Date: Tue, 17 Mar 1998 13:45:32 -0600 From: Karl Denninger <karl@mcs.net> To: Paulo Cesar Pereira de Andrade <paulo@fiscodata-pr.netpar.com.br> Cc: "John S. Dyson" <dyson@FreeBSD.ORG>, michaelh@cet.co.jp, current@FreeBSD.ORG Subject: Re: WARNING WILL ROBINSON! Risk of severe filesystem damage suspected Message-ID: <19980317134532.55817@mcs.net> In-Reply-To: <Pine.BSF.3.96.980317160648.24088A-100000@fiscodata-pr.netpar.com.br>; from Paulo Cesar Pereira de Andrade on Tue, Mar 17, 1998 at 04:41:39PM %2B0000 References: <199803170731.CAA11588@dyson.iquest.net> <Pine.BSF.3.96.980317160648.24088A-100000@fiscodata-pr.netpar.com.br>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Mar 17, 1998 at 04:41:39PM +0000, Paulo Cesar Pereira de Andrade wrote: > On Tue, 17 Mar 1998, John S. Dyson wrote: > > > Karl Denninger said: > > > > > > Yes, and now I've done it twice more while trying to isolate this thing. > > > > > > The corruption comes back within minutes of starting a "make world". > > > > > Whomever is trying to work with late Sun, early Mon -current should > > immediately stop, fsck their drives (if there is anything left), and > > go back to a Sat kernel. If you are very aggressive, try the -current > > as of now (as committed by me at around 01:00 EST on Tue.) > > I have cvsupped early saturday (about 02:00am Brazil), and while > building world, received a error building gcc: > <some-file>.o: malformed input file (not rel or archive) > > Looking the files, (one of them was > /usr/obj/usr/src/gnu/usr.bin/cc/cc1plus/except.o) I saw it was fragments > of c source code. This was after a crash while making world. Restarted a > clean make world (/var/tmp as 8M MFS, and -j3) and it build with no > errors. > > I got corrupted kernel .o files when compiling a library and the kernel > at the same time. When compiling the kernel alone ld does not complain. > > The two times I thought it was operator error, because restarting did > not repeat the problem. Folks, please, please listen closely to John on this. I will say it LOUDLY as well. -CURRENT, as of AT LEAST Saturday, is critically broken. If you run it on a system which has volatile data, you risk unrecoverable and permanent filesystem damage. PLEASE DON'T, at least for now, unless you're doing so in a controlled environment or don't mind a FULL reload! There ARE conditions under which it appears to be "safe" to use. HOWEVER, I'm not about to elucidate those under which it works here and doesn't destroy data, because (1) I don't understand why I get bit on some machines and not others, and (2) A situation in which I *THINK* is safe might turn out NOT to be safe. -- -- Karl Denninger (karl@MCS.Net)| MCSNet - Serving Chicagoland and Wisconsin http://www.mcs.net/ | T1's from $600 monthly / All Lines K56Flex/DOV | NEW! Corporate ISDN Prices dropped by up to 50%! Voice: [+1 312 803-MCS1 x219]| EXCLUSIVE NEW FEATURE ON ALL PERSONAL ACCOUNTS Fax: [+1 312 803-4929] | *SPAMBLOCK* Technology now included at no cost To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980317134532.55817>