[omniORB] ominORB memory Leaks

[Duncan Grisby]

On Friday 25 August, Mike Olson wrote:

>   I have a server running omniORB 3.01 and omniORBPy 1.1 on RedHat 6.2
> that is leaking memory.  I've spent a lot of time looking into it and I
> think the leak is coming from omniORB.  However, I am currently unable
[...]

I just did a quick experiment with a server which just returns a 2000
character string, being called from 10 concurrent clients, and it
doesn't leak at all. That doesn't mean that you haven't found a leak,
but it's not something totally elementary.

One thing to check -- omniORB starts a new thread for each client.
You aren't using a tool which adds together the total size of each
thread are you?  gtop reports a size of 49MB for my little server,
even though it's only really using 3MB, for example. That doesn't
explain the differences you see with the size of the string, though.

Unless you can come up with an example which shows the problem, I
don't really know what to suggest. You could try attaching to the
server process with gdb and looking at the heap. You might be able to
identify the contents enough to have some idea of what's leaking.

Sorry I can't be of more help,

Duncan.

-- 
 -- Duncan Grisby  \  Research Engineer  --
  -- AT&T Laboratories Cambridge          --
   -- http://www.uk.research.att.com/~dpg1 --