OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

tosca message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: [OASIS Issue Tracker] (TOSCA-244) Need to add clarity around orchestrator's copy behavior for script


    [ https://issues.oasis-open.org/browse/TOSCA-244?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=59763#comment-59763 ] 

Luca Gioppo commented on TOSCA-244:
-----------------------------------

Artifacts should only copyed in the place that is needed.

Consider a DB dump: I have 800 MB of Db dump that represent the initial state of my newly created application I will place it in the only place I need it and not on all the VM created for sure.

In my case since I generate the Docker image on the fly everything stays in the orchestrator machine.

Artifacts as I see in the examples are too simple for my use case; I have to generate artifacts dynamically fron the orchestrator since they represent puppet's manifests or dockerfiles and I build the docker image in the orchestrator so that the docker container will run on this new image.

This trick removes the need to do complex operations on remote systems.

> Need to add clarity around orchestrator's copy behavior for script
> ------------------------------------------------------------------
>
>                 Key: TOSCA-244
>                 URL: https://issues.oasis-open.org/browse/TOSCA-244
>             Project: OASIS Topology and Orchestration Specification for Cloud Applications (TOSCA) TC
>          Issue Type: Task
>          Components: Profile-YAML
>            Reporter: Sahdev Zala 
>            Assignee: Luc Boutier
>             Fix For: CSD04
>
>
> Per discussion on 05/07/15 TOSCA TC Special 4 Hour Meeting:
> (Luc) : 
> Default behavior: Assumed that orchestrators execute relative to the CSAR and removes on completion.
> Now that we have get_artifact() function which allows template author to control placement and persistence of an artifact if they want to control it directly.
> Luc: Perhaps we do not want to copy on every compute node if for example we have 5 compute nodes.  
> Note: orchestrators should only copy artifacts for that node to that compute not all artifacts to all nodes.
> (keep wordpress scripts on its compute and mysql on its compute).
> Needs to reference the 'hosting' (or containment) hirarachy. 



--
This message was sent by Atlassian JIRA
(v6.2.2#6258)


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]