What the questions in my mind
The basic question is if oracle SOA suite 10.1.3.1.0 needs Oracle DB 10.2.02 or higher. Why is there no Windows Xp compatible 10.2.0.3 or higher. Why only vista?
To get SOA suite onto my Xp, looks like i have to go through a lot of pain to just get the dehydration database up and running. Not at all elegant.
What did i end up doing
Well I end up download the database upgrade for 10.2.0.1 to 10.2.0.2.
How to upgrade 10.2.0.1 to 10.2.0.2
You will find all the information on how to upgrade from 10.2.0.1 database to version 10.2.0.2 follow read me of patch 4547817 or metalink note 316900.1
Basically you will end up downloading the path set for oracle database and applying it.
Just be sure to follow post installation instructions.
Link that helped me was
http://becomeappsdba.blogspot.com/2007/06/upgrade-oracle-database-to-10202-soa.html
I just tried to install SOA suite 10.1.3.1.0 onto my PC.
So i went onto download the oracle DB. That when I was surprised.
What is the version of oracle you want???
The SOA suite advanced install database version support recommended by oracle is
Oracle Database 10g Release 2 (10.2.x) -->10.2.0.2 or later
OK. Thatz cool. Just that I can not download one for the version of OS i have.What is this to do with the OS !! :(
I have got Windows XP professional running on my PC.
So naturally I want to download oracle DB version 10.2.0.2 or later for windows.
But I can not find it. Oracle only gives these to download.
Oracle Database 10g Release 2 (10.2.0.1.0) for Microsoft Windows - This I can't use
Oracle Database 10g Release 2 (10.2.0.3/10.2.0.4) for Microsoft Windows Vista and Windows 2008 - This I don't want as this is for Vista
Oracle Database 10g Release 2 (10.2.0.4) for Microsoft Windows Vista x64 and Microsoft Windows Server 2008 x64 - This I don't want as this is for Vista
What the questions in my mind
The basic question is if oracle SOA suite 10.1.3.1.0 needs Oracle DB 10.2.02 or higher. Why is there no Windows Xp compatible 10.2.0.3 or higher. Why only vista?
To get SOA suite onto my Xp, looks like i have to go through a lot of pain to just get the dehydration database up and running. Not at all elegant.
What did i end up doing
Well I end up download the database upgrade for 10.2.0.1 to 10.2.0.2.
How to upgrade 10.2.0.1 to 10.2.0.2
You will find all the information on how to upgrade from 10.2.0.1 database to version 10.2.0.2 follow read me of patch 4547817 or metalink note 316900.1
Basically you will end up downloading the path set for oracle database and applying it.
Just be sure to follow post installation instructions.
Link that helped me was
http://becomeappsdba.blogspot.com/2007/06/upgrade-oracle-database-to-10202-soa.html
Thursday, October 30, 2008
Default Retired BPEL process
How did I find this?
What was the setup?
Say version 1.0 and version 1.1
Version 1.1 was deployed.
Version 1.0 was retired.
What I wanted to do?
So I retired version 1.1
Set version 1.0 as default
All this was done using the Java API
Why java API?
If you ask me why java API, well there were quite a few BPEL process that need to be rolled back.
And getting this done manually is a monumental pain.
What I expected and found?
What I expected was, once I retire version 1.1
And set version 1.0 as default, version 1.0 will become active
But that did not happen. After version 1.0 was set as default it still was in retired state.
How to do this?
Once I found this it was all logical to say yes this is the expected behaviour.
But this never crossed the mind when preparing the automated rollback process.
Now I know this, possibly other might face or some thing to be aware of.
Monday, October 13, 2008
Create instance with Pick
Break the receive idea!
Well if you think that receive activity is the only way to create BPEL instances, think twice. You can do this with Pick onMessage activity as well. Well if this is a surprise to you then read on else dump this stupid post.
Why do this with PICK
Well if you think about it, it makes it clear that you can have multiple Pick onMessage activities at the same time. What that means in plain english is, you can have multiple operations calling in, to create a instance in BPEL. Is that not cool.
You create BPEL process for each operation in onMessage Pick.
Generally you have single operation to create a instance in BPEL.
How to create instance with PICK
Just check the create instance in the activity, that it. You are done.
What can not be done in this
You can not use a onAlarm activity in this Pick onMessage activity.
If you think about why, it is not logical at all.
How do you expire time on some thing that was not created.
Word of caution
At design time there should be strong reason that you are trying to simulate this at operation level. Use this with caution and for some strong reasons.
For example if you want a single web service for each operation from a source system.
Well if you think that receive activity is the only way to create BPEL instances, think twice. You can do this with Pick onMessage activity as well. Well if this is a surprise to you then read on else dump this stupid post.
Why do this with PICK
Well if you think about it, it makes it clear that you can have multiple Pick onMessage activities at the same time. What that means in plain english is, you can have multiple operations calling in, to create a instance in BPEL. Is that not cool.
You create BPEL process for each operation in onMessage Pick.
Generally you have single operation to create a instance in BPEL.
How to create instance with PICK
Just check the create instance in the activity, that it. You are done.
What can not be done in this
You can not use a onAlarm activity in this Pick onMessage activity.
If you think about why, it is not logical at all.
How do you expire time on some thing that was not created.
Word of caution
At design time there should be strong reason that you are trying to simulate this at operation level. Use this with caution and for some strong reasons.
For example if you want a single web service for each operation from a source system.
Tuesday, October 07, 2008
How do you relate BPEL dehydration tables
Why I write this?
The dehydration database table need to be house kept to clear production data.
What helps you in these scenarios are:
Well if you want to write some real clever applications on to top your BPEL dehydration database, you need to know some key BPEL tables inside out.
The dehydration database table need to be house kept to clear production data.
What helps you in these scenarios are:
- Understanding of what the key tables are.
- What the key fields are.
- How to relate these table.
What is the key table?
There are a few key tables that can be used across to relate together BPEL tables.
The first table that you should know is CUBE_INSTANCE
The most important column that you should know as a BPEL developer is cikey
What is the significance of cikey?
This is the instance ID that you see in your BPEL console.
This gets incremented in a sequence with creation of BPEL instances.
If you don't know what cikey is then here I go. Each BPEL instance is assigned a unique ID.
This ID is visible in the BPEL console.
And guess what BPEL console does allow you to filter using instance ID.
This is one of the key columns that you should know to do a lot of stuff.
And this key cuts across a lot of tables in the dehydration tables.
The article will continue to grow....
Continuation to cover..
What are the other tables that have cikey
What are the other Key columns in CUBE_INSTANCE
Sunday, October 05, 2008
SOA does not start with error "An unknown OPMN error has occured"
When did I get this?
I have a fresh installation of SOA suite on my stupid Vista laptop.
I had a hard time getting this into it in the first place.
If you want to find how hard it was see this
I was damn happy getting this into it and I though wow here I go.
But damn stupid thing, after re-starting my system I could not get the SOA suite working.
I capture what I did to get this wonderful thing working.
The first bolt I got was this
"An unknown OPMN error has occurred"
"Error connecting to OPMN (is it running?): Connection refused: connect"
This was when I tried to start SOA suite from the "start SOA suite" from start menu.
When I looked at opmn.log file it said some thing that did not make any sense initially.
It read "[ons-connect] Local connection 127.0.0.1,6100 invalid form factor"
Tried to see opmnctl status -app but got back
"Unable to connect to opmn.
Opmn may not be up."
How did I solve this?
Solving is very simple process.
Stop all opmn and related process forcefully if does not stop from "stop SOA suite" script
Rename persistence to persistence.old in ${ORACLE_HOME}/j2ee/${OC4J_Insatnce}
Note you will not be able to rename the folder if opmn or any related process is still running.
Use the following command to start SOA suite
opmnctl start
opmnctl startproc process-type=HTTP_Server
opmnctl startproc process-type=${oc4j_instance_name}
Where did I find the solution?
Found the most helpful solution here
Damn why does it suck so much to get SOA suite working on my laptop.
