Hi Kurt, I sent a message just prior to this that contained a forward from the linux kernel mailing list that explains this further, but:
The method you are using to check changed with 2.6.10 and may readily account for the changes you are seeing in your figures. In the end, you should not be using '/proc/<pid>/statm' shared column derived figures to determine how shared apache is with it's children's via COW pages. That figure in 2.4 was polluted with other data, and in 2.6 it has changed. In 2.6.10 and later, and for the "foreseeable" future it will NOT contain private pages shared with forked children. This leaves us at a loose end though. Best, On Fri, 2005-02-11 at 17:34 -0500, Kurt Hansen wrote: > Perrin Harkins wrote: > > > > >I wonder if it has something to do with SMP. It would be good to try it > >on a kernel that is not compiled for SMP. It might be worth compiling > >your own kernel to test it. > > > >Running a non-threaded perl is always a good idea if you are not using > >threads. Performance will be better too.mercy > > > > > I re-booted into the 2.6.10-1.12_FC2 (non-SMP) kernel. No difference. > > Then, I tried going back to an earlier kernel, 2.6.8-1.521smp to be > exact. In that case, the shared memory jumped from 4M to 15M! That still > leaves 55M unshared, but the kernel seems to be part of the equation. > Thanks for the clue, Perrin! I'm not sure if there is a problem with > 2.6.10 or if 2.6.8 was the kernel on the machine when I originally > compiled perl. > > Has anyone else seen problems with the 2.6.10 kernel? > > I guess my next step is to re-compile perl and see where that leads me. > > Take care, > > Kurt Hansen > [EMAIL PROTECTED] -- Richard F. Rebel <[EMAIL PROTECTED]> WhenU.com
signature.asc
Description: This is a digitally signed message part
