dolfin team mailing list archive

[Johannes Ring <johannr@xxxxxxxxx>]

On Wed, Dec 5, 2012 at 9:10 AM, Johannes Ring <johannr@xxxxxxxxx> wrote:
> On Tue, Dec 4, 2012 at 11:17 PM, Garth N. Wells <gnw20@xxxxxxxxx> wrote:
>> On Tue, Dec 4, 2012 at 2:34 PM, Anders Logg <logg@xxxxxxxxx> wrote:
>>> The buildbots were green yesterday but some of them are failing
>>> again. The problem seems to be a an error appearing "at random" for
>>> some demos using PETSc in combination with Hypre. Johannes and I
>>> discussed it and the problem is likely a partly broken OpenMPI package
>>> in Ubuntu Precise (12.04).
>>>
>>> What is the best way to handle this? Tuning the demos to avoid the
>>> problems with the Ubuntu Precise OpenMPI package does not seem to be
>>> the best solution.
>>>
>>> My suggestion would be to disable parallel testing on the Ubuntu
>>> Precise buildbot. Are there other solutions?
>>>
>>
>> Disabling sounds good.
>
> I have rebuilt the Ubuntu PETSc package without Hypre and that seems to have
> fixed the problem. Maybe it's a bug in the Hypre package on Precise. I can
> try backporting the Hypre package from Quantal and see if that works, but if
> not, I guess Precise users must live without Hypre supportere.

Sorry, I'm not sure why I thought it was working. I guess I was
tricked by my home computer, which showed the buildbot to be all
green. Maybe I didn't refresh the page correctly. Now, at work, I see
that everything is failing at the compile stage.

Johannes