"The procedure entry point snlinAddrLocalhost could not be located in the dynamic link library oranl10.dll" post SOA suite installation
When did I see this?
When I tried to fire up sqlplusw I got this stupid error.
Well I had done nothing wrong but only installed SOA Suite in my laptop.
And here this error presents itself to me.
No clue why some thing should not work when I have not touched the original pieces of the application but installed entirely different application.
What is this stupid error?
This error has nothing to do with your software being corrupted but just a little paying with some thing called the PATH.
Once you install SOA suite, the installer tries to get itself as the first thing in the path. Pushes the oracle database bin path backwards.
How to solve this?
Well if SOA suite can push the path back for oracle database you can do the reverse and have the path come forward in the path list. Volla there you go, thatz it, nothing big but difficult to find.
Just another of those oracle's issues that you discover in the forum and fix.
Links I liked?
The link that sort of gave me the idea was
The post by [dragosv] in the last few entries.
Wednesday, October 01, 2008
Install SOA suite on Windows Vista
Why I write this post?
I have a laptop at home and I want to install SOA suite in that.
Now this piece of equipment that I have got has Windows Vista Home in it.
When I first tried to get this software into it, it was a failure. The setup process itself did not run to completion.
I originally thought that it might take a few hours. But it was days before I could get this working.
How did it start?
I could not get the installation process to complete.
And the error message that I saw was
"Error: Missing ormi[s]://host:port"
That is when I started digging for information.
It turned out that I was not the only one struggling.
I have a laptop at home and I want to install SOA suite in that.
Now this piece of equipment that I have got has Windows Vista Home in it.
When I first tried to get this software into it, it was a failure. The setup process itself did not run to completion.
I originally thought that it might take a few hours. But it was days before I could get this working.
How did it start?
I could not get the installation process to complete.
And the error message that I saw was
"Error: Missing ormi[s]://host:port"
That is when I started digging for information.
It turned out that I was not the only one struggling.
And even the installation of Oracle Database faced issue with Vista with DHCP settings.
Is Windows Vista Home the only standalone?
Well having read the different flavours of Vista had similar issues in forum, my impression was that Vista was the culprit. And when I did install the same software in XP in a flash.
What all did I do?
I found a host of things that need to be done, or at least I need to be aware of before I hit the road. Some are pre-installation and some during installation and some post installation.
Is Windows Vista Home the only standalone?
Well having read the different flavours of Vista had similar issues in forum, my impression was that Vista was the culprit. And when I did install the same software in XP in a flash.
What all did I do?
I found a host of things that need to be done, or at least I need to be aware of before I hit the road. Some are pre-installation and some during installation and some post installation.
If you do not find these in your system, do not be surprised. Every time I have done this I have faced a different issue. Thought not completely sure why so. Might be my system got bored of me with repeated installation and de-installations.
Pre-Installation Issues
Under score in Computer Name
Remove any under scores that you find in your computer name
If you ask me how to do this in Vista? Here are the steps:
Follow the signs and you are done.
Get rid of ::1 localhost from hosts file. To work this out you need to type in some thing in RUN.
Thatz it. I think you are ready to rock DHCP now.
Do not use white space character in directory structure that has the SOA setup files.
Run as Admin
While firing up you setup file, be sure to right click the setup file and choose "run as Administrator". I guess this gives the setup rights to do stuff that running in normal mode does not.
Version of OS
Another post with quite a few issues covered was
Pre-Installation Issues
Under score in Computer Name
Remove any under scores that you find in your computer name
If you ask me how to do this in Vista? Here are the steps:
Right click on MyComputer
Click Properties
Locate some thing called "Computer name, domain, and workgroup settings"
Under there Click "Change Settings"
Some thing called "System Properties" pop up and you will be in a tab called "Computer Name"
You should find a button "Change", click that
Change the Computer Name there
Use any name that does not contain a underscore(_)
Configure loop back adapter
You need to get this loopback adapter into your Vista.
And to get this in do the following
Go to Control Panel (use classic view)
Click "Add Hardware"
Click "Add Hardware"
Click Next
Choose "Install hardware manually"
Choose "Network Adapters" in Common Hardware Type choice
Choose "Microsoft" as Manufacturer and "Loopback" as the adapter
Follow the signs and you are done.
DHCP configuration
Disable IPv6 completely is important for successful installation
Have a look at the Metalink Note 444112.1
If you do not do this installation goes smooth till a point booom. Fails with error unable to locate OC$J URI
Have a look at the Metalink Note 444112.1
If you do not do this installation goes smooth till a point booom. Fails with error unable to locate OC$J URI
Use the following link to disable IPv6
Hosts file
Get rid of ::1 localhost from hosts file. To work this out you need to type in some thing in RUN.
In RUN type in "drivers"
Go into the folder "etc"
Edit the file "hosts"
Comment out part "::1 localhost"
The changed value will look like
#::1 localhost
Completely Disable IPv6
I missed this and I could not complete my installation till I figured that this was the issue.
This is to do with adding a new entry into registery to disable IPv6.
This is documented in
The setups would be like
Run regedit to open registry
Go to HKEY_LOCAL_MACHINE, SYSTEM, CurrentControlSet, Services, Tcpip6, Parameters
Create a new DWORD called DisabledComponents with value 0xFF
Thatz it. I think you are ready to rock DHCP now.
No WhiteSpace Directory
Do not use white space character in directory structure that has the SOA setup files.
Though I have not tested this I could find some people who seem to have faced these issues.
What all this white space actually means is if you get the directory to SOA installation file, there should be no blank space there.
JDK version
JDK version
I tried to run the IRCA script with java 1.6. To my surprise the IRCA complained that I have to go to a lower version of JDK.
I expected the higher version of JDK to work with out issue. But it did not. I had to uninstall and go back to a older version of JDK i.e 1.5.
Installation Steps
Run as Admin
While firing up you setup file, be sure to right click the setup file and choose "run as Administrator". I guess this gives the setup rights to do stuff that running in normal mode does not.
But frankly I have not tried running otherwise.
Version of OS
Do not be surprised if you see some error message stating that version of OS is 6.0.
And installtion does not go any further.
Gives details for getting this working.
Could not get DeploymentManager
Do not be surprise if you see this error message
Failed at "Could not get DeploymentManager".
This is typically the result of an invalid deployer URI format being supplied, the target server not being in a started state or incorrect authentication details being supplied.
More information is available by enabling logging -- please see the Oracle Containers for J2EE Configuration and Administration Guide for details.
If you get this then you have missed the disabling of IPv6. each and every step is important and you can not skip even one of them.
What are the links I liked
Oh my god, at last this is done. Took me a while to get here.
The link that I could find that gave the most details was
Another post with quite a few issues covered was
Tuesday, September 30, 2008
Specify Time in onAlarm
What forced me into it?
This was in a onAlarm activity that I wanted to specify the duration.
This was to set an alarm and retry interval in BPEL.
I saw some documentation providing me incorrect info, and I did have a lot of BPEL instances go to invalid state once they hit the onAlarm.
Now I can not go wrong, I have this post to help me.
So what is the notation?
Well this is the notation as in 10.1.3.3 MLR 5
P1Y1M1DT1H1M1S
Explaination!
This is straight forward.
Duration is
1 Year
1 Month
1 Day
This is for the P1Y1M1D part of the expression
Duration is
1 Hour
1 Minute
1 Second
This is for the T1H1M1S part of the expression
Links I liked!
To be frank I did a few trial and errors in BPEL before I was sure that the doc was not correct.
And I could not Google this up in the first 2 pages I was looking at.
So here is the post that has it. Now I know where to look. I hope it helps others as well.
This was in a onAlarm activity that I wanted to specify the duration.
This was to set an alarm and retry interval in BPEL.
I saw some documentation providing me incorrect info, and I did have a lot of BPEL instances go to invalid state once they hit the onAlarm.
Now I can not go wrong, I have this post to help me.
So what is the notation?
Well this is the notation as in 10.1.3.3 MLR 5
P1Y1M1DT1H1M1S
Explaination!
This is straight forward.
