*BSD News Article 5830


Return to BSD News archive

Newsgroups: comp.unix.bsd
Path: sserve!manuel!munnari.oz.au!sgiblab!swrinde!cs.utexas.edu!torn!maccs!mcshub!csx.cciw.ca!u009
From: u009@csx.cciw.ca (G. Stewart Beal)
Subject: Re: BEWARE:  stupid placement of swap space...
Organization: Canada Centre for Inland Waters
Date: Wed, 30 Sep 1992 13:41:21 GMT
Message-ID: <1992Sep30.134121.23744@csx.cciw.ca>
Keywords: (or:  Don't be stupid like me :-)
References: <47897@shamash.cdc.com>
Lines: 27

In article <47897@shamash.cdc.com> pbd@runyon.cim.cdc.com (Paul Dokas) writes:
>Just a word of caution to anyone who installs swap space on extra
>disk drives.  I learned this the hard way and don't want anyone
>to be as stupid as I am ;-)
>
>I borrowed a 1.0GB SCSI drive so that I could build X11R5 and have some
>
>...... deleted
>
>Everything is now fine.  By the way, the reason for not putting swap
>first was given to me by a gentleman from Germany (sorry I've forgotten
>your name :-(  It turns out that the first few sectors of a SCSI drive
>are used for partition information.  A filesystem will not use the first
>few sectors of it's partition (knowing about the reserved blocks).  But
>a swap space will eventually overwrite the partition information,  thus
>causing strange things to happen.
>
Does this mean even if the swap partition is specified to start at cylinder
1 or 2? Does the swap creep, or just want to use block 0 of the partition
starting at cyl 0? If so, it sounds like an "off by one" boundary bug.
 
Regards, Stu Beal, VE3MWM, U009@CS.CCIW.CA,
National Water Research Institute, Burlington, Ontario, Canada.
 
"We'd made it through yet another nuclear winter and
 the lawn had just trapped and eaten it's first robin." - Kyle J. Spiller