On Tuesday 31 December 2002 02:23 pm, Ripunjay Bararia (ILUG-MUM) wrote:
- LUG meet on 12 Jan. 2003 @ VJTI
Happy new year to all....
I have almost the same problem,
more in depth explanation from all the great masters will be appreciated
What happens in some memory management algorithms is that even if the dirty buffers (bufferes which are modified in the RAM) are written to the HDD the RAM is not freed. So if a process again wants the same page (memory segment) then it need not be brought from the HDD to the RAM. So your RAM usage which is shown as 70% (assume) then not all that memory is actually used by the kernel at that time. How ever if you stop some memory hogging processes like httpd, mysqld, named, squid then the buffers will be forced to be written to the disk and the memory will definately be vacated. But if these processes are running (as it seems to be the case with you - server process usually never die) then there is no use of vacating the RAM. To just verify that swap space is working (dont you trust Linus) what you probably can do is that start approx 20 instances of mozilla (approx 15*20 = 300MB RAM) or a VMware session and ask it to use 80% of the existing RAm, then hope that the RAM is completely used up. Other thing to verify this concept is to remove some RAM say 128 MB and still the RAM usage should be around 90% so that will proove that the RAM utilization (70% of 256MB = 180 MB) wasnt actually required and the kernel could live with less RAM and give the same performance. I hope you understand this crappy little explaination.
thanks
Ripunjay Bararia
-----Original Message----- From: linuxers-admin@mm.ilug-bom.org.in [mailto:linuxers-admin@mm.ilug-bom.org.in]On Behalf Of Rajesh Deo Sent: Monday, December 30, 2002 10:42 PM To: linuxers@mm.ilug-bom.org.in Subject: Re: [ILUG-BOM] [ILUG-BOM]Linux swap not working.
Hi,
On Mon, 30 Dec 2002, The Linuxer wrote: [snip]
http://people.debian.org/~psg/ddg/node81.html (more info on swap)
[snip]
Best regards,
Rajesh