Duration is
1 Year
1 Month
1 Day
This is for the P1Y1M1D part of the expression
Duration is
1 Hour
1 Minute
1 Second
This is for the T1H1M1S part of the expression
Links I liked!
To be frank I did a few trial and errors in BPEL before I was sure that the doc was not correct.
And I could not Google this up in the first 2 pages I was looking at.
So here is the post that has it. Now I know where to look. I hope it helps others as well.
Whatz with the default Services TaskManager and TaskActionHandler
Why I wanted to write this?
When I first started to use BPEL console, I saw 2 service that were already deployed.
That did not make sense to me, and I wanted a clean environment.
I had not deployed anything. And here I see 2 services already deployed.
I wanted to undeploy those, but I decided to find out why.
Then I started to ask question to myself as to why these were there.
What is this TaskManager Service for?
TaskManager service is basically for managing the tasks.
Initiate Task
Update Task
Complete Task
Also does a few things like send callback for
Results for the task and
on expire of task
Basically operations on taks are performed through this service.
What is this TaskActionHandler for?
Task Action Handler as the name states handles actions.
Put items to user's worklist
Get response from the users
Maintain timer events
If you are wondering what timer events are, they are task expiration, task reminder and other such related timer based events.
Links that I liked!
The link you would like to read are
http://www.oracle.com/technology/products/ias/bpel/pdf/orabpel-Tutorial6-TaskManagerServiceTutorial.pdf
When I first started to use BPEL console, I saw 2 service that were already deployed.
That did not make sense to me, and I wanted a clean environment.
I had not deployed anything. And here I see 2 services already deployed.
I wanted to undeploy those, but I decided to find out why.
Then I started to ask question to myself as to why these were there.
What is this TaskManager Service for?
TaskManager service is basically for managing the tasks.
Initiate Task
Update Task
Complete Task
Also does a few things like send callback for
Results for the task and
on expire of task
Basically operations on taks are performed through this service.
What is this TaskActionHandler for?
Task Action Handler as the name states handles actions.
Put items to user's worklist
Get response from the users
Maintain timer events
If you are wondering what timer events are, they are task expiration, task reminder and other such related timer based events.
Links that I liked!
The link you would like to read are
http://www.oracle.com/technology/products/ias/bpel/pdf/orabpel-Tutorial6-TaskManagerServiceTutorial.pdf
Monday, September 29, 2008
Undeploy BPEL with JAVA
What forced me into it?
Undeploying BPEL process from BPEL Console is a nightmare especially when you have many environments. And with each test cycle, the environment is refreshed and recreated.
It takes hell a lot of time and AS admin time to get this. (Well these admin guy are always short of time with their hand full of work)
Best option, automate the process. So I have small piece of code that I have pulled.
You will be surprised how short this is?
package bpelUndeployment;
import java.util.Properties;
import com.oracle.bpel.client.BPELProcessId;
import com.oracle.bpel.client.IBPELDomainHandle;
import com.oracle.bpel.client.Locator;
import com.oracle.bpel.client.ServerException;
/**
* @author Kalidass Mookkaiah
*/
public class UnDeployBPELProcess {
public static void main(String[] args) throws ServerException {
//Properties with BPEL server connection information
Properties props = new Properties();
props.put("orabpel.platform", "ias_10g");
props.put("java.naming.factory.initial", "com.evermind.server.rmi.RMIInitialContextFactory");
props.put("java.naming.provider.url", "opmn:ormi://hostname:6008:home/orabpel");
props.put("java.naming.security.principal", "oc4jadmin");
props.put("java.naming.security.credentials", "oc4jadmin");
props.put("dedicated.connection","true");
//Get a locator in default domain
Locator locator = new Locator("default","welcome1",props);
//Get a handle to the domain
IBPELDomainHandle iBPELDomainHandle = locator.lookupDomain();
//Undeploy this process from default domain
iBPELDomainHandle.undeployProcess(new BPELProcessId("default","MyUndeployedBPELProcess"));
}
}
Explanation!!!
If you do not know that the properties mean, see this post
http://oraclebpelindepth.blogspot.com/2008/08/bpel-context-properties-for-client-api.html
Well if you have read the above post then you know what the Properties are for.
In one word it is values to connect to the AS server.
Use the locator to get a handle to the domain.
Use this handle to perform undeployment.
Here I have used the BPELProcessId class to identify the domain and ProcessName.
The ProcessName should match exactly.
What about versioned BPEL process?
Help your self.
But what I can hint you with is this.
The Class BPELProcessId has overloaded constructor.
3 of them.
One takes domain and processID
Other takes domain, processID and revision Tag.
There is another one tooo with domain, processID, revision Tag and process GUID.
Happy hunting with the overloaded constructors.
And don't forget to run the code ;)
What Jars to include?
And if you are looking to find the jars to include to run this see the link
http://oraclebpelindepth.blogspot.com/2008/09/jars-for-bpel-java-api.html
Undeploying BPEL process from BPEL Console is a nightmare especially when you have many environments. And with each test cycle, the environment is refreshed and recreated.
It takes hell a lot of time and AS admin time to get this. (Well these admin guy are always short of time with their hand full of work)
Best option, automate the process. So I have small piece of code that I have pulled.
You will be surprised how short this is?
package bpelUndeployment;
import java.util.Properties;
import com.oracle.bpel.client.BPELProcessId;
import com.oracle.bpel.client.IBPELDomainHandle;
import com.oracle.bpel.client.Locator;
import com.oracle.bpel.client.ServerException;
/**
* @author Kalidass Mookkaiah
*/
public class UnDeployBPELProcess {
public static void main(String[] args) throws ServerException {
//Properties with BPEL server connection information
Properties props = new Properties();
props.put("orabpel.platform", "ias_10g");
props.put("java.naming.factory.initial", "com.evermind.server.rmi.RMIInitialContextFactory");
props.put("java.naming.provider.url", "opmn:ormi://hostname:6008:home/orabpel");
props.put("java.naming.security.principal", "oc4jadmin");
props.put("java.naming.security.credentials", "oc4jadmin");
props.put("dedicated.connection","true");
//Get a locator in default domain
Locator locator = new Locator("default","welcome1",props);
//Get a handle to the domain
IBPELDomainHandle iBPELDomainHandle = locator.lookupDomain();
//Undeploy this process from default domain
iBPELDomainHandle.undeployProcess(new BPELProcessId("default","MyUndeployedBPELProcess"));
}
}
Explanation!!!
If you do not know that the properties mean, see this post
http://oraclebpelindepth.blogspot.com/2008/08/bpel-context-properties-for-client-api.html
Well if you have read the above post then you know what the Properties are for.
In one word it is values to connect to the AS server.
Use the locator to get a handle to the domain.
Use this handle to perform undeployment.
Here I have used the BPELProcessId class to identify the domain and ProcessName.
The ProcessName should match exactly.
What about versioned BPEL process?
Help your self.
But what I can hint you with is this.
The Class BPELProcessId has overloaded constructor.
3 of them.
One takes domain and processID
Other takes domain, processID and revision Tag.
There is another one tooo with domain, processID, revision Tag and process GUID.
Happy hunting with the overloaded constructors.
And don't forget to run the code ;)
What Jars to include?
And if you are looking to find the jars to include to run this see the link
http://oraclebpelindepth.blogspot.com/2008/09/jars-for-bpel-java-api.html
JARs for BPEL Java API
What forced me into it?
To get the BPEL Java API working the first thing to do is to have the java class be available for you to code.
And to get these classes into your work space, what else but include the libs.
Every time I do this I Google. I am either very bad at Google or there are not many people who write about this.
Now I want this as a reference for myself, and also for those who want the list.
What are the jars that I included? And why in that Order?
The first jar that I included was orabpel.jar
For what other reason that the fact that it has what I need.
And what I need to start is Locator.
To be more specific I need com.oracle.bpel.client.Locator class
Is this jar enough?
Well if you think this is enough then it is time for you to run and see for yourself.
Do not be surprised to get error that looks like this
Exception in thread "main" java.lang.NoClassDefFoundError: javax/ejb/EJBException
Make this error disappear?
