[X4U] how to view "pageouts?"

[David Ledger]

>From: Ed Gould <edgould1948 at comcast.net>
>On Jun 19, 2008, at 11:47 AM, David Ledger wrote:
>>  ---------------------
>>  SNIP----------------------------------------------------------
>>  Other than memory mapped file I/O I'd be interested to know what
>>  these other reasons are.
>The one or two that come to the top of my mind (there are others of
>course) are
>1. The application tells the OS that the page will not be
>used in quite a long time (perhaps an error routine that is never(?)
>to be invoked)
>2. The OS sees a wait issued by the application and
>due to other issues (like cpu busy, etc) decided to swap the entire
>application out. Some of the types of wait are either I/O and or
>waiting on input from the user.

If the cpu is busy but memory is not tight, swapping out an 
application would only make things worse - more for the cpu to do.

>If the system is overloaded say by
>cpu usage then the system looks at the the applications that are not
>doing anything and frees up memory by swapping those applications out
>or another reason is that the OS keeps track of a a "routine"  and if
>its no longer in use it it moves it to the ready "que" to be paged
>out que and so long as the use count does not go up by the time its
>ready to be paged it is paged out or depending on how the OS is
>written  maybe just the page is freed and when the routine is need
>again it is paged in from the swap file, again it depends on how
>sophisticated the OS is.

What does this mean? Ready queue has nothing to do with pageouts.

>It maybe just because the application says
>I am stopped and can go no further until an external event occurs
>like a time expiring at say 2 hours later(ICAL might be an example
>here).

Why should this cause a pageout if there's plenty of memory?

>The MAC internals on scheduling is not externally published so
>there may be others as well, its difficult to say what was
>architected into the OS do I can only guess based on how other
>systems that *ARE* externalized.
>
>I posted a lengthy pageout response yesterday and it was held for
>length restrictions. Maybe that will answer the question as well.

It's not through yet. I'd still be interested to know what version of 
Unix this is.

David


-- 
David Ledger - Freelance Unix Sysadmin in the UK.
HP-UX specialist of hpUG technical user group (www.hpug.org.uk)
david.ledger at ivdcs.co.uk
www.ivdcs.co.uk