Date: Sun, 29 May 2011 17:28:08 +0200 From: rank1seeker@gmail.com To: hackers@freebsd.org Subject: Re: Active slice, only for a next boot Message-ID: <20110529.152808.781.1@DEV> In-Reply-To: <4DE139C9.2080808@freebsd.org> References: <20110527.124553.718.1@DEV> <20110527134754.GA94769@freebsd.org> <20110527.164723.750.2@DEV> <496B0C04-7777-458D-A116-27944A4006BB@bsdimp.com> <4DE0BA7C.8080707@freebsd.org> <BANLkTinnZyGoi9RCZRX0JQNvBhHnQwEOSg@mail.gmail.com> <20110528.130639.921.1@DEV> <4DE139C9.2080808@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
----- Original Message -----=0D=0AFrom: Julian Elischer = <julian@freebsd.org>=0D=0ATo: rank1seeker@gmail.com=0D=0ACc: = hackers@freebsd.org, Doug Ambrisko <ambrisko@freebsd.org>=0D=0ADate: Sat, = 28 May 2011 11:07:05 -0700=0D=0ASubject: Re: Active slice, only for a = next boot=0D=0A=0D=0A> On 5/28/11 6:06 AM, rank1seeker@gmail.com = wrote:=0D=0A> > And how about this:=0D=0A> >=0D=0A> > # boot0cfg -o = noupdate -s 1=0D=0A> > Now when you choose to hit slice 2, it is only for = a this one boot.=0D=0A> > Next and each boot, defaults to slice 1=0D=0A> = >=0D=0A> > Problem is, that you must see, early bootstrap, to manually = choose, so this won't work on a remote server.=0D=0A> > This = requires:=0D=0A> > a) physicall access=0D=0A> > or=0D=0A> > b) ssh = access to the remote box, which is conected via serial cable, to your = server.=0D=0A> >=0D=0A> > Anyone has any idea, for a case of a remote = server, which is accessible over ssh, only when it is "up"?=0D=0A> = =0D=0A> pull the old bootblocks from about 2000 and use them.=0D=0A> and = nextboot as well=0D=0A> they do exactly what you want.=0D=0A> =0D=0A> = OR=0D=0A> =0D=0A> ask Doug Ambrisko (cc'd) for a copy of them that he = still uses at work.=0D=0A> He may have updates to make them work with = modern systems that would =0D=0A> save you time.=0D=0A> =0D=0A> the old = nextboot(8) stored instructions as to what to do on block 1 of =0D=0A> = the drive=0D=0A> (you can make it a small 1 block partition if you = want). Actually it =0D=0A> stored a series of them, NULL = separated.=0D=0A> On each boot the boot 0 bloter would read the first = (after skipping =0D=0A> any nulls) and then write Nulls over=0D=0A> what = it just read and write it back to block 1.=0D=0A> so it would progress = gradualy boot by boot over the sequence written =0D=0A> by = nextboot.=0D=0A> =0D=0A> it would pass on the stack, what it had read to = boot1.=0D=0A> =0D=0A> =0D=0A> the format was "hd(1,a)/boot/loader" = (for example)=0D=0A> personally I would like to hav ethis capabiltiy back = because it's =0D=0A> stupid rely on a=0D=0A> possibly dead filesystem to = get around booting from the possibly dead =0D=0A> filesystem.=0D=0A> = =0D=0A> by default we used to have a /etc/rc entry that would rewrite the = =0D=0A> 'current' setup several times on successful boot,=0D=0A> followed = by a couple of alternate boot targets.=0D=0A> If boot failed a coupke of = times it would automatically boot from the =0D=0A> second drive or from = another partition,=0D=0A> ..=0D=0A> =0D=0A=0D=0AWell, if he has a working = code and you have an idea, then it should be modernized and integarted = into FreeBSD's src tree.=0D=0ASo everyone could use it.=0D=0A=0D=0A> = =0D=0A> =0D=0A> >=0D=0A> > Domagoj Smol=E8i=E6=0D=0A> > = _______________________________________________=0D=0A> > = freebsd-hackers@freebsd.org mailing list=0D=0A> > = http://lists.freebsd.org/mailman/listinfo/freebsd-hackers=0D=0A>; > To = unsubscribe, send any mail to = "freebsd-hackers-unsubscribe@freebsd.org"=0D=0A> >=0D=0A> >=0D=0A> = =0D=0A>
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20110529.152808.781.1>