How else but include the jar that has this class. And the jar is ejb.jar
I think now it is time to run this and see for your self.
Whatz the next error?
Well you have gussed it right. There is another error that comes up.
This time round it looks like this
Exception in thread "main" java.lang.NoClassDefFoundError: com/collaxa/common/util/NonSyncPrintWriter
at com.oracle.bpel.client.util.ExceptionUtils.handleServerException(ExceptionUtils.java:82)
at com.oracle.bpel.client.Locator.getFinder(Locator.java:962)
.......
Caused by: java.lang.NoClassDefFoundError: com/collaxa/common/util/NonSyncPrintWriter
at com.oracle.bpel.client.util.BeanRegistry.lookupFinderBean(BeanRegistry.java:351)
at com.oracle.bpel.client.Locator.getFinder(Locator.java:956)
...
Make this error disappear? And next error appears.
How else but include the jar that has this class. And the jar is orabpel-common.jar
You know what time this is. Yes run time. So letz see te next error.
The next error you might be seeing is
Exception in thread "main" java.lang.Exception: Failed to create "ejb/collaxa/system/FinderBean" bean; exception reported is: "javax.naming.NoInitialContextException: Cannot instantiate class: com.evermind.server.rmi.RMIInitialContextFactory [Root exception is java.lang.ClassNotFoundException: com.evermind.server.rmi.RMIInitialContextFactory]
....
Caused by: java.lang.ClassNotFoundException: com.evermind.server.rmi.RMIInitialContextFactory
...
Make this error disappear? And...
This time round I include the jar oc4j-internal.jar
The next error you might be seeing is
Exception in thread "main" java.lang.NoClassDefFoundError: oracle/ias/opmn/optic/OpticException
...
Caused by: java.lang.NoClassDefFoundError: oracle/ias/opmn/optic/OpticException
...
Make this error disappear? And...
This time round I include the jar optic.jar
The next error you might be seeing is nothing....
What was the code that I ran?
There might be other jars that need to be included based on what is that you want the Java do perfom.
But the java code that I ran is for undeploying BPEL process.
The code is in the post
http://oraclebpelindepth.blogspot.com/2008/09/undeploy-bpel-with-java.html
To get the BPEL Java API working the first thing to do is to have the java class be available for you to code.
And to get these classes into your work space, what else but include the libs.
Every time I do this I Google. I am either very bad at Google or there are not many people who write about this.
Now I want this as a reference for myself, and also for those who want the list.
What are the jars that I included? And why in that Order?
The first jar that I included was orabpel.jar
For what other reason that the fact that it has what I need.
And what I need to start is Locator.
To be more specific I need com.oracle.bpel.client.Locator class
Is this jar enough?
Well if you think this is enough then it is time for you to run and see for yourself.
Do not be surprised to get error that looks like this
Exception in thread "main" java.lang.NoClassDefFoundError: javax/ejb/EJBException
Make this error disappear?
How else but include the jar that has this class. And the jar is ejb.jar
I think now it is time to run this and see for your self.
Whatz the next error?
Well you have gussed it right. There is another error that comes up.
This time round it looks like this
Exception in thread "main" java.lang.NoClassDefFoundError: com/collaxa/common/util/NonSyncPrintWriter
at com.oracle.bpel.client.util.ExceptionUtils.handleServerException(ExceptionUtils.java:82)
at com.oracle.bpel.client.Locator.getFinder(Locator.java:962)
.......
Caused by: java.lang.NoClassDefFoundError: com/collaxa/common/util/NonSyncPrintWriter
at com.oracle.bpel.client.util.BeanRegistry.lookupFinderBean(BeanRegistry.java:351)
at com.oracle.bpel.client.Locator.getFinder(Locator.java:956)
...
Make this error disappear? And next error appears.
How else but include the jar that has this class. And the jar is orabpel-common.jar
You know what time this is. Yes run time. So letz see te next error.
The next error you might be seeing is
Exception in thread "main" java.lang.Exception: Failed to create "ejb/collaxa/system/FinderBean" bean; exception reported is: "javax.naming.NoInitialContextException: Cannot instantiate class: com.evermind.server.rmi.RMIInitialContextFactory [Root exception is java.lang.ClassNotFoundException: com.evermind.server.rmi.RMIInitialContextFactory]
....
Caused by: java.lang.ClassNotFoundException: com.evermind.server.rmi.RMIInitialContextFactory
...
Make this error disappear? And...
This time round I include the jar oc4j-internal.jar
The next error you might be seeing is
Exception in thread "main" java.lang.NoClassDefFoundError: oracle/ias/opmn/optic/OpticException
...
Caused by: java.lang.NoClassDefFoundError: oracle/ias/opmn/optic/OpticException
...
Make this error disappear? And...
This time round I include the jar optic.jar
The next error you might be seeing is nothing....
What was the code that I ran?
There might be other jars that need to be included based on what is that you want the Java do perfom.
But the java code that I ran is for undeploying BPEL process.
The code is in the post
http://oraclebpelindepth.blogspot.com/2008/09/undeploy-bpel-with-java.html
Thursday, September 25, 2008
FTP file name: Both Get fileName and dynamic Put filename
How did it make into my workspace?
What else but a requirement.
A file is dynamically created based on the web service message received.
The file that needs to be FTPed has a name, that is dynamic.
The file name is one of the fields in the message received.
Well this was about Put.
How the hell did Get file name come to my workspace?
Well it was not a requirement in my project. But was a help that I did for a friend.
This person wanted to read files from a server.
And drop this off to another server after some enrichment's to the content.
And wanted the Get and Put file names to match.
So this adds a new thing of getting the filename of the file that was read from FTP directory.
How does it work?
If you have created FTP Get and Puts you should see
ftpAdapterInboundHeader.wsdl and ftpAdapterOutboundHeader.wsdl file in your workspace.
These are the 2 key file to get this cracking.
Both of these have a element called fileName.
What to do with this element?
You have to either play with these elements to get the work done.
To play with these elements you first need to create a variable from the wsdl.
How to get the file name of the get file?
Just use the variable that you have created from the ftpAdapterInboundHeader.wsdl.
Read the fileName element. Yes it is that easy.
How to assign file name to put file?
Well nothing different from the above one.
But this time round you need not read but assign value to the fileName element.
The variable that you have created from the ftpAdapterOutboundHeader.wsdl, use that.
Assign the file name to the fileName element in this variable.
Thatz it. You have done it.
What Does it do in the background?
What you are trying to do here is either read or write to the FTP Adapter Header Message.
For Inbound messages the header has the fileName, FTP server and FTP port.
For Outbound message the header has the fileName.
You can play with these.
From Inbound message you can read the values.
For Outbound if you populate these values the Adapter will behave accordingly.
What else but a requirement.
A file is dynamically created based on the web service message received.
The file that needs to be FTPed has a name, that is dynamic.
The file name is one of the fields in the message received.
Well this was about Put.
How the hell did Get file name come to my workspace?
Well it was not a requirement in my project. But was a help that I did for a friend.
This person wanted to read files from a server.
And drop this off to another server after some enrichment's to the content.
And wanted the Get and Put file names to match.
So this adds a new thing of getting the filename of the file that was read from FTP directory.
How does it work?
If you have created FTP Get and Puts you should see
ftpAdapterInboundHeader.wsdl and ftpAdapterOutboundHeader.wsdl file in your workspace.
These are the 2 key file to get this cracking.
Both of these have a element called fileName.
What to do with this element?
You have to either play with these elements to get the work done.
To play with these elements you first need to create a variable from the wsdl.
How to get the file name of the get file?
Just use the variable that you have created from the ftpAdapterInboundHeader.wsdl.
Read the fileName element. Yes it is that easy.
How to assign file name to put file?
Well nothing different from the above one.
But this time round you need not read but assign value to the fileName element.
The variable that you have created from the ftpAdapterOutboundHeader.wsdl, use that.
Assign the file name to the fileName element in this variable.
Thatz it. You have done it.
What Does it do in the background?
What you are trying to do here is either read or write to the FTP Adapter Header Message.
