On 17-7-2016 14:44, Willem Jan Withagen wrote: > On 14-7-2016 22:50, Willem Jan Withagen wrote: >> On 7-7-2016 00:46, Brad Hubbard wrote: >>> On Thu, Jul 7, 2016 at 3:22 AM, Willem Jan Withagen <wjw@xxxxxxxxxxx> wrote: >>>> Hi, >>>> >>>> I'm trying to run cephtool-test-mon.sh, and it gets there quite far >>>> but somewhere down the test I have this error: >>>> Now python is not my strong point. >>> >>> Not mine either but monitor_log should be defined in >>> src/pybind/rados/rados.pyx and within the >>> "Rados" object definition. >>> >>> 519 cdef class Rados(object): >>> ... >>> 1280 def monitor_log(self, level, callback, arg): >>> >>> I guess just checking they are there is a start :) >>> >>> HTH, >>> Brad >> >> Fixed by recompilation. >> So could be a faulty leftover from previous build attempts. > > Not quite.... > > Turns out that I had ceph-libs in /usr/local/lib from a leftover (wrong) > setup. > So that means that the search path of python libraries prefers > /usr/local/lib over the libs in buidl/lib... > I guess that that is a serach path setting, but would it be wise move > ./lib more up front? Nothing ever seems to be as easy as it looks like to be... This required a fix in cmake/modules/AddCephTest.cmake Which contains a hardcode path with lib.linux-x86_64-2.7 in it, which is the location where cython leaves certain results. Making this behave a bit nicer seems to work. --WjW -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
