We have 3-4 strong linux machines with access to a shared file system (
samba) and 5 users. During every work day, all users need to run heavy ad-hoc jobs (C++, shell and python scripts), usually the same job is executed multiple times, each execution with different parameters.
Currently we're using
condor, and we love its simplicity - we just
condor_submit command line scripts which is great, and condor is load balancing our jobs on all available machines. However, our condor agents keep crashing, and for several months we've been trying to triage and fix the problem and to no avail (it's not helpful that the support level and data in stack-exchange is very low for condor :( ).
What modern alternative would you recommend for such simple needs? It should be reliable, we need just the minimal functionality, and most of all it should have a strong community/knowledge-base somewhere on the web (or it could be so bug-free we won't need any help :)) Do we go to dkron? bistro? Some other tool with
Thanks, and please don't regard this question as an open invitation to a non-constructive flame war on the pros and cons of those tools - I'm just asking what's the right tool for us.
preferably free and open-source of course. However, if there's something better that requires a license, I'd give it a try and attempt to persuade my company that its worth paying for.
(related to https://serverfault.com/questions/277821/outgrowing-cron-whats-the-next-scheduler but they were looking for a single machine scheduler, not distributed)