For Inbound messages the header has the fileName, FTP server and FTP port.
For Outbound message the header has the fileName.
You can play with these.
From Inbound message you can read the values.
For Outbound if you populate these values the Adapter will behave accordingly.
Monday, September 22, 2008
The requested URL /integration/services/deploy/deployHttpClientProcess.jsp was not found on this server
How the hell did I get it?
I tried deploying BPEL process from automated deployment script that I have got.
This script uses ANT process to do the job.
What I see is the below error.
--------------------- Error Message Start ------------------------
BUILD FAILED
/nas/orasoft/orbit/oracle/soasuite_pc_deploy_v1.2/bpel_deploy/build.xml:223: The following error occurred while executing this line:
/nas/orasoft/orbit/oracle/soasuite_pc_deploy_v1.2/bpel_deploy/src/ErrorLogging/build.xml:79: A problem occured while connecting to server "orbit" using port "7777":
<HTML><HEAD>
<TITLE>404 Not Found</TITLE>
</HEAD><BODY>
Not Found
The requested URL /integration/services/deploy/deployHttpClientProcess.jsp was not found on this server.
<HR>
<ADDRESS>Oracle-Application-Server-10g/10.1.3.1.0 Oracle-HTTP-Server Server at orbit Port 7777</ADDRESS>
</BODY></HTML>
--------------------- Error Message End ------------------------
Where did I find the solution?
Well first I read a few posts in oracle forum.
The post that helped me solve or at least pointed towards a solution was
http://kr.forums.oracle.com/forums/thread.jspa?threadID=631670
Healthy Green hw_service!!
My hw_services was working and was green in colour.
Well that meant that it was healthy.
But later I found it was not configured for deployment (read to find where I found the jsp)
Next I wanted to find the directory. /integration/services/deploy/deployHttpClientProcess.jsp
I could not. So I started a search to find it! And yes I did find it. Not as a directory but as a configuration in Active Servlets/JSPs.
Where did I find /integration/services/deploy/deployHttpClientProcess.jsp?
I tried to find this in the hw_services from EM (enterprise manager). And yes I did.
Thanks to the post there.
These are the steps that I followed.
Go to EM.
From the home page click the + sign of OC4J instance used.
Click hw_services to open up the properties.
Click the deploy Web Module.
You will find "deployHttpClientProcess.jsp" under the Active Servlets/JSPs.
That is where I found this guy registered.
What happened to deployment from JDEV?
The deployment from JDev went with out any issues.
This issues was with ANT deployment only.
What I found and how I solved it?
Before I tried to solve the issue. I wanted to find this. And I did as above.
My DEV ENV was working so I wanted to see that before I did any thing else.
I found this configuration present in my DEV environment.
I found this missing in my SYS TEST ENV and that had this above error.
What I found was "deployHttpClientProcess.jsp" was not not present in the Active Servlets/JSPs list at all. It was blank in DEV.
The very next thing I did was to stop and start hw_services.
And when I did the deployment went smoothly.
And the "deployHttpClientProcess.jsp" came back to the Active Servlets/JSPs.
I tried deploying BPEL process from automated deployment script that I have got.
This script uses ANT process to do the job.
What I see is the below error.
--------------------- Error Message Start ------------------------
BUILD FAILED
/nas/orasoft/orbit/oracle/soasuite_pc_deploy_v1.2/bpel_deploy/build.xml:223: The following error occurred while executing this line:
/nas/orasoft/orbit/oracle/soasuite_pc_deploy_v1.2/bpel_deploy/src/ErrorLogging/build.xml:79: A problem occured while connecting to server "orbit" using port "7777":
<HTML><HEAD>
<TITLE>404 Not Found</TITLE>
</HEAD><BODY>
Not Found
The requested URL /integration/services/deploy/deployHttpClientProcess.jsp was not found on this server.
<HR>
<ADDRESS>Oracle-Application-Server-10g/10.1.3.1.0 Oracle-HTTP-Server Server at orbit Port 7777</ADDRESS>
</BODY></HTML>
--------------------- Error Message End ------------------------
Where did I find the solution?
Well first I read a few posts in oracle forum.
The post that helped me solve or at least pointed towards a solution was
http://kr.forums.oracle.com/forums/thread.jspa?threadID=631670
Healthy Green hw_service!!
My hw_services was working and was green in colour.
Well that meant that it was healthy.
But later I found it was not configured for deployment (read to find where I found the jsp)
Next I wanted to find the directory. /integration/services/deploy/deployHttpClientProcess.jsp
I could not. So I started a search to find it! And yes I did find it. Not as a directory but as a configuration in Active Servlets/JSPs.
Where did I find /integration/services/deploy/deployHttpClientProcess.jsp?
I tried to find this in the hw_services from EM (enterprise manager). And yes I did.
Thanks to the post there.
These are the steps that I followed.
Go to EM.
From the home page click the + sign of OC4J instance used.
Click hw_services to open up the properties.
Click the deploy Web Module.
You will find "deployHttpClientProcess.jsp" under the Active Servlets/JSPs.
That is where I found this guy registered.
What happened to deployment from JDEV?
The deployment from JDev went with out any issues.
This issues was with ANT deployment only.
What I found and how I solved it?
Before I tried to solve the issue. I wanted to find this. And I did as above.
My DEV ENV was working so I wanted to see that before I did any thing else.
I found this configuration present in my DEV environment.
I found this missing in my SYS TEST ENV and that had this above error.
What I found was "deployHttpClientProcess.jsp" was not not present in the Active Servlets/JSPs list at all. It was blank in DEV.
The very next thing I did was to stop and start hw_services.
And when I did the deployment went smoothly.
And the "deployHttpClientProcess.jsp" came back to the Active Servlets/JSPs.
Sunday, September 21, 2008
Arrays in BPEL
Why it made itself into my Project?
What else but FlowN. That was the culprit.
FlowN will use index to create separate flows for each of the parallel flows using the index variable.
And the index was used to read data from where else but array.
How to create a array?
Actually it is not as complex as I originally though it to be.
Array in XML can be any element that can store multiple repetition of some data.
How to add value into the array?
What else but use append. I you have not seen this, it looks like a + in the assign activity. Just see an assign.
Behind the scenes the bpel code looks like:
<assign name="Assign_CDM_Contents_Append">
<bpelx:append>
<bpelx:from variable="Variable_Add_Array_Element"
query="/ns1:ABC"/>
<bpelx:to variable="Variable_Array"
part="payload"
query="/ns1:ArrayList"/>
</bpelx:append>
</assign>
Here Variable_Array is the variable to hold the array of Variable_Add_Array_Element
Whatz with XPath?
There are 2 options to read data from XML arrays.
One is to use another XPath variable to build the dynamic XPath query.
Second option is to not use a new variable but dynamically insert the XPath in the bpws:getVariableData()
I prefer the first option as it gives a little clarity over the second option.
Why Dynamically build XPath?
Well what ever the case be you need to build XPath dynamically.
But why? Just to use the changing index into the Array.
How to Read data?
If you have a running index over either a flowN or a while loop that does this for you, either way you need the index to access the array.
Now the way to read is through dynamic XPath. That need to be built.
Here I have used option 1.
The below code shows how to use assign to build the dynamic XPath.
<assign name="Assign_XPath_For_Array">
<copy>
<from expression="concat('/ns1:ArrayList/ns2:ArrayElem[',bpws:getVariableData('Variable_Index_Counter'),']') "/>
<to variable="Variable_Xpath"/>
</copy>
</assing>
Note have used [index] to represent index element here in the array.
Here Variable_Xpath acts as a temporary string variable that holds the intermediate Xpath as string.
Now use this XPath string Variable while reading array index elements.
This would look like
<assign name="Assign_Read_Array">
<copy>
<from expression="bpws:getVariableData('Variable_Array','payload',bpws:getVariableData('Variable_Xpath'))"/>
<to variable="Variable_Read_Array_Element"
query="/ns1: ABC"/>
</copy>
</assign>
How to do with out another variable?
