[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [fab] Filesystems support in Fedora



On Tue, 2006-07-18 at 02:44 +0530, Rahul wrote:
> seth vidal wrote:
> > On Tue, 2006-07-18 at 02:33 +0530, Rahul wrote:
> >> Hi
> >>
> >> Ext3(ext2) are the only formally supported on disk filesystems in Fedora 
> >> apart from GFS for clustering. There is a limited option for other 
> >> filesystems like reiserfs, xfs and jfs which doesnt go much beyond 
> >> packaging what upstream provides and which is only meant to be 
> >> transiently used for people migrating off such filesystems.
> >>
> >> Moving such tools to Fedora Extras doesnt help since Anaconda wouldnt 
> >> support it during installation time. I am not sure this would make sense 
> >> even if Anaconda gets support for Fedora Extras. Community participation 
> >> on such things tend to be limited due to the amount of expertise 
> >> required. The large impact of such a core piece of technology needs to 
> >> be taken into account here.
> >>
> >> I believe it is better to get them supported or drop them completely 
> >> instead of the current status quo which leads to a less tested and 
> >> potentially dangerous option being provided to provided to end users 
> >> (though they have to enable it explicitly). Without good support, things 
> >> like SELinux in reiserfs to pick a example would end up being broken now 
> >> and then and that isnt a good thing at all.
> >>
> >> Realistically we need to make a hard choice on this. explain that well, 
> >> stick to that and make sure that support well what we do rather than 
> >> provide a multitude of half baked options (kernel-unsupported comes to 
> >> mind for those aware of the pains we had with it).
> >>
> > 
> > Why wouldn't this sort of technical decision be made by the installer
> > maintainers?
> > 
> 
> This has similarities to the RPM discussion that came up recently. 
> Though it is a specific package, any changes has a large impact. It is 
> better to have a wider look at this IMO.
> 

I disagree.

We're not debating the technical merits with rpm. We're debating the
political and policy issues with regard to the current interaction with
the upstream rpm maintainer.

In all cases where there is not a significant policy clash with the
upstream maintainer then we defer to the judgment of the upstream.

In this case upstream is the maintainers/developers of anaconda and the
kernel.

-sv



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]