Rajneesh Kumar Rajput

just a guess, did you server has temp directory already created in its root?
if not please create temp folder manually, and restart server.

if it does not have, it will not create it automatically but will not upload big files.

try setting this variable too...
<param-name>uploadThresholdSize</param-name> <param-value>100k</param-value>

This is from my faces 1.2.6 if you are using old version, variable name will be different,

Images are contains directly binary data, so the answer is yes.

Well I would say, over here closing/not closing stream could not be cause of out of Memory. most of output stream source code does not do anything while you close stream. it is blank implemented in sub classes.


And look for your example, that depends on the scope where image is being created/loaded.

If it is per user (per request) a different instance, will reach max usage of memory 200*10 = 2000 MB (~2GB).
is this right?

either look for current memory allocated to your JVM?

further analysis you need a tool like JProve etc or try to check jheap and jstate map to see which and how many objects are lying in the memory with this operation.

Are you using some standard pagination component or want to implement your known text fields and buttons as pagination?

1) either use tomhawk table, provides automatic pagination.

2) Use rows and first attribute of the data table to control no. of records in the table.

well if you have posted this, you must have seen OutOfMemory as well while running the code?

The code that you are showing should not cause out Of Memory at all, but that different API that you are using to create an image of 10 MB will be the main suspect if you encountered an Out Of Memory.

Thanks to all of you in participating and providing your valuable thoughts.
really very helpful.

yes, the explanation is further clarified other doubts had in mind.

Thapliyal Akshat wrote:
It's the mediator between your build tools and the deployment environment. Either you do it manually or through CC.

If I just see this point, either using Maven / Ant some tasks definition can be written which will do the automatic deployment work also after checking out code and doing compilation on the build.
so, would you clarify this like how CC is better than maven/Ant in this specific area?

except above, like David told commits merges and build failure reporting, I have found some good points on top is that in Cruise Control you get a web control to see, compare and find out various commits differences and can easily review the changes done and let CC to control the stability of the builds also.

further inputs and suggestions are appreciated, and looking forward to take a good suggestion based on this conversation.

Thanks David, it gave me a broader idea.

One more quick question, if I get it correct?

CI server (Cruise Control), will itself behave as some version control like CVS, SVN
or CI server still needs a separate version control to be set up for the project and will just provide merge commits feature to derive report for build failure etc?

Ant/Maven responsibilities, I understand.

what are the other tasks suppose to be performed/expected by setting up continuous integration server?

Can someone please help me to find the key differences, so that I can choose either Maven or Cruise Control as the standard tool for managing the project?

We do want to standardize the project build management, test case execution, documentation and release deployment automation all together.

Hold the records, of the query, possible in the List by defining a variable in the page bean.
Now, in the getter of this variable just add a null check to verify that call the query method if list is not initialized.

private List<O> records;

public List<O> getRecords (){
if(records == null){
records = /call the query method.
}
}

Actually, my faces table calls value attribute twice in Apply Request and Update Model phase. The above will solve your problem.
Also, in order to refresh records you need to again set records = null in the actions of buttons so that new data will be picked after the action.

Displaying results are totaly depends on this variables...
value="#{myprojectBeanCommunicationBackingBean.localDataModelCategoryLOV}"

So, let user to type in the search box and you can use the onblur event to call a java script which will submit your form (or a hiddnen link action) and just simply set this above value to the new required list of results based on new value what is found in the filter. (Ofcourse this you can do with AJAX also, this sumbit a page on typing in filter)

Could you let me know, what exactly are you thinking and what problem are you facing to filter the new results?

Hi Anil,
Changing scope to the session, was not the fix it was just an alternate.
So, you have assumed that for request scope bean a selected row data cannot be fetched.

Hi Pallam,
This seems to me java script behaviour which is already known in JSF radio button. Please use onClick intead of onChange.

see...
http://wiki.apache.org/myfaces/Radio_Buttons_and_Browsers