Just do not use the new variable. Build your XPath dynamically right in the bpws:getVariableData() 3rd parameter.
What else but FlowN. That was the culprit.
FlowN will use index to create separate flows for each of the parallel flows using the index variable.
And the index was used to read data from where else but array.
How to create a array?
Actually it is not as complex as I originally though it to be.
Array in XML can be any element that can store multiple repetition of some data.
For example if A represents sequence of element B then A can be used as array for B.
What is that I am going to show you below?
In short this is what I am attempting to show you.
1. Add elements into the array using BPEL append.
Most of the cases people already have a array from the source system.
Use this when you want to build your ow array i the fly.
2. Build the XPath
If you have decided to use array then build a dynamic XPath.
Store that in a temporary variable.
And use that XPath temporary variable while reading elements from the array.
If you ask me what type of variable, I prefer String type.
2. Read data from array
Just as plain as use the temporary XPath variable and read data from the array element.
Just like saying A[X] i ay other lang. Here A is the array and X is the temporary XPath.
Movig in rounds to read like A[1], A[2] etc.
How to add value into the array?
What else but use append. I you have not seen this, it looks like a + in the assign activity. Just see an assign.
Behind the scenes the bpel code looks like:
<assign name="Assign_CDM_Contents_Append">
<bpelx:append>
<bpelx:from variable="Variable_Add_Array_Element"
query="/ns1:ABC"/>
<bpelx:to variable="Variable_Array"
part="payload"
query="/ns1:ArrayList"/>
</bpelx:append>
</assign>
Here Variable_Array is the variable to hold the array of Variable_Add_Array_Element
I am trying to add elements into the Array (Variable_Array).
I am doing this by appending the elements to be end of it.
Whatz with XPath?
There are 2 options to read data from XML arrays.
One is to use another XPath variable to build the dynamic XPath query.
Second option is to not use a new variable but dynamically insert the XPath in the bpws:getVariableData()
I prefer the first option as it gives a little clarity over the second option.
Why Dynamically build XPath?
Well what ever the case be you need to build XPath dynamically.
But why? Just to use the changing index into the Array.
How to Read data?
If you have a running index over either a flowN or a while loop that does this for you, either way you need the index to access the array.
Now the way to read is through dynamic XPath. That need to be built.
Here I have used option 1.
The below code shows how to use assign to build the dynamic XPath.
<assign name="Assign_XPath_For_Array">
<copy>
<from expression="concat('/ns1:ArrayList/ns2:ArrayElem[',bpws:getVariableData('Variable_Index_Counter'),']') "/>
<to variable="Variable_Xpath"/>
</copy>
</assing>
Note have used [index] to represent index element here in the array.
Here Variable_Xpath acts as a temporary string variable that holds the intermediate Xpath as string.
Now use this XPath string Variable while reading array index elements.
This would look like
<assign name="Assign_Read_Array">
<copy>
<from expression="bpws:getVariableData('Variable_Array','payload',bpws:getVariableData('Variable_Xpath'))"/>
<to variable="Variable_Read_Array_Element"
query="/ns1: ABC"/>
</copy>
</assign>
How to do with out another variable?
Just do not use the new variable. Build your XPath dynamically right in the bpws:getVariableData() 3rd parameter.
Tuesday, September 16, 2008
How to perform manual copy of bpel code for faster development!! With out issues
What forced it?
Well as part of the BPEL development I had used a set of Enterprise Integration Patterns that each particular type of BPEL scope code should follow.
When each of those BPEL processes are created, they follow the same Pattern, a lot of code looked similar. But are not the same.
That where the catch was!!
SIMILAR but not the SAME!!!
What I wanted to do?
Well what I originally planned to do was to copy some of the code from one of the already built BPEL scope onto the new one I was building.
And I would be seamlessly able to reuse the stuff.
What that meant from code perspective was, I would like a scope from another BPEL code into my current BPEL code.
Well originally I though it was a brilliant idea. And finally it did turn up that way.
But it was damn hard till I figured out the trick. Not a trick actually but divide and conquer.
What are the problems I faced?
The first time I did that. I saw ERROR. Actually a brilliant one.
"XML parsing failed"
Nothing else. How the hell is that supposed to help me?
How did I make it work?
Well that is when I put my experimenting hand into the BPEL xml code.
The first thing you need to worry about are the non-declared variables that have come over as part of the copy process.
Resolution : declare them.
The second thing would be the name space that are new or namespaces that already exist.
Resolution : Either create the new namespace or when appropriate use the old namespace.
The third and the best helping hand is. Comment out groups of code. Leave a logical complete unit and compile. Follow the above steps. Fix apparent errors are they come.
Then go on to the next logical unit and repeat till finally you have completed the full Scope.
Just be aware of the Catch, if you have them. I hope you will as part of the scope. Copy them as well.
Why I like this better than drag and drop?
Some people advice not to do that but I prefer this better than the GUI to build.
If you ask me why? I would say this is damn faster to build.
And that is what brings me back into the BPEL xml code.
And the second reason is the way JDeveloper behaves.
Some times it does weired thing. Does not save thing that I have changed.
Some times goes back to the old code after I save.
It catches you by surprise some times as you are damn sure that you made the change that does not exist.
What is it with templates?
Some people recommend templates but they give complete BPEL template.
Not scopes to copy. But templates are good options if the BPEL processes are damn alike.
Well as part of the BPEL development I had used a set of Enterprise Integration Patterns that each particular type of BPEL scope code should follow.
When each of those BPEL processes are created, they follow the same Pattern, a lot of code looked similar. But are not the same.
That where the catch was!!
SIMILAR but not the SAME!!!
What I wanted to do?
Well what I originally planned to do was to copy some of the code from one of the already built BPEL scope onto the new one I was building.
And I would be seamlessly able to reuse the stuff.
What that meant from code perspective was, I would like a scope from another BPEL code into my current BPEL code.
Well originally I though it was a brilliant idea. And finally it did turn up that way.
But it was damn hard till I figured out the trick. Not a trick actually but divide and conquer.
What are the problems I faced?
The first time I did that. I saw ERROR. Actually a brilliant one.
"XML parsing failed"
Nothing else. How the hell is that supposed to help me?
How did I make it work?
Well that is when I put my experimenting hand into the BPEL xml code.
The first thing you need to worry about are the non-declared variables that have come over as part of the copy process.
Resolution : declare them.
The second thing would be the name space that are new or namespaces that already exist.
Resolution : Either create the new namespace or when appropriate use the old namespace.
The third and the best helping hand is. Comment out groups of code. Leave a logical complete unit and compile. Follow the above steps. Fix apparent errors are they come.
Then go on to the next logical unit and repeat till finally you have completed the full Scope.
Just be aware of the Catch, if you have them. I hope you will as part of the scope. Copy them as well.
Why I like this better than drag and drop?
Some people advice not to do that but I prefer this better than the GUI to build.
If you ask me why? I would say this is damn faster to build.
And that is what brings me back into the BPEL xml code.
And the second reason is the way JDeveloper behaves.
Some times it does weired thing. Does not save thing that I have changed.
Some times goes back to the old code after I save.
It catches you by surprise some times as you are damn sure that you made the change that does not exist.
What is it with templates?
Some people recommend templates but they give complete BPEL template.
Not scopes to copy. But templates are good options if the BPEL processes are damn alike.
Create Empty Tag in Tranformations
What forced me into it?
The message received from a web service had less data than what the XML (cannonical Data model needed) had. Down the line there were conditions that check the value of a element in the XML. These could be empty or some XYZ value.
Now the transformation had to map these value onto the XML, but if I did not map them the XML will miss the elements.
What I wanted to see?
What I wanted to see was, I should be able to perform the transformation from the Input Variable to the Output Variable.
Those elements that were required should not be lost.
These elements should be as empty tags in the Output Variable.
What did I get?
The transformation lost those elemements that I did not map.
That left me with only a few elements in the Output Variable.
Less than the required elements.
How I fixed this?
First of all I could not change the XML definition as the XMl is shared my a host of systems.
And it would be a massive effort to get this done.
