what process is eating my swap?

[Stijn Hoop]

On Fri, May 09, 2003 at 06:14:41PM +0200, Alexander Haderer wrote:
> At 16:32 09.05.2003 +0200, Stijn Hoop wrote:
> >On Fri, May 09, 2003 at 04:18:17PM +0200, Alexander Haderer wrote:
> >> What does "top -o size" show? Look at column size. What is on top? How 
> >many
> >> processes?
> >
> >Here's a random sample. Of course mount_mfs is using 256M, that's to be
> >expected.
> >
> >last pid: 39401;  load averages:  0.15,  0.14,  0.09    up 
> >1+03:32:39  16:31:15
> >117 processes: 1 running, 116 sleeping
> >CPU states:  0.0% user,  0.0% nice,  1.2% system,  1.2% interrupt, 97.7% 
> >idle
> >Mem: 146M Active, 445M Inact, 124M Wired, 32M Cache, 86M Buf, 4984K Free
> >Swap: 1536M Total, 1536M Used, 4K Free, 99% Inuse
> >
> >  PID USERNAME PRI NICE  SIZE    RES STATE    TIME   WCPU    CPU COMMAND
> >   24 root      10   0   257M 16512K mfsidl   0:01  0.00%  0.00% mount_mfs
> >  262 mysql      2   0 32828K 16504K poll    48:17  1.17%  1.17% mysqld
> >  263 root       2   0 18356K 17644K select   0:03  0.00%  0.00% perl
> >38861 www       18   0 10952K  7336K lockf    0:01  0.00%  0.00% httpd
> >34040 www        2   0 10900K  7340K poll     0:08  0.00%  0.00% httpd
> >37664 www       18   0 10876K  7304K lockf    0:03  0.00%  0.00% httpd
> >38249 www       18   0 10860K  7296K lockf    0:02  0.00%  0.00% httpd
> >38203 www        2   0 10852K  7264K poll     0:02  0.00%  0.00% httpd
> >37490 www        2   0 10848K  7272K poll     0:05  0.00%  0.00% httpd
> >39005 www       18   0 10840K  7224K lockf    0:00  0.00%  0.00% httpd
> >38001 www       18   0 10836K  7256K lockf    0:02  0.00%  0.00% httpd
> >37942 www       18   0 10828K  7240K lockf    0:03  0.00%  0.00% httpd
> >38864 www       18   0 10812K  7220K lockf    0:01  0.00%  0.00% httpd
> >38862 www       18   0 10680K  7104K lockf    0:01  0.00%  0.00% httpd
> >
> >I see nothing wrong in the output so far.
> 
> Well, in this case you have to much (107) memery consuming processes 
> running. httpd is 10MB, for example. How many httpd processes are active?

Good question.

[stijn at sandcat] <~> ps auxww | grep httpd | wc -l
      13

so that's 130 MB of httpd, 256 MB of mfs and then other processes
averaging around 5 MB. But as far as I can see, the top 'SIZE' column
is the same as the 'ps auxww' VSZ column, right?

Because, as I posted in my other email, if I total that column, I have
processes taking about 700 MB of virtual memory (and the mfs is included).

Since this machine has 768 MB of RAM and 1536 MB swap, it could not be
full :(

I really wonder where this memory is allocated.

Thanks for thinking with me so far, both to you and Anthony.

--Stijn

-- 
The rain it raineth on the just
	And also on the unjust fella,
But chiefly on the just, because
	The unjust steals the just's umbrella.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 187 bytes
Desc: not available
Url : http://lists.freebsd.org/pipermail/freebsd-questions/attachments/20030509/3c55ebd6/attachment.bin