Personally if I were building a new cluster I would probably go with slurm. Even though its architecture seems a bit more kludgey than SGE, it's much more actively maintained. I'm not sure what others here think. Of course there is also Univa GridEngine, the proprietary version, which is better maintained and (I believe) comes with support.
Dan On Fri, May 17, 2019 at 1:56 PM Müller, Marcus (CEL) <muel...@kit.edu> wrote: > October '18. From the kind of activity, I'd say Dave Love is a time- > restricted yet dedicated FOSS maintainer. I'd assume he's busy > administrating an SGE installation, or simply doing something > satisfying with his life. Is that not the case? (Looking at the amount > of effort put into SGE, I really hope at least the latter is the case.) > > To be honest, your question is not exactly trust-inspiring; I'm just > trying to build SGE to have a way of running it on modern Linux > distros. > This is just to evaluate it as cluster management tool (a class of > software I've never used, leave alone set up, before), because there is > in-house experience using SGE, so that was the first system I'm looking > at. > > So, what's up with SGE, Daniel? Still recommended for production/new > designs? > > Best regards, > Marcus > > On Fri, 2019-05-17 at 12:16 -0400, Daniel Povey wrote: > > So is Dave Love still maintaining SGE? When is the last commit? > > > > On Fri, May 17, 2019 at 11:50 AM Marcus Müller <muel...@kit.edu> > > wrote: > > > Dear Maintainers / list, > > > > > > trying to build SGE for OpenSuSE leap 15.0 and Fedora >= 28, I > > > encountered quite a few obstacles, so I started by trying to do > > > what the > > > .spec does (fixes for that coming up, if you're interested). > > > > > > However, when the thing finally built under OS leap, the sanity > > > checker > > > beat me with a stick for a missing n that slipped through Dave > > > Love's > > > fingers in 9dcf22749e, "Change many sprintf to snprintf". > > > > > > Looks like that would actually lead to segfaults if the function > > > containing that typo was used. > > > > > > Marcus Müller (1): > > > qmon_arsub: actually use sNprintf when specifying length > > > > > > source/clients/qmon/qmon_arsub.c | 2 +- > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > _______________________________________________ SGE-discuss mailing list SGE-discuss@liv.ac.uk https://arc.liv.ac.uk/mailman/listinfo/sge-discuss