Now a simple work around for this would be to assign a blank to the element in the transformation.
Now if you are think what is that supposed to mean, here is what I mean.
In the transformation right click the element that you dod not want to loose as that is not mapped.
Choose "setText" option in there.
Give a space in the "Text" field that you see there.
Congrats you have done this. Was simple aint it.
What does it look like back there?
Well inthe transaformation it will look like this
<xsl:text disable-output-escaping="no"> </xsl:text>
If you want to try?
If you have noted this, the xsl will have a space between the tags.
Delete that. And see what you get.
I have seen no difference on my instance.
The message received from a web service had less data than what the XML (cannonical Data model needed) had. Down the line there were conditions that check the value of a element in the XML. These could be empty or some XYZ value.
Now the transformation had to map these value onto the XML, but if I did not map them the XML will miss the elements.
What I wanted to see?
What I wanted to see was, I should be able to perform the transformation from the Input Variable to the Output Variable.
Those elements that were required should not be lost.
These elements should be as empty tags in the Output Variable.
What did I get?
The transformation lost those elemements that I did not map.
That left me with only a few elements in the Output Variable.
Less than the required elements.
How I fixed this?
First of all I could not change the XML definition as the XMl is shared my a host of systems.
And it would be a massive effort to get this done.
Now a simple work around for this would be to assign a blank to the element in the transformation.
Now if you are think what is that supposed to mean, here is what I mean.
In the transformation right click the element that you dod not want to loose as that is not mapped.
Choose "setText" option in there.
Give a space in the "Text" field that you see there.
Congrats you have done this. Was simple aint it.
What does it look like back there?
Well inthe transaformation it will look like this
<xsl:text disable-output-escaping="no"> </xsl:text>
If you want to try?
If you have noted this, the xsl will have a space between the tags.
Delete that. And see what you get.
I have seen no difference on my instance.
Monday, September 15, 2008
Change my Partnerlink Property Dynamically
What forced me into it?
This was in production. And there were 2 releases planned into production.
This was a property that was defined in bpel.xml file in the partnerLinkBinding.
As part of the second release the property had to be changed.
How did I do it?
Well I used java embedding to get this done.
Created a java embedding to change the property value in the partnerLinkBinding
Sample bpel.xml file
<?xml version = '1.0' encoding = 'UTF-8'?>
<bpelsuitcase>
<bpelprocess id="UpdateThis" src="UpdateThis.bpel">
<partnerlinkbindings>
<partnerlinkbinding name="client">
<property name="wsdlLocation">UpdateThis.wsdl</property>
<property name="myPartnerlinkProperty">myValueOfPartnerlinkProperty</property>
</partnerlinkbinding>
</partnerlinkbindings>
<configurations>
<property name="configProperty" encryption="plaintext">myBPMSystemAdmin</property>
</configurations>
</bpelprocess>
</bpelsuitcase>
Sample java embedding code to change these
try {
getLocator().lookupProcess ("UpdateThis").getDescriptor ().
getConfigurations().
setPropertyValue ("configProperty",
"BPMSystemAdmin");
getLocator().lookupProcess ("UpdateThis").getDescriptor ().
getPartnerLinkBindings().
getPartnerLinkBinding ("client").
setPropertyValue ("myPartnerlinkProperty",
"myChangedValueOfPartnerlinkProperty");
}
catch(Throwable ex) { .... }
Whatz the issue with restarting AS?
The first time I saw this it looked stupid. I changed the property from using another dummy bpel process. Used that happily till I restarted the AS.
And when I did the values in the partnerlink went back to those present earlier.
What I tried but failed?
And then I realized that the AS will reload all the BPEL process from the SUITCASE_BIN
The changed functionality with 10.1.3.X
A option that I tried but failed was to change the DESCRIPTOR column value in the PROCESS_DESCRIPTOR table.
But that did not help with restart. It always went back to the values that were present as part of the original deployment.
If you are thinking where can I find these tables, they are in the ORABPEL schema.
This was in production. And there were 2 releases planned into production.
This was a property that was defined in bpel.xml file in the partnerLinkBinding.
As part of the second release the property had to be changed.
How did I do it?
Well I used java embedding to get this done.
Created a java embedding to change the property value in the partnerLinkBinding
Sample bpel.xml file
<?xml version = '1.0' encoding = 'UTF-8'?>
<bpelsuitcase>
<bpelprocess id="UpdateThis" src="UpdateThis.bpel">
<partnerlinkbindings>
<partnerlinkbinding name="client">
<property name="wsdlLocation">UpdateThis.wsdl</property>
<property name="myPartnerlinkProperty">myValueOfPartnerlinkProperty</property>
</partnerlinkbinding>
</partnerlinkbindings>
<configurations>
<property name="configProperty" encryption="plaintext">myBPMSystemAdmin</property>
</configurations>
</bpelprocess>
</bpelsuitcase>
Sample java embedding code to change these
try {
getLocator().lookupProcess ("UpdateThis").getDescriptor ().
getConfigurations().
setPropertyValue ("configProperty",
"BPMSystemAdmin");
getLocator().lookupProcess ("UpdateThis").getDescriptor ().
getPartnerLinkBindings().
getPartnerLinkBinding ("client").
setPropertyValue ("myPartnerlinkProperty",
"myChangedValueOfPartnerlinkProperty");
}
catch(Throwable ex) { .... }
Whatz the issue with restarting AS?
The first time I saw this it looked stupid. I changed the property from using another dummy bpel process. Used that happily till I restarted the AS.
And when I did the values in the partnerlink went back to those present earlier.
What I tried but failed?
And then I realized that the AS will reload all the BPEL process from the SUITCASE_BIN
The changed functionality with 10.1.3.X
A option that I tried but failed was to change the DESCRIPTOR column value in the PROCESS_DESCRIPTOR table.
But that did not help with restart. It always went back to the values that were present as part of the original deployment.
If you are thinking where can I find these tables, they are in the ORABPEL schema.
Friday, September 12, 2008
Database Adapters JNDI name in JDev and Application server
Why I used it?
Well in all 3 of big projects that I worked in, Db adapters was required. I had no choice but use it.
Why JNDI?
Well JNDI itself provides one level of indirection from the physical database connection.
Decouples the actual database connection code from the application logic.
If you are thinking cut the theory and tell me the real thing, how to do this in BPEL, that's exactly what I am about to do.
Where is this done for DBAdapter?
If you had a look at DB Adapter, JNDI is a configuration that comes up while following the wizard for DBAdapter. Now where the hell did I set up JNDI connections and how will AS (application server [for those who might not guess AS is this]) figure this out.
Well actually you need to do this in 2 places. One in DBAdapter and another in AS.
Why I need this in AS?
Well how can AS figure out which database and other connection details.
What else but you need to do this in application server as well!
If you are asking what all I need to setup in AS just read on!!
What is the order to do this?
Both of these are independent and can be done in any order. (By both I mean JDev setup and AS setup)
Just be sure that both of these are done before you execute any code in BPEL else your opmn logs will show thing that you never like to see, ERRORS.
How to relate them? Whatz this use the same name thing?
Both the JDev and AS JNDI names should be the same or else AS will not be able to relate these together.
What's the AS setup?
There are 3 basic things that need to be setup in AS before AS JNDI can be used.
Number 1 the physical connection pool with all the service details.
Number 2 the Data source that points to the connection pool.
Number 3 setup the JNDI name as in JDev and pointing to the Data Source.
What do I setup in JDev?
In the connections in JDev create a connection to the database that was created in AS connection pool. Else you will not be able to explore the stuff in the Database.
What is with this mcf thing?
If you setup connection in JDev there is some thing that comes on board but not directly visible.
What else but the mcf properties.
Does not harm anything but if the JNDI are missing in AS starts using this to connect.
So for production it is best that these mcf details are removed.
If you don't know that where it is located just create a project with JNDI see it for your self.
It will be present in the Database Adapter generated files. Do you want clues!! I guess not.
Friday, August 22, 2008
Correspond manual recovery back to the BPEL process
Looks complex, but is it?
Manual recovery of BPEL process does not look like a trivial job at all.
