Up Prev Next
From: Stephen dot Sprague at morganstanley dot com
Subject: Re: blk_rowxfer and leakage
Date: Sep 30 2002 3:10PM
Thanks for the assist and let me know if I can help. Feel free to hack
up the script I included in the original mail. I tried to make it easy
to plug in connection data.
On 30 Sep 2002 @ 10:44am, an entity claiming to be Michael Peppler scribbled:
mpeppl :On Sat, 2002-09-28 at 14:01, Stephen.Sprague@morganstanley.com wrote:
mpeppl :> It clearly shows when the blk_rowxfer is commented out the mem size tops
mpeppl :> out in my case at 43MB. When reinstated it just keeps ballooning until I
mpeppl :> kill it. To me, that pretty much isolates it to that routine. My guess
mpeppl :> something inside it is not deallocating something it should.
mpeppl :I took a look at the C code and I can't see any obvious places where
mpeppl :memory gets allocated. It *could* be a problem with references in your
mpeppl :perl script, causing data to not get freed properly.
mpeppl :Alternatively it's a problem with memory fragmentation.
mpeppl :I'll try to set up an example script and see if I can reproduce it.