[one-users] appFlow vs appStage for hadoop

Daniel Molina dmolina at opennebula.org
Thu May 9 07:45:19 PDT 2013


On 2 May 2013 11:59, Zeeshan Ali Shah <zashah at pdc.kth.se> wrote:

> Hi, For a use case of Hadoop cluster (just e.g) with few clicks . I looked
> into appFlow and appStage both what i noticed is that
> 1. appStage is for (single instance installation) i.e. to have all service
> in one vm .. Is this true ?
> 2. For multi tiered application you have to use appFlow but than you are
> not using chef rather a custom JSON format.
> are above true or i missed some points ?

You are right. Currently you can:

1. Customize your instances through the appStage tool (#1 instance). [1]
2. Define dependencies among different instances using appFlow to build
your service (#multiple instances) [2]

But you cannot use both at the same time. This is a feature that is in our
roadmap and we will consider it for the next release.

Meanwhile you can use appFlow to define your service and customize each
instance using the OpenNebula contextualization system [3]

PS: Note that OpenNebulaApps have not been updated to work with the last
OpenNebula 4.0 release. There will be an updated version of OpenNebulaApps
in a few day fully integrated with the new Sunstone


[1] http://docs.opennebula.pro/appstage
[2] http://docs.opennebula.pro/appflow
[3] http://opennebula.org/documentation:rel4.0:context_overview

Join us at OpenNebulaConf2013 <http://opennebulaconf.com/> in Berlin, 24-26
September, 2013
Daniel Molina
Project Engineer
OpenNebula - The Open Source Solution for Data Center Virtualization
www.OpenNebula.org | dmolina at opennebula.org | @OpenNebula
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opennebula.org/pipermail/users-opennebula.org/attachments/20130509/0cad3f31/attachment-0002.htm>

More information about the Users mailing list