On 2007-12-10 00:54, Ivan Voras <ivoras_at_freebsd.org> wrote: >On 10/12/2007, Giorgos Keramidas <keramida_at_ceid.upatras.gr> wrote: >> Some of these are in the PEBKAC range of problems, though, i.e.: >> >> => tmpfs and fstab/fsck errors >> >> Since tmpfs /tmp is a memory file system, which is created from >> scratch every time, it makes no sense to fsck it on boot. > > And this gives it the right to block system from booting? I'd at least > like a symlink from "true" to "fsck_tmpfs". Heh, no, there's no reason to block the boot process because of a missing binary (although some may argue that botching fstab is not something that should be taken lightly). A *real* fix is probably something we should discuss with Julio M. Merino Vidal <jmmv at NetBSD.org> and the people who ported tmpfs to FreeBSD. For example, does it make sense to fsck a tmpfs, ever? When it's not mounted, is it accessible in any way? What are the `use cases' of the tmpfs filesystem, other than memory-backed /tmp mounts? > > => SCTP depends on IPv6 > > > > This is clearly documented in NOTES. I don't think it should count > > as a surprising fact. > > I know it's in the NOTES - I've added it when it bit me since I didn't > read the notes. Maybe I'll remove it. If it bit you, an experienced FreeBSD developer, it can bite others too. Let's work with re_at_ to get it in the release notes, if it's not already there (instead of deleting it from the list) :-) My last CVSup is from Wed Dec 05 19:32:59 2007 +0000 (right before I started preparing for Friday's FreeBSD talk at a nearby university), so I'll CVSup again tomorrow morning, and check how we can fit this into the release notes in a nice, non-intrusive but still useful manner.Received on Sun Dec 09 2007 - 23:22:01 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:23 UTC