Re: CURRENT: WARNING! r273914 leaves filesystems in inconsistent/corrupted condition!

On Sat, 2014-11-01 at 15:21 +0100, Dag-Erling Smørgrav wrote:
> Tomoaki AOKI <junchoon_at_dec.sakura.ne.jp> writes:
> > Dag-Erling Smørgrav <des_at_des.no> writes:
> > > Manfred Antar <null_at_pozo.com> writes:
> > > > Then for some reason /var started to being mounted mfs.  [...]  If
> > > > I have varmfs="NO" and cleanvar_enable="NO" everything works fine.
> > > Not really.  The default for varmfs is AUTO, which mounts a memory
> > > file system on /var if, after mounting all "early" file systems,
> > > /var is not writeable.
> > For me, Manfred's workaround actually helped.
> 
> It helped that particular issue, more or less by accident.  It was not
> in any way a correct fix or even a correct workaround.
> 
> > In single user mode, actual /var (in root partition) appears as
> > before.  So there can be some mis-ordering within rc scripts.
> > (Remounting of / is delayed? Check for /var too early?)
> 
> Exactly right; the check for a writeable /var occurred before / was
> mounted r/w, so it mounted an mfs instead.  Xin fixed this in r273919.
> 
> > For me, [unblocking /dev/random] takes nearly 2 minutes each boot
> > after r273872.  No specific rc.conf setting for it.
> 
> That means we're not getting enough entropy during early boot, or we're
> underestimating the amount of entropy we're getting.  We added entropy
> harvesting to device_attach() about a year ago, which in most cases
> provides enough entropy to unblock /dev/random before we even run
> init(8).
> 
> DES

And I vaguely remember being promised that things like that would NEVER
happen, even on systems with little or no entropy available during early
startup (which describes quite nicely the embedded systems we build at
work).

-- Ian