And the BPEL Console for manual recovery does do seem to reveal a lot of information as to which process corresponds to which instance of BPEL process in the first glance.
Why do instances end up in manual recovery?
Well this is to do with how BPEL engine handles in coming messages.
BPEL delivery service does 2 things
And when the instance is complete the dehydration stores goes to HANDLED state.
If the server shuts down or crashes or engine times out and comes back unable to find the JMS message as it might be already consumed.
What does Manual Recovery show?
Manual recovery shows all instances that are in UNRESOLVED state.
So it would also show instance going from UNRESOLVED to HANDLED states in dehydration store. But are not yet marked as HANDLED.
From what I have seen in production box, a lot of instances come into manual recovery and disappear in a short while.
Should I recover everything that I see in Manual Recovery?
One should not recover every message that you see in manual recovery.
Some of these could be genuine messages that are in flight.
Recover only those messages that are in manual recovery for at least a duration of X which is larger than expected in that particular enterprise system.
What and where do you recover?
Well when you look at the manual recovery area there are 3 different tabs that you generally see in the BPEL console. Most often you would only recover from the first 2 that are represented invoke and callbacks.
Generally I use the instance ID of the BPEL process to identify manual recovery instance corresponds to which BPEL instance.
If there are instances in staled state in BPEL, there there would be a correspoonding manual recovery process instance.
This manualt recovery instance conversation ID will have also contain the instance ID of the BPEL process along with other information with which the conversation Id is build.
What if there is no instance ID but some thing like MD5{xyz...}
These are messages for which instances are not yet created. Can be safely recovered.
Manual recovery of BPEL process does not look like a trivial job at all.
And the BPEL Console for manual recovery does do seem to reveal a lot of information as to which process corresponds to which instance of BPEL process in the first glance.
Why do instances end up in manual recovery?
Well this is to do with how BPEL engine handles in coming messages.
BPEL delivery service does 2 things
- Use JMS to register message to be processed
- Save message in dehydration store
And when the instance is complete the dehydration stores goes to HANDLED state.
If the server shuts down or crashes or engine times out and comes back unable to find the JMS message as it might be already consumed.
What does Manual Recovery show?
Manual recovery shows all instances that are in UNRESOLVED state.
So it would also show instance going from UNRESOLVED to HANDLED states in dehydration store. But are not yet marked as HANDLED.
From what I have seen in production box, a lot of instances come into manual recovery and disappear in a short while.
Should I recover everything that I see in Manual Recovery?
One should not recover every message that you see in manual recovery.
Some of these could be genuine messages that are in flight.
Recover only those messages that are in manual recovery for at least a duration of X which is larger than expected in that particular enterprise system.
What and where do you recover?
Well when you look at the manual recovery area there are 3 different tabs that you generally see in the BPEL console. Most often you would only recover from the first 2 that are represented invoke and callbacks.
Generally I use the instance ID of the BPEL process to identify manual recovery instance corresponds to which BPEL instance.
If there are instances in staled state in BPEL, there there would be a correspoonding manual recovery process instance.
This manualt recovery instance conversation ID will have also contain the instance ID of the BPEL process along with other information with which the conversation Id is build.
What if there is no instance ID but some thing like MD5{xyz...}
These are messages for which instances are not yet created. Can be safely recovered.
Relate BPEL process instance parent to child relationship
Why I want to relate these?
Before I start I consider this scenario that prompted me to write this stuff.
Well I have a sequence of BPEL processes that get called.
A
--> B
----> C
-------> D
-------> E
----> F
I want a way to relate all the process together in the BPEL console or bespoke application built on top of the BPEL Database.
How to do this?
There are multiple ways to relate these BPEL process.
Title approach
Well the one I prefer is to set the title of the BPEL process instances as some logical value that makes sense to business. What that would mean is all the BPEL processes from "A" to "F" will all be name as "XYZ".
But this is good for the bespoke applications.
Well some times before the actual title of the process is set the BPEL process might go into manual recovery. How do you solve?
Database approach
Well go into the BPEL database and look.
Now how do I relate all the processes in the database.
Well first you need to go the table "ORABPEL"."CUBE_INSTANCE"
There are 3 specific columns that you should know about.
"TITLE" - sets the title of the BPEL process in this column as seen in the BPEL console
"ROOT_ID" - CIKEY (instance id as in BPEL console) of the root parent basically instance ID of A as in BPEL console
"PARENT_ID" - Immediate parents ID i.e the caller process
Sample Scenario
A is the parent
A calls B
B calls C and F
C calls D and E
-------------------------------------------------------
Process Name - Instance ID - Root ID - Parent ID --> comment
-------------------------------------------------------
A - 1 - 1 - --> parent Id blank as no parent for it but root Id is 1
B - 2 - 1 - 1 --> both parent and root ID as same as A
C - 3 - 1 - 2 --> root is still 1, but parent now is B
D - 4 - 1 - 3 --> root is still 1, but parent now is C
E - 5 - 1 - 3 --> root is still 1, but parent now is C
F - 6 - 1 - 2 --> root is still 1, but parent now is B
Sample Query
So sample queries to find all the children in a parent BPEL tree would be to use the root ID
================ Query START ==================================================
select * from cube_instance where root_id = {ROOT aprent instance number}
================ Query END ====================================================
A query to find all root BPEL process i.e BPEL process that are the initiators
================ Query START =================
select * from cube_instance where parent_id is NULL
================ Query END ===================
Before I start I consider this scenario that prompted me to write this stuff.
Well I have a sequence of BPEL processes that get called.
A
--> B
----> C
-------> D
-------> E
----> F
I want a way to relate all the process together in the BPEL console or bespoke application built on top of the BPEL Database.
How to do this?
There are multiple ways to relate these BPEL process.
Title approach
Well the one I prefer is to set the title of the BPEL process instances as some logical value that makes sense to business. What that would mean is all the BPEL processes from "A" to "F" will all be name as "XYZ".
But this is good for the bespoke applications.
Well some times before the actual title of the process is set the BPEL process might go into manual recovery. How do you solve?
Database approach
Well go into the BPEL database and look.
Now how do I relate all the processes in the database.
Well first you need to go the table "ORABPEL"."CUBE_INSTANCE"
There are 3 specific columns that you should know about.
"TITLE" - sets the title of the BPEL process in this column as seen in the BPEL console
"ROOT_ID" - CIKEY (instance id as in BPEL console) of the root parent basically instance ID of A as in BPEL console
"PARENT_ID" - Immediate parents ID i.e the caller process
Sample Scenario
A is the parent
A calls B
B calls C and F
C calls D and E
-------------------------------------------------------
Process Name - Instance ID - Root ID - Parent ID --> comment
-------------------------------------------------------
A - 1 - 1 - --> parent Id blank as no parent for it but root Id is 1
B - 2 - 1 - 1 --> both parent and root ID as same as A
C - 3 - 1 - 2 --> root is still 1, but parent now is B
D - 4 - 1 - 3 --> root is still 1, but parent now is C
E - 5 - 1 - 3 --> root is still 1, but parent now is C
F - 6 - 1 - 2 --> root is still 1, but parent now is B
Sample Query
So sample queries to find all the children in a parent BPEL tree would be to use the root ID
================ Query START ==================================================
select * from cube_instance where root_id = {ROOT aprent instance number}
================ Query END ====================================================
A query to find all root BPEL process i.e BPEL process that are the initiators
================ Query START =================
select * from cube_instance where parent_id is NULL
================ Query END ===================
Why not use Tree Finder?
Itz a good question to ask why not use tree finder.
In the projects that I worked, access to the BPEL console is restricted and access to the BPEL is only through a bespoke J2EE application. But that application does not allow viewing of BPEL instances are that is for functinal people to see work flow tasks related to business errors.
The alternative that was implemented was a read only user to the BPEL database from where these queries are helpful to relate BPEL instances. Especially when performing manual recovery of stales instances and to identify the tree chain that has stalled in the flow.
Hope this answers question raised as a commnet here.
Subscribe to:
Posts (Atom)
