How to setup puppet to autoinstall a Sun Grind Engine execute node
Dear All, i am new to puppet and IT automation, since in my setting software is typically installed manually ... I need to extend my Sun Grid Engine 6.2u7 with additional execute nodes, perhaps on EC2 with VPC -VPN connectivity, to scale out with the bioinformatics "Galaxy" open source framework.
AWS CloudFormation was my first attempt, however it does not impresses me since - it does not provision a proper develop/test environment (one goes "live" immediately!) - it is very much AWS specific (what about moving to another provider..)
Since in the IAAS stack I need the SGE, the Univa Unicloud is appealing, however I feel it being not fully matured. Nevertheless, studying their specs I've learned about Puppet.
However, I've searched in the PuppetLab and could not find any trace of a config neither for the SGE nor for Galaxy.
For SGE, the manual installation procedure is rather cumbersome, with several manual steps..
For Galaxy, the work that needs to be done is well documented (and I did couple of times by hand) and basically involves a) on the master node - add the compute node name in the list of execute nodes b) on the file server - export to the execute node specific directories c) on the compute node - create a specific user account - import some directories from b)
I'd appreciate if anyone could share experience on autoinstalling SGE and Galaxy with puppet, and the advantage / disadvantage compared to native AWS Cloud Formation.