Hya,

Background: I have several Debian's (atom and Core, all running Debian 64bit) sitting around me. I need to run quite a few batch jobs. (Each job runs 1-5 min)

My idea: I guess it is a good idea to build some sort of cluster. I do not think I need a sophisticated cluster.

Web search: In the past, several threads were posted in LQ.

Question: What shall I do?

Any suggestion will be appreciated.

Hya,

I guess there's not much change recently, I will try slurm. I will post if there is any news.

Happy Penguins!

Hya,

I started with slurm, ran into trouble. (slurm did not like .local domain assigned by avahi)

Then I went to gridengine, which did not work well. qmon was OK, then "unable to run job, your job is not allowed to run in any queue" I could not resolve this condition, because I was unable to locate good documentation. (thanks to oracle?)

I went back to slurm, set up static IP, by historical /etc/hosts file. It is currently working with minor trouble, which is only two jobs run on 4 thread-CPU. Does anybody know how to resolve this condition?

Happy Penguins!