Re: excessive swap-in time

On Mon, 17 Mar 2008 13:55:23 +0000 (UTC) Michel Talon <talon@xxxxxxxxxxxxxxxx> wrote:

| phil-news-nospam@xxxxxxxx wrote:
|> | You can think of obvious ways to tweak this to make this one case
|> | perform better, but you can think of cases that are made worse by
|> | those same tweaks.
|>
|> Something has to be forced back out to swap in a process that becomes
|> active again. The question is which pages to select. Those of processes
|> that are recently inactive would be what ultimately would be swapped out.
|> Might as well go straight to them directly, rather than the newly active
|> process fighting against its own pages.
|>
|
| Nothing new in this discussion. The Linux people have always been of the
| advice that, "if there is not enough memory, buy some more", while it is

I wish it were that easy. Some companies (e.g. Intel) like to make chipsets
that artificially limit memory size so they can force you to update the whole
system just to get more memory.

It is easier today, aside from new issues indroduced by the new systems.


| well known and demonstrated that a system can very well function fine
| and dandy with a constant stream of paging to swap. Of course, this
| depends on proper selection of the pages to swap, and one of the main
| point in Linux kerner 2.6 was the inclusion of a reasonable VM system
| doing that correctly. A small discussion of the ideas around such
| choices can be found here:
| http://www.freebsd.org/doc/en_US.ISO8859-1/articles/vm-design/freeing-pages.html

Actually, my target is to have enough RAM to do without any swap space at all.
I've heard arguments why I should not do this, but so far I have not heard any
that make sense. As soon I can solve the video issues on the new machine that
has no swap space with its 8 GB of RAM, I will see how well it performs with
the biggest of obese beasts: firefox after it has loaded 100 websites

--
|---------------------------------------/----------------------------------|
| Phil Howard KA9WGN (ka9wgn.ham.org) / Do not send to the address below |
| first name lower case at ipal.net / spamtrap-2008-03-17-1439@xxxxxxxx |
|------------------------------------/-------------------------------------|
.

Relevant Pages

  • Re: WARNING: Sorry, no swap space to grow stack for pid 4456
    ... > What I meant was on another machine, swap -l shows ... Virtual memory is made up of three "resources" that are put together ... if RAM isn't available, the data is written to the file instead of swap. ... state of affairs might be to have a little swap space used every now ...
    (comp.unix.solaris)
  • [HPADM] RE: -SUMMARY- Clarification on swapinfo/glance output
    ... Memory allocation under HP-UX can be a little odd, ... the system needs to swap out a portion of that program, ... configured swap space, you will no longer be able to run more programs (they ... The pseudo swap is enabled by default by ...
    (HP-UX-Admin)
  • Re: Slow FC4, perhaps swap is to blame
    ... kind of memory leak that builds up in the swap space over time. ... I did manage to find a command which I somehow got to work at least ...
    (Fedora)
  • Summary swap rules.
    ... So it looks as if swap is mostly not used anymore and can be kept pretty ... Tim Bradshaw stated what I privately felt- that 1:1 swap to memory is ... so it's really only around for memory leaks and core dumps. ... think) which are written to swap space. ...
    (SunManagers)
  • Re: swap space size
    ... I believe this was just a rule of thumb when memory was so expensive. ... And if it stops swapping, ... "two times 'so much'" swap space have? ...
    (Debian-User)