atom feed20 messages in org.freebsd.freebsd-stableRe: More problems with new slice code
FromSent OnAttachments
Burton SampleyMar 13, 1998 12:39 pm 
Burton SampleyMar 14, 1998 7:54 am 
Robert WatsonMar 14, 1998 10:06 am 
Michael V. HardingMar 14, 1998 11:04 am 
Robert WatsonMar 14, 1998 11:12 am 
Burton SampleyMar 14, 1998 12:11 pm 
Robert WatsonMar 14, 1998 2:53 pm 
Jordan K. HubbardMar 14, 1998 4:10 pm 
Jordan K. HubbardMar 14, 1998 4:11 pm 
Robert WatsonMar 14, 1998 4:14 pm 
Jordan K. HubbardMar 14, 1998 4:33 pm 
Greg ShenautMar 14, 1998 5:27 pm 
StuddedMar 14, 1998 7:19 pm.comproot
Mike SmithMar 14, 1998 8:09 pm 
Chad R. LarsonMar 15, 1998 12:16 am 
Cy Schubert - ITSD Open Systems GroupMar 15, 1998 10:46 am 
Cy Schubert - ITSD Open Systems GroupMar 15, 1998 10:57 am 
Mike SmithMar 15, 1998 2:19 pm 
Robert WatsonMar 15, 1998 2:27 pm 
Cy Schubert - ITSD Open Systems GroupMar 16, 1998 8:01 am 
Subject:Re: More problems with new slice code
From:Robert Watson (rob@cyrus.watson.org)
Date:Mar 14, 1998 2:53:58 pm
List:org.freebsd.freebsd-stable

On Sat, 14 Mar 1998, Burton Sampley wrote:

To the best of my knowledge (if I'm wrong, will someone please correct me), the entries are not deleted, they just become invalid, ie, they *act* like they point to /dev/null or somewhere in outerspace. I have been bitten by this in the past and have been quickly educated in the use of the fixit floppy/CDROM, which is actually a really cool tool to have! The pair have saved my butt a few times. It seems kinda odd when I explicitly give "/dev/MAKEDEV sd0s2a" (and the rest of my file systems individually) that it does what it needs to do to allow me to use that slice successfully, but "/dev/MAKEDEV all" makes the same slice no longer function.

That is what puzzled me, actually. The device names were really gone from /dev -- /dev/wd0a which had been happily there just sort of disappeared after I did a MAKEDEV of the new devices. I haven't looked enough at the MAKEDEV script to know its depths though.

I believe this has been corrected in -current, but I probably have my facts wrong.

I think all of us a one point in time have been a member of the head-scratching club. That's part of the fun of using FreeBSD. I just have a bad habit of learning at the worst possible time, like blowing up my system without a backup and all of my code for the quarter being on the unaccessible hard drive the week before finals. :-)

Oh, I would certainly rather deal with just about any head-scratchy moments in FreeBSD than run anything else. We have over 8 machines running FreeBSD of various forms in our apartment, and I do all development for work under FreeBSD. :) On the other hand, while I was bitten by this bug, and knew how to deal with it, I pity a more novice user who gets in that situation. There are, of course, risks to running -STABLE, but as it is currently a release candidate, we should try to minimize those risks. Having poked about at the code a bit where the changes were, I accept that they were simplifying and helpful; it is unfortunate, however, that the correction they provide to the code has some side-effects. sysinstall under 2.2.6 should definitely know how to deal with this config and auto-magically do it (or request confirmation first, but notify the user :).

Robert N Watson

Carnegie Mellon University http://www.cmu.edu/ SafePort Network Services http://www.safeport.com/ rob@fledge.watson.org http://www.watson.org/~robert/

To Unsubscribe: send mail to majo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message