[blivet:all][PATCH] Don't try to get no profile's name (#1151458)

Vratislav Podzimek vpodzime at redhat.com
Wed Oct 22 05:37:31 UTC 2014


On Tue, 2014-10-21 at 08:27 -0400, Anne Mulhern wrote:
> 
> 
> 
> ----- Original Message -----
> > From: "Vratislav Podzimek" <vpodzime at redhat.com>
> > To: anaconda-patches at lists.fedorahosted.org
> > Sent: Tuesday, October 21, 2014 3:45:48 AM
> > Subject: [blivet:all][PATCH] Don't try to get no profile's name (#1151458)
> > 
> > If no profile is specified, trying to get its name results in a traceback.
> > 
> > Signed-off-by: Vratislav Podzimek <vpodzime at redhat.com>
> > ---
> >  blivet/devices.py | 7 ++++++-
> >  1 file changed, 6 insertions(+), 1 deletion(-)
> > 
> > diff --git a/blivet/devices.py b/blivet/devices.py
> > index 56f8dcc..75afcc2 100644
> > --- a/blivet/devices.py
> > +++ b/blivet/devices.py
> > @@ -3584,10 +3584,15 @@ class LVMThinPoolDevice(LVMLogicalVolumeDevice):
> >          """ Create the device. """
> >          log_method_call(self, self.name, status=self.status)
> >          # TODO: chunk size, data/metadata split --> profile
> > +        if self.profile:
> > +            profile_name = self.profile.name
> > +        else:
> > +            profile_name = ""
> > +
> >          lvm.thinpoolcreate(self.vg.name, self.lvname, self.size,
> >                             metadatasize=self.metaDataSize,
> >                             chunksize=self.chunkSize,
> > -                           profile=self.profile.name)
> > +                           profile=profile_name)
> >  
> >      def dracutSetupArgs(self):
> >          return set()
> > --
> > 1.9.3
> > 
> > _______________________________________________
> > anaconda-patches mailing list
> > anaconda-patches at lists.fedorahosted.org
> > https://lists.fedorahosted.org/mailman/listinfo/anaconda-patches
> > 
> 
> It looks to me like it might make sense to fix this class's
> populateKSData() method at the same time so that data.profile
> is the str name, not the ThPoolProfile object. Or you could
> make fix related to rhbz#869456.
Good catch! Fixing locally.

> 
> Also, right next to them metaDataSize and chunkSize definitely
> have similar unit/type problems...but those difficulties seem
> of long standing and a bit harder to fix. Since you know
> more about lvm though, you might want to take a look.
These are okay, I think, because they default to 0 which is then ignored
by LogVolData when producing its string representation.

> 
> Also, I can't find any mention of the --profile option in current
> RHEL7 installation guide. Maybe it's worth filing a bug so
> that they can put it in?
I could clone the bug rhbz#869456 to documentation, but I feel like we
need some more general approach here for new stuff in 7.1.

> 
> I concur w/ Samantha on hasattr...but I think getattr with
> a default would be a bit better.
I've discussed this with Samantha on IRC I think that anyone passing
something else then ThPoolProfile as the profile argument (and I expect
me to be the first one) deserves that to blow out into their face as it
means something wrong is happenning somewhere.

-- 
Vratislav Podzimek

Anaconda Rider | Red Hat, Inc. | Brno - Czech Republic



More information about the anaconda-patches mailing list