We got these two clusters with eight nodes each, and we are looking for a good cluster framework that would allow us to launch jobs, has an inbuilt scheduler with different scheduling policies and a monitoring system with web frontend. Each of the nodes are running on Ubuntu 11.04. Both commercial and opensource are OK.
Some of them i saw were,
TORQUE and MAUI.(Not sure if it has a web frontend for monitoring)
SLURM and MAUI.
GEXEC and GANGLIA.(Doesn’t have a scheduler)
Which product(s) would you recommend? Also is there any advantage using cluster operating systems like MOSIX instead of tools?
The paid version of Maui is called Moab (it normally uses TORQUE as the RM). It also can be sold with monitoring tools as well. I think Moab is a really good product, but I am strongly biased towards it (I work for the company that develops TORQUE/Moab).