Daniel Veillard wrote: > I realized that one of the NUMA API entry point didn't had a >binding in Python, the enclosed patch adds it, it's a method >on a virConnect class, taking the startCell and maxCells integer >parameters and returning a list of available memory for that >range of cell, using it should be as simple as launching python >as root after reinstallation of the libvirt library and bindings >and doing the following: > >>>>import libvirt >>>>conn = libvirt.open(None) >>>>conn.getCellsFreeMemory(0, 20) Beat you to that binding by several weeks then :-)

http://libvirt.org/ocaml/html/Libvirt.Connect.html#VALnode_get_cells_free... +1 for the patch.

Daniel Veillard wrote: >On Thu, Dec 06, 2007 at 06:19:07PM +0000, Richard W.M. Jones wrote: > > >>Daniel Veillard wrote: >> >> >>>I realized that one of the NUMA API entry point didn't had a >>>binding in Python, the enclosed patch adds it, it's a method >>>on a virConnect class, taking the startCell and maxCells integer >>>parameters and returning a list of available memory for that >>>range of cell, using it should be as simple as launching python >>>as root after reinstallation of the libvirt library and bindings >>>and doing the following: >>> >>> >>> >>>>>>import libvirt >>>>>>conn = libvirt.open(None) >>>>>>conn.getCellsFreeMemory(0, 20) >>>>>> >>>>>> >>Beat you to that binding by several weeks then :-) >> >> > > heh :-) > > > >>http://libvirt.org/ocaml/html/Libvirt.Connect.html#VALnode_get_cells_free_memory >> >>+1 for the patch. >> >> > > If it could get a bit of testing I would feel more confident ! > >Daniel > > > Hi Daniel. I was able to test on a 2-node (the best I could do at the moment :-) and it worked!