Jeevan Sunkersett

Ranch Hand
+ Follow
since Jul 03, 2007
Merit badge: grant badges
For More
Cows and Likes
Total received
In last 30 days
Total given
Total received
Received in last 30 days
Total given
Given in last 30 days
Forums and Threads
Scavenger Hunt
expand Ranch Hand Scavenger Hunt
expand Greenhorn Scavenger Hunt

Recent posts by Jeevan Sunkersett


Mine is a j2EE, application working in a j2EE server.

For tracing execution I use logback/ sl4j instead of System.out.println() ... so that execution can be diagnosed by reading the log file

However, in my application I use a 3rd party jar, that internally uses System.out.println(), even exceptions were being printed using e.printStackTrace().

How can I redirect the stdout and stderr of this 3rd party jar, to the RollingFileAppender defined use in my SL4J.

I cannot update/ edit the 3rd party jar.

8 years ago
there are multiple - jSonde, MaintainJ (commercial), Eclipse-Diver.

My issue: I want to create/ define Queue and Topic Connection factory in IBM WebSphere MQ, with support for XA Transactions
BUT using the JMSAdmin tool, provided - I do NOT want to use the UI (MQExplorer)

I could create, Q and Topic Connection Factory, after launching JMSAdmin console and using

DEFINE qcf(myName-of-QConFactory) qmgr(my-QMgr-name)
but this way they get created using OR

<< I want What should I do ? >>

I can get the desired, if create/ define using the MQExplorer UI,
and check the "Support XA Transactions" check box (in the dialog)

How to do the same, using the JMSAdmin prompt and DEFINE keywords

10 years ago

I am fairly new to JSR 303, Bean validation.
I wish to know how I can use it for validating certain complex conditions, for instance

I have a object:

1. Person (name, age)
2. Account (type, minimumBalance, dailyWithdrawalLimit)

I think we can use JSR 303, to validate Person or Account alone, say
a. person.age must be more than 5.
b. cannot be blank
c. account.type can have only 2 values - 'Savings' or 'Checking'

Is it possible to use JSR 303, for something more complex;

a. If account.type = 'Savings' then
account.minimumBalance IS $1000/-
account.dailyWithdrawalLimit IS $ 5000/-

but If account.type = 'Current' then
account.minimumBalance IS $ 1.0/-

Is it possible to use JSR 303, alone (nothing else) to validate across objects?

eg. If person.age < 17, then account.type CAN-Only-be 'Savings'

If it is not possible, to use JSR 303 alone, what will I have to use in addition to JSR 303?

Would appreciate if can point me to links/ blogs/ articles, which show case, how such validations are achieved.

thank you

We are trying to setup a cluster of Tomcat servers.

We are using Tomcat 7.x and a mix of Windows and Linux servers.

As shown in the attached picture, we have been able to get 3 nodes in cluster, but cannot get the 4th one - what could be the issuse?

What could be so special on the L-95 linux server, because of which it does not join the cluster ?

Any tips, to debug will be appreciated.


The server.xml of tomcat\conf ... on all machines is identical

I am consider W-82 as the primary node of the clusster; so the <cluster> element inside server.xml is bit more elaborate; viz.
10 years ago
My app is a Web application, employs Spring, Apache CXF, ... and is deployed on Tomcat-7.
In META-INF\context.xml I specify an additional class-path element

At initialization time, it loads up a property file, (using PropertyPlaceholderConfigurer)
from within the jar/ war as well as from the file system (if found in the additional class-path)
For this I have used the following

The logs specifically say;
Loading properties file from URL [jar:file:/D:/WMSandbox/tools/apache-tomcat-7.0.40/wtpwebapps/Myserver-web/WEB-INF/lib/Myserver-1.0.0-SNAPSHOT.jar!/]
Loading properties file from URL [file:/D:/opt/xyz/config/]

But the path(s) are (is) only getting printed in the log.
At run time, I need the exact file:/ path in a variable - how can I get that?


gave me only the exact paths of the jar's used by my application but not the file:/ path of the property file (loaded from the file system/ outside the .jar)

Any clues, how I can get the full path of the property file (loaded from the file system)

Thank you
11 years ago
Thank you Paul.

My client reverted with a specific questions .... please can you help?

For Hibernate to work, do you need both before and after image for the journaled files?

Please can you help me answer that?


As I could not figure out the appropriate forum for my question posting here
- Dear Moderator, Please move as appropriate.

My query:
How can a handheld RFid device (used for scanning bar codes on products etc) be interfaced with a web/ browser based application/ page.

The user of my application, would be viewing the data entry page in a browser.

He would type in data for a few fields manually and for a couple would use a hand held RF gun to scan the bar-code label of the item; with the expectation to get that value populated in the web page.

I think this should be easy with a desktop based (thick client) application but in my case the data entry form is browser based.

So requesting some tips in this area.

thanking you
11 years ago

I want a bit of advice.

My current project involves a legacy modernization - the current app is developed on AS/400 (iSeries) using the DB2/400 database (operating system: OS/400)

My to-be system: A modern j2EE based application, with a browser based UI, utilizing jQuery, Spring MVC, coupled with Spring POJO and Hibernate ORM

However my client has made substantial investment in hardware and so wants to retain the AS/400 machine and wants the new modern application on the same DB2/400 database.

