> > can someone take a short amount of time to outline what some of these
> > "mustfix" bugs are/1.2 prerelease "issues" remaining to be fixed? Are
> > the 1.2 prereleases "safe" to use on production data (i.e. outstanding
> > bugs only pop up under really rare conditions, other "issues" are
> > nice-to-haves but don't affect functionality) or should we be careful
> > with them?
> There was a filesystem force shutdown on nfs-exported filesystems, and
> very hard to hit data corruption case with fs blocksize < pagesize, mmap
> IO, and high memory pressure... I don't think anyone ever hit it outside
> of SGI. This is fixed (crossed fingers!) in -pre3.
i'm with 6 november CVS code (2 day uptime) and (keeping the fingers
crossed ;)) the first bug you've mentioned seems to be solved (usually
oops raised within few hours).