Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

We have jobs that run on single servers on various schedules, but we need to run them on a cluster of servers regardless of which server is going to run them. It is as if we need a cluster-aware CRON (or Windows Scheduler for our Windows servers).

Does anyone have any experience in this? How would you handle such scenario?

share|improve this question

4 Answers 4

Check out mcollective. Their introductory screencast is actually quite good.

share|improve this answer

On windows clusters I am used to schedule on both nodes. At the starting of the batch I check for shared volume existance. If current node own the volume, the batch goes on, otherwise it exit without do anythig.

IF NOT EXIST I:\NUL GOTO NOVOLUMEI

CD I:\

:NOVOLUMEI

share|improve this answer

How about a cluster resource manager ? (open-source ones are Torque, Oracle Grid Engine, or Slurm)

share|improve this answer
1  
A cluster resource manager seems like a massive overkill for something as simple as running a task on any of the available machine. Aren't cluster resource managers for running code on clusters? –  Khash Dec 22 '11 at 23:11
    
Well, you didn't say how large your cluster is nor how many jobs you have. And what Torque does is essentially to keep track of free machines and dispatch shell scripts to them, according to rules you can optionally define. –  b0fh Jan 6 '12 at 10:47

What about Puppet or OCS Inventory NG?

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.