Cloud computing offers IT resources on-demand. The IaaS (Infrastructure-as-a-Service) Cloud type provides virtual machines, network and storage to run applications in a customised environment. In order to realize long term preservation of the Planck algorithm, the Cloud computing services are used.
The Planck algorithm is on a disk image within a cutomized environnement. The disk image is referenced on the catalog of image onf two academic Clouds i.e. the general MarketPlace of StratusLab and the local Glance catalog of OpenStack@CC-IN2P3.
The Planck data are on persistent disks of both infrastructures.
More help on Cloud client can be found [here](https://www.apc.univ-paris7.fr/FACeWiki/pmwiki.php?n=Cloud.Cloud).
Ask for an account creation on the Cloud infrastructure of your choice.
https://cloud.lal.stratuslab.eu/register/
https://cc-usersupport.in2p3.fr/otrs/customer.pl Note: this Cloud is a community Cloud and is not open yet (early 2015) to projet. To ask a Cloud account, you need to already have a CC-IN2P3 account and to use the ticket system.
Install locally the command line client.
pip install –user stratuslab-client edit ~/.stratuslab/stratuslab-user.cfg
pip install –user python-novaclient python-glanceclient python-cinderclient edit ~/.novacreds/novarc.sh
Instanciate a VM with specific disk image containing all the environnement and ressources in order to run the Planck pipeline.
ID = BJILII-tFu00rKKM-enj9l83rsn stratus-run-instance -t m1.large –vm-name=planck ID stratus-describe-instance stratus-connect-instance id
ID = c0de1a4a-ba7d-443c-a50d-b22170495b96 nova boot –key-name=cloudkey –image=ID –flavor=m1.small.2 planck nova add-floating-ip planck IP edit .ssh/config Host vm_cloud_cc
IdentityFile ~/.novacreds/cloudkey HostName IP ProxyCommand ssh user@ccage.in2p3.fr nc %h %p
ssh root@vm_cloud_cc
Attach and mount persistent disk in order to access to the Planck data.
stratus-attach-volume -i id UUID mount /dev/vdc /mnt
nova volume-attach planck UUID /dev/vdc mount /dev/vdc /mnt
Run the pipeline from the VM as user.
pipeweb track planck /mnt/pipelet/.sqlstatus pipeutils -a user -l 2 /mnt/pipelet/.sqlstatus
pipeweb start ...