*BSD News Article 8777


Return to BSD News archive

Path: sserve!manuel.anu.edu.au!munnari.oz.au!spool.mu.edu!uwm.edu!ogicse!news.u.washington.edu!ns1.nodak.edu!plains.NoDak.edu!tinguely
From: tinguely@plains.NoDak.edu (Mark Tinguely)
Newsgroups: comp.unix.bsd
Subject: Re: Occasional system hangs
Message-ID: <Bz495r.Byn@ns1.nodak.edu>
Date: 11 Dec 92 22:29:51 GMT
Article-I.D.: ns1.Bz495r.Byn
References: <andrewh.724059111@bruce.cs.monash.edu.au> <1gaclpINN84v@hrd769.brooks.af.mil> <1992Dec11.210430.17335@coe.montana.edu>
Sender: usenet@ns1.nodak.edu (News login)
Organization: North Dakota State University
Lines: 24
Nntp-Posting-Host: plains.nodak.edu


 Did all of you with 16+ Megs of RAM apply the patch to restrict the size of
 bufpages (Terry Lambert's patchkit patch #2. The calculations for
 bufpages were being done interchangedly in pages and bytes, the fix
 did not correct the calculation problem, but limited the error of the
 calculation; without this correction, too much RAM is lost to bufpages).

 Besides, there are small memory leaks in the kernel; Programs that fork a lot,
 cause the system to hang (make and apparently news); Big programs like X
 also make hanging very likely; Even a couple simple looping mallocing
 processes can hang a system.

 I agree that a kernel fix is better than a symptom avoidance technique.
 The kernel change is massive. The VM allows over extention of the swap
 backstore (and this can lead to the system freeze -- all of the swap
 blocks are used or fragmented to a point a certain size block cannot be
 allocated, new request for VM -- page a new part of a program, use of an already
 malloced **but memory in swap backstore is not reserved until paged out**
 memory, forking (only small part is new on a fork), and exec-ing a new program
 will look for physical memory, but cannot get it until the dirty pages
 are put on swap backstore, but backstore is full ... system will work the
 drive a little but will eventually freeze.

--mark.