Date: Thu, 15 Apr 2004 13:13:53 -0700 From: Brooks Davis <brooks@one-eyed-alien.net> To: Kent Stewart <kstewart@owt.com> Cc: Robert Watson <rwatson@freebsd.org> Subject: Re: very current won't mount /tmp Message-ID: <20040415201353.GA11772@Odin.AC.HMC.Edu> In-Reply-To: <200404151300.30275.kstewart@owt.com> References: <16507.9094.803648.85615@roam.psg.com> <Pine.NEB.3.96L.1040412195402.19174B-100000@fledge.watson.org> <20040415061146.GD4415@ip.net.ua> <200404151300.30275.kstewart@owt.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--+HP7ph2BbKc20aGI Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Thu, Apr 15, 2004 at 01:00:30PM -0700, Kent Stewart wrote: > On Wednesday 14 April 2004 11:11 pm, Ruslan Ermilov wrote: > > On Mon, Apr 12, 2004 at 07:54:26PM -0400, Robert Watson wrote: > > > On Tue, 13 Apr 2004, Randy Bush wrote: > > > > >> very current (08:00gmt the 11th) > > > > >> thinkpad t40p > > > > >> > > > > >> make world and kernel > > > > >> boot single abuser to do the installworld > > > > >> > > > > >> can mount drives one at a time until i get to /tmp which is > > > > >> defined as > > > > >> > > > > >> md /tmp mfs -s131072,rw,nosuid,nodev,noatime 0 0 > > > > >> > > > > >> locks up > > > > >> > > > > >> as installworld has not happened, it should not be the new > > > > >> automagic /tmp mfs. or am i misunderstanding that? > > > > > > > > > > The new auto /tmp creation code should only kick in after you > > > > > run mergemaster. Even then, it shouldn't have any effect in > > > > > single user mode because it hasn't been run yet at that point.=20 > > > > > If you're being really paranoid, you simply verifying that > > > > > nothing is mounted on /tmp before trying to mount it yourself > > > > > should eliminate that from the equation. > > > > > > > > i tried removing md /tmp from fstab and rebooting -sw > > > > > > > > mount -a went fine > > > > > > > > vi locks up immediately > > > > > > > > make installworld locks up immediately > > > > > > > > all this is new and very disturbing > > > > > > Any chance you can get into DDB eitgher using console break or > > > serial console break? > > > > The vi(1) will be shown as blocked. One now needs to manually feed > > /dev/random, while in single-user mode. > > > > Mark, please do something about it in a short term, add it to > > UPDATING at least. > > > I think updating UPDATING is a good idea. I tried "feed /dev/random" and= =20 > it couldn't find "feed". What are we supposed to do to get installworld= =20 > to work? Here's a possiably UPDATING entry, how's it look? 20040409: Due to changes in the the Yarrow initialization process, /dev/random needs to be feed before operations requiring temp files can succeed in single user mode. This includes running "make installworld". /dev/random can be feed by running "/etc/rc.d/initrandom start" or with 20040315 source by running "/etc/rc.d/preseedrandom". Appropriate entries also need to be added to the build instructions. -- Brooks --=20 Any statement of the form "X is the one, true Y" is FALSE. PGP fingerprint 655D 519C 26A7 82E7 2529 9BF0 5D8E 8BE9 F238 1AD4 --+HP7ph2BbKc20aGI Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (GNU/Linux) iD8DBQFAfuz9XY6L6fI4GtQRAvazAKCv0hYlDeXV6ZmrsIrNSQpH2yKzVgCeKTqd /Ujx/BERZrGjilLOUn//pRo= =YfCY -----END PGP SIGNATURE----- --+HP7ph2BbKc20aGI--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040415201353.GA11772>