RE: [flalug] amount of allocated memory that's swapped out

From: Larry Brown (larry.brown@dimensionnetworks.com)
Date: Tue Mar 16 2004 - 20:59:29 EST


Sounds like a difficult problem to solve since at any time win2k could load
portions of what the vm machine believes to be hd into physical memory. The
only given is that the portion of memory allocated by design in vmware to
the vm will be physical memory. The little script I was writing about the
other day can tell you how much a particular process is using and you could
modify it to add which ever element you want of the status page.

By the way, that page you sent me the link for stated that the vmsize is
supposed to be the sum of four of the other figures and it is not. I still
haven't found out why. I did notice that several of the processes that
where children of the same parent had the exact same difference. IE:
several httpd children all had 6228 kB as the difference. Most of those
children had differing vmsizes but the difference between the vmsize and the
other 4 figures where all the same. Now there were other children of httpd
that had a different value as the difference and I am thinking that maybe
they were spawned by a different parent and that that difference indicates
the amount of shared memory. Still looking.

Larry

-----Original Message-----
From: flalug@nks.net [mailto:flalug@nks.net]On Behalf Of Eben King
Sent: Tuesday, March 16, 2004 1:56 PM
To: Florida Linux Users' Group
Subject: [flalug] amount of allocated memory that's swapped out

How can I find how much memory allocated to a process is swapped out? Or
is the only way to kill the process and see how much swap usage goes down?
I run W2K inside of VMware, and VMware (according to top) is currently
using 162 MB (out of 768 MB). Often, tasks take a long time to start in
VMware (raise a menu, start a process, etc.), but once started, proceed at
normal speed. I think that's because VMware is partially swapped out,
which in turn is because I don't have enough RAM. I suppose I could infer
which "memory" accesses are really swap reads by the time taken for them.

When W2K accesses part of its 600 MB of real (it thinks) RAM, part of that
RAM has been swapped out by the host OS. Had it been swapped out by W2K,
presumably W2K could've optimized things so it acted better.

Should I lower VMware's memory allocation to match what it really gets in
the mean time?

--
-eben    ebQenW1@EtaRmpTabYayU.rIr.OcoPm    home.tampabay.rr.com/hactar

Every normal man must be tempted at times to spit upon his hands, hoist the black flag, and begin slitting throats. -- H.L. Mencken



This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 20:16:31 EDT