[Pkg-gridengine-devel] Bug#815754: Bug#815754: Bug#815754: [PATCH] gridengine: proposal to add ppc64el architecture

[Wouter Verhelst]

On Sun, Feb 28, 2016 at 01:32:47AM -0800, Afif Elghraoui wrote:
> Hi, Fred,
> 
> على الجمعـة 26 شباط 2016 ‫01:18، كتب Frederic Bonnard:
> > sorry for that, indeed I patched the version in unstable.
> > So I worked on the git tree above, and made a new debdiff.
> > Some comments on the diff :
> > - there was some work done on ppc64el as you said : 
> >   https://arc.liv.ac.uk/trac/SGE/changeset/4712/sge
> >   Though as said in the commit upstream, this was not tested.
> >   So I changed a few little things (check patch header) that may be errors.
> >   Maybe that patch could go upstream.
> > - on the debian/control, I added the architecture ppc64el and there was also
> >   a typo in debian/rules I guess.
> > 
> 
> Excellent. Yes, this could also be forwarded upstream. I'll bring it up
> with him.
> 
> > I also see that gridengine in debian will be based soon on SoGE (yes, from the
> > git tree obviously :) and https://bugs.debian.org/703256 ).
> > Though, I'd like to know if SoGE is fully compatible with orginal Sun GE.
> 
> Based on upstream's notes [1], it should be, but I don't have firsthand
> experience in making the jump.

I do.

At an ex-customer they were using N1 Gridengine 6 with some proprietary
software that had builtin support for gridengine. Switching that to SoGE
made things fail. I tried to figure out why, but didn't manage to track
down the exact issue before my contract there ended.

The fact that qstat output now defaults to not showing jobs of other
users might be related, but I'm not sure if it is.

At any rate, the differences are fairly minimal, and mostly limited to
"new features" and similar things. I don't think this should stop us. If
something pops up within Debian, we can always find workarounds or
solutions; and if external proprietary applications limit themselves to
gridengine implementations that haven't been changed in five or so
years, they get what they deserve :-)

-- 
< ron> I mean, the main *practical* problem with C++, is there's like a dozen
       people in the world who think they really understand all of its rules,
       and pretty much all of them are just lying to themselves too.
 -- #debian-devel, OFTC, 2016-02-12