Steffen Neumann <[email protected]> writes: > Hi, > > I am a lurker here, so just my 2c. We are running gridengine > inside a number of virtual machines on a small sized cluster > managed by opennebula.org.
It's worth hearing about. Have you seen previous posts about ON which will be in the gridengine.markmail.org archive? > A question I have is what do people do before shutdown/poweron: > in our setup, nodes do self-register upon poweron, and completely > de-register on poweroff. This way, we avoid that SGE sees many nodes > in alarm (au) state. Does that make sense ? I think so. I don't know how best to deal with it without support from GE itself. It may be necessary to tie into more of the normal cluster admin tools, such as monitoring, if that's relevant. > What's left (and beyond the scope of SGE) is the actual scheduling > of VMs, and eventually powering off the physical nodes. Isn't there other stuff under the Reservoir umbrella (Stratuslab?) that deals with that sort of thing? _______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
