home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.unix.aix
- Path: sparky!uunet!cs.utexas.edu!milano!cactus.org!chukran
- From: chukran@cactus.org (Rudy E. Chukran)
- Subject: Re: Virtual Memory curiosity
- Message-ID: <1992Nov19.063242.6978@cactus.org>
- Keywords: VM overflow response
- Organization: Capital Area Central Texas UNIX Society, Austin, Tx
- References: <bmarlowe.721422470@ka> <BxM5C5.Jyw@csn.org>
- Distribution: usa
- Date: Thu, 19 Nov 1992 06:32:42 GMT
- Lines: 26
-
- In article <BxM5C5.Jyw@csn.org> losburn@teal.csn.org (Lynn Osburn) writes:
- >>
- >>I've heard rumor that AIX will pick processes to kill off to avoid
- >>running out of virtual memory? Can anyone verify this and if this is
- >>true a) is there a way to disable this "feature" and b) what idiot
- >>decided that randomly killing processes was a good idea??!!
- >
- >True in v3.1.x
- >
- >False in v3.2.x
- >In 3.2, he just picks the larger process(es) and puts them to sleep (and
- >swaps them) for a while until things improve. A much better solution to
- >the problem, eh?
-
- Not quite. You confuse real memory shortage with virtual memory shortage.
- When virtual runs out, ie paging space runs out, then processes are
- signalled. The choice of which processes are killed first changed from
- 3.1 to 3.2. There is no option to swap anybody out because there is
- no place to swap them to.
-
- The new algorithm to which you refer is pertinent when real memory is
- overcommitted. This is called the thrashing detection algorithm.
- When paging reaches a certain threshold, the scheduler will suspend the
- suspected culprits (not signal them) and they will be paged out as a
- result.
-
-