On Wed, Jun 25, 2014 at 11:23 AM, 'Klaus Aehlig' via ganeti-devel < [email protected]> wrote:
> ...and warn that this option usually leads to an overestimation > of the cluster capacity. > > Signed-off-by: Klaus Aehlig <[email protected]> > --- > man/hspace.rst | 7 +++++++ > 1 file changed, 7 insertions(+) > > diff --git a/man/hspace.rst b/man/hspace.rst > index c887f97..a2c55a0 100644 > --- a/man/hspace.rst > +++ b/man/hspace.rst > @@ -239,6 +239,13 @@ The options that can be passed to the program are as > follows: > groups. The default is to not try allocation at all, if some not N+1 > happy node is found. > > +\--accept-existing-errors > + This is a strengthend form of \--independent-groups. It tells hspace > Nit: s/strengthend/strengthened/ > + to ignore the presence of not N+1 happy nodes and just allocate on > + all other nodes without introducing new N+1 violations. Note that this > + tends to overestimate the capacity, as instances still have to be > + moved away from the existing not N+1 happy nodes. > + > -l *rounds*, \--max-length=*rounds* > Restrict the number of instance allocations to this length. This is > not very useful in practice, but can be used for testing hspace > -- > 2.0.0.526.g5318336 > > LGTM
