I keep getting this OutOfMemory Exception. Sometimes it is consistent, it occurs everytime I try to fetch data from the backend. Sometimes it is inconsistent, it does not happen always.

As a result of this error message, no records are available at the frontend. This affects production significantly.

Kindly advice me how to combat this problem.

Here are some solutions I have tried and was unsuccessful:

* Transferring the records in batches. Instead of sending 100 records in one shot, I send 10 records at a time (or) instead of sending around 50 columns I fetch 10 columns at a time (with all the records)

* Multithreading. I am working on it currently.

Sometimes this problem occurs, even if the number of records is very less (less than 50). It is not clear what creates this problem.

It would stand to reason that your hardware is lacking memory. What kind of data are you retrieving? How large are the records? I have retrieved hundreds of thousands of records in past projects and not encountered this kind of problem.

Multithreading is unlikely to solve this. Splitting a process up into multi threads isn't going to reduce the memory required by the process as a whole. It just reduces the memory demand of each single thread. In fact, it would probably INCREASE the memory demands because you are introducing thread management to the process.

-Peter

Dear Peter,

I agree. In fact I have a different screen that fetches around 7000 records and works fine. I do not get Out of Memory exception.

I do not know what causes this exception. It is unlikely to be hardware problem, because the other screen that performs well, retrieves so many records and yet doesn't complain.

I just do not have any lead towards a solution. Please give me ideas/suggestions.

Please specify your backend and the mathod you fatch your records. A piece of code causing erorr will be useful too.

-Ash

Positiveness is in my blood.. I'm O+.

I provide some more information regarding the problem that could be useful

* I have a middle tier between the front end and the backend. There is no problem when the records move from the back end to the middle tier. The problem seems to be when the records come to the front end.

* Oracle backend.

* I invoke the middle tier through remoting. The middle tier methods in turn invoke Oracle.

* The records move from the middle tier to front end in compressed format. Actually the return value is a dataset, that is compressed using DeflateStream, and in the front end it is decompressed. The compressed format is a byte array. The retrieved byte array is restored to dataset in the front end.