Newsgroups: comp.sys.sgi Path: utzoo!censor!geac!maccs!apcah2!root From: root@apcah2.Eng.McMaster.CA (Apcah2 Super-User) Subject: Re: slow boot-time fsck -y -D, invoked by mount(!) Message-ID: <1990Nov9.203004.473@maccs.dcss.mcmaster.ca> Reply-To: root@apcah2.eng.mcmaster.ca (Apcah2 Super-User) Organization: Dept. of Chem. Eng., McMaster University, Hamilton, Ontario References: <1990Oct28.073438.8627@utstat.uucp> <73531@sgi.sgi.com> <1990Nov1.102132.19777@maccs.dcss.mcmaster.ca> <74515@sgi.sgi.com> In article <74515@sgi.sgi.com> daveh@xtenk.asd.sgi.com (David A Higgen) writes: >> In article <73531@sgi.sgi.com> daveh@xtenk.asd.sgi.com (David A Higgen) writes: >> >I'm currently in the midst of a major overhaul of fsck which will appear >> >in the next release. I've been able to make a speed improvement of a factor >> >2 to 2.5 in the program itself, and there will be logic added to do >> >parallel fsck's during bootup. >> >> How about a -p option, instead of -y? Is there a good reason to have fsck >> run with -y always? [ ... chop chop ...] >Yes, a -p type option will be available. Unlike BSD, this will NOT be >hardwired to parallel behaviour (a rotton design decision, IMHO), but will >be an independant option. I thought the -p meant "preen", and the parallelism was invoked through careful arrangement of the fsck pass number in /etc/fstab. In any case I'm glad to see the -p option. >We will probably still ship systems with -y behaviour by default, the reason >being that for the non-UNIX-experienced, you want the system to just >COME UP, not require cryptic messing with some scary fsck-thing! > >But -p type behaviour will be available for those who want it. Sounds reasonable to me. The more choice I have the better (and the more work I have to do :-( E. John Benjamins BITNET: JOHNB@MCMASTER Computer and Information Services, Internet: johnb@sscvax.cis.mcmaster.ca ABB 131, McMaster University, Hamilton, Ontario, Canada "You can't chop down a symmetry" -- Jane Siberry