(the application server, would be: IBM WebSphere 6.x or whichever is latest)

I chanced across some old posts, in which the writer has expressed problems when using Hibernate with DB2/400 (which is normally non-journaled in production)

( )
( )

I had thought, Hibernate was database independent; simply change the dialect and it can work with "any" database.

But these posts have left me thinking -
* these posts are pretty old (5 year old), so does the same still hold true?
* given these findings, I have a serious doubt - should or should-I-not adopt Hibernate, for my migration project; if not what ORM can I use?

Please advise.



I am contemplating if I should employ a workflow or BPM suite (like jBPM/ Intalio/ joget) or build my app using a simple Struts+Spring+Hibernate OR JSP+Servlet+database technology stack.

Looking at the several examples; available with the above products; like
* Leave appliation and approval
* Self appraisal, followed by super-visor & HR evaluation and then computing a result.

I could very well implement such use cases, using a Struts+Spring+Hibernate OR JSP+Servlet+database technology stack.

On the UI, the requestor submits a requests.
In the controller layer/ Action class it can be validated and then in the DAO layer, inserted into a database.
(followed by sending an email to the supervisor/ hr) to also login and approve/ reject the request.

When supervisor or HR login, in a table I can display the items pending (giving it l&f of a Inbox), (be retrieving from the database)

If they wish to proceed, display the details (in a JSP) with a approve/ reject button.
The same will be submitted to the Action class and updated into the database; with an email sent to the requestor.

If I were to use a workflow product and model this in BPEL; where is the benefit?
I can find several problems though
1. Steep learning curve
2. I will have to develop UI (and invest time for making it aesthetically appealing) .... let alone the costs of integrating into a BPM.
3. Deployment and huge infrastructure costs (to support the BPM suite), as approsed to a simpler App server + database setup.

Please can you help me decide?

thank you

12 years ago

I completed my SCEA in 2009.

Please can you suggest the next certification - that I could pursue? TOGAF or ?

thank you


for the benefit of all,

the root cause was identified as bug in Oracle

Bug 8769746 - Describe returns wrong type for column added by ALTER TABLE [ID 8769746.8] . You can see the details in the

To avoid this thing in future never add a new column to an existing table with CHAR datatype with a default value.

The issue is fixed in :
  • (Server Patch Set)
  • Patch 23 on Windows Platforms

  • thank you.
    12 years ago

    We are experiencing an odd issue in production; related to Golden gate replication. Request feedback to debug and find the root cuase.
    (please excuse the formatting as this is the best I could achieve and make it more readable)

    Production environment:
  • 1. Red Hat Linux 5 servers with Oracle 11 G, one primary and another DR.
  • 2. with Golden Gate (Version Build 078) for real time replication between the two.

  • In the primary database,
    We have a table "COUNTER" in schema 'DPS', in which column 'PHOTOCAPTURED' has data type CHAR(1)
    from our change request history, we know this column was added using the ALTER table command

    This ALTER table command was perfectly replicated by GG (as determined by doing a DESC on DR database)

    But when data is added to this column GG gives error, when replicating;

    The REPLICAT job is in ABENDED state; error

    ERROR OGG-01163 Bad column length (5) specified for column PHOTOCAPTURED in table DPS.COUNTER, maximum allowable length is 3."

    This error occurs only when data is added to the above mentioned COUNTER table.

    After a bit of exploration it was found that data in DR database
    table 'COUNTER' column PHOTOCAPTURED (with CHAR data type) was having 2 extra non-printable characters.

    The same non-printable characters were missing in the primary database.

    To resolve, we executed a ALTER table on primary database and increased CHAR(1) to CHAR(3) and inserted data in COUNTER table again.

    The GG error repeated, but this time on primary database;

    ERROR OGG-01163 Bad column length (7) specified for column PHOTOCAPTURED in table DPS.COUNTER, maximum allowable length is 5"


    In the primary database,
    We have other tables in the same (and other schemas) which have column data type as CHAR(1)

    GG replication works fine for these columns and we have not faced the above issue.

    (only difference seems to be:: the columns were added using the CREATE table command and not ALTER table)

    for now, the only work around seems to be
    whenever the error occurs,
    manually execute ALTER table and increase the column size on the affected database.

    The DPS.COUNTER table is a very infrequently updated table so the work around suffices.

    Any clues to debug further and find the root cuase, will be appreciated.

    thank you

    12 years ago

    From where does an application (EAR) deployed in Jboss 5.1 GA, load properties from, during start-up.

    We curently maintain properties files used by our applcation (EAR) in the \conf folder inside the appropriate server folder, inside Jboss

    Typically this is d:\jboss5.1\server\default\conf
    The EAR/WAR is deployed in d:\jboss5.1\server\default\deploy

    The jboss auto-deployer expands it in d:\jboss5.1\server\default\tmp

    Can the properties file be maintained as part of the EAR or WAR, instead of in a separate folder ?

    (our application is a j2ee application, using Struts, Spring and Hibernate .... and the database connection pool provided by jboss)

    thank you.



    Folder structure of a jboss installation

    12 years ago