Alfredo Ferrari wrote:
... indeed we have RHEL on the master member of the cluster and our web
server, but it would be unpractical to pay a license for each slave, or
for each personal dekstop / laptop in the group.
Are you kidding me? Are you seriously saying that your data is not
worth $79/node - list? Have a look at:
http://www.redhat.com/rhel/details/hpc/
If your cluster is not worth $79 per node, why are you bitching so loud?
If you are a sysadmin on a production environment, you need to set up
an environment that looks something like:
Dev -> QA -> Production
You need to roll updates into Dev and pound the Hell out of them with
your specific workload. Then you roll those changes into QA and have
your users do the same thing. Finally, after user acceptance, *then*
you roll into Production.
If you bypass Dev and QA, I'm afraid I can't really be too terribly
sympathetic.
Thomas
