Re: Some feedback + EC2 Help
Patrick Armstrong
On 2010-04-12, at 4:54 PM, Richard Hosking wrote:
> First of all i think i may have found a bug with the cloud_status command, i had a bit of a look through the source and think i might have found the problem:
>
> When running the command "cloud_status -c NAME" it comes up with an error "Unexpected error: <class 'xmlrpclib.Fault'> <Fault 1: "<type 'exceptions.AttributeError'>:ResourcePool instance has no attribute 'get_cluster'">"
>
> The resource pool class in the cloud_management file doesnt seem to have a method called get_cluster so causes cloud_status -c NAME to fail (as far as i can tell).
Great. I reported the bug on the github issues:
http://github.com/hep-gc/cloud-scheduler/issues#issue/76
> Another thing, im having a few problems getting the nimbus setup to work trying out KVM and trying to figure out the networking issues, but for the moment i was thinking of trying out EC2 for provisioning the VM's. Was just wondering if you have any advice about the best way of going about that? havnt really used EC2 before, but have created an account and it seems easy enough.
It's pretty easy. I wrote up a EC2 Getting Started Guide on our group wiki at: https://wiki.gridx1.ca/twiki/bin/view/Main/EC2GettingStarted
That will get you started. There's lots of good information on the web about EC2.
Once you boot the Cloud Scheduler Test Drive VM, you shouldn't really have to do anything to provision your VMs, Cloud Scheduler will do it automatically to suit your VMs. Or it should anyway. :)
--patrick
PS. I took the liberty of CCing this to our new Cloud Scheduler mailing list.