Re: fetch size too big causes hang

From: jgar the jorrible <>
Date: Mon, 30 Mar 2009 10:30:21 -0700 (PDT)
Message-ID: <>

On Mar 29, 1:01 pm, Jim <> wrote:
> Latest update: I installed a clean new client on a new
> machine which had never had Oracle client on it before.  The problem
> persists.
> One thing I didn't mention before: I am going through a VPN to get to
> the server.
> It seems to me since I see the same thing on two different client
> machines, that it must be a problem somewhere between me and the
> server, maybe in the VPN.  Could be I need to go to some other forum
> to continue on this problem.

Try googling:
VPM memory limitations

I think you are seeing something like (from a Cisco security advisory found with the google):

"Information Leakage via Windows Paging File This limitation is the inability to prevent data from leaking to the Windows virtual memory file, which is commonly referred to as the paging file and is called pagefile.sys. This file is normally located in the root directory of the hard drive where Windows is installed, but it can also be a group of files stored in various locations, across hard disks and partitions.

The paging file is used to store the contents of physical memory that have been swapped out by the Windows kernel when there is pressure to provide additional physical memory for some application, and no physical memory is available. In this case, the Windows kernel swaps out memory used by idle processes to the paging file and gives the deallocated  memory to the application that is asking for more memory.

As a consequence of how the Windows virtual memory subsystem operates, the physical memory contents used by any application, including those running in a Secure Desktop, may end up in the paging file. The Windows paging file stores "paged out" physical memory contents without encryption, and therefore information "paged out" by the operating system may be recovered using data forensic tools. Because of this process, CSD may not be able to remove from the system all data produced and accessed during the SSL VPN session after the VPN session terminates.

This item is not a CSD product defect. It is, rather, a CSD product limitation resulting from how the Microsoft Windows operating system interacts with applications. "

In other words, I'm hypothesizing when you increase the amount of memory the client needs, at the inflection point where you notice the problem something important (like maybe packet buffering) in the VPN is getting swapped out, resulting in the "hang." Some of the other google results appear to suggest this with more detail.

Of course, I have no idea what I'm talking about, I likely wouldn't have responded to a thread that mentioned VPN. I've had it foisted upon me, and have had to do way too much extra work fixing production data as the result of problems like this. And this explains some of the problems I've seen - finger pointing between VPN and OS vendors is not acceptable.


-- is bogus.
Received on Mon Mar 30 2009 - 12:30:21 CDT

Original text of this message