A good operating system should attempt to use all available memory. Why
have it when you don't use it? The way it uses it is to create a cache
of things seen on disk. Next time you need the same data from disk an IO
(a few milliseconds) is saved. Naturally it can throw this away in an
instant if a more pressing need (like an application program) comes
along. It makes sense to do repeat this process with all available
memory.
While I can't directly speak for Linux, I have noticed that most other
systems will slowly drift up to 80-90% of installed memory in use and
just sit there (the little bit of free is essentially for emergency
use).
That's not to say that you don't have a memory leak in the kernel that
others have suggested, it just means that there is a perfectly normal
reason for it that is also plausible. A far better measure of memory
adequacy is in paging rates, but even that is suspect when normal IO is
managed through the paging system with mmap most of the time these days.
-----Original Message-----
From: Thomas [mailto:user-97316fb2dd2a@xymon.invalid]
Sent: Wednesday, 23 May 2007 5:06 PM
To: user-ae9b8668bcde@xymon.invalid
Subject: [hobbit] OT: CentOS mem usage.
Hi all,
Totally off topic question but I guess here are at least 1 person with
the right answer.
On my CentOS server I have 2 GB memory, however hobbit shows me that the
used memory is slowly being used and never freed. How can I determine
which application has the mem leak ? Doing a lot of ps commands only
produces a list of processes with less than 1% mem usage.
TIA
/Thomas
**********************************************************************
This message is intended for the addressee named and may contain
privileged information or confidential information or both. If you
are not the intended recipient please delete it and notify the sender.
**********************************************************************