Bug #381

EC: support "all" topology on the cmd line (omf load)

Added by Christoph Dwertmann over 1 year ago. Updated 2 months ago.

Status:Closed Start date:02/09/2010
Priority:Normal Due date:
Assignee:Christoph Dwertmann % Done:

100%
Category:OMF Software
Target version:OMF 5.4

Description

cdw@norbit:~/omf/3$ omf-5.3 load -t all -d
DEBUG: Executing '/usr/bin/ruby -I/usr/share/omf-expctl-5.3/ -I/usr/share/omf-common-5.3 /usr/share/omf-expctl-5.3/omf-expctl.rb -r -S pxe_slice -s system:exp:imageNode -- --nodes 'all' --image 'baseline.ndz' --timeout '800''

 INFO NodeHandler: OMF Experiment Controller 5.3.633
 INFO NodeHandler: Slice ID: pxe_slice 
 INFO NodeHandler: Experiment ID: pxe_slice-2010-09-02t21.19.03+10.00
 INFO NodeHandler: Message authentication is disabled
 INFO NodeHandler: Web interface available at: http://10.0.0.200:4000
 INFO Experiment: load system:exp:stdlib
 INFO property.resetDelay: value = 210 (Fixnum)
 INFO property.resetTries: value = 1 (Fixnum)
 INFO Experiment: load system:exp:eventlib
 INFO Experiment: load system:exp:imageNode
 INFO property.nodes: value = "all" (String)
 INFO property.image: value = "baseline.ndz" (String)
 INFO property.domain: value = nil (NilClass)
 INFO property.outpath: value = "/tmp" (String)
 INFO property.timeout: value = 800 (Fixnum)
 INFO Topology: Loading topology 'all'.
FATAL service_call: ------------
FATAL service_call:   A fatal error was encountered while making a request to an AM Service.
FATAL service_call:   Request: 'http://norbit.npc.nicta.com.au:5053/pxe/setBootImageNS?domain=norbit&ns=all'
FATAL service_call:   Exception: 'ServiceException'
FATAL service_call: ------------
FATAL run: ----------
FATAL run:   A fatal error was encountered while processing your experiment description.
FATAL run:   Exception: OEDLException
FATAL run:   Exception: (ServiceException) ServiceException
FATAL run:   In file: /usr/share/omf-expctl-5.3/omf-expctl/nodeHandler.rb:273:in `service_call'
FATAL run: ----------
 INFO NodeHandler: 
 INFO NodeHandler: Shutting down experiment, please wait...
 INFO NodeHandler: 
 INFO NodeHandler: Shutdown flag is set - Turning Off the resources
FATAL service_call: ------------
FATAL service_call:   A fatal error was encountered while making a request to an AM Service.
FATAL service_call:   Request: 'http://norbit.npc.nicta.com.au:5053/pxe/clearBootImageNS?domain=norbit&ns=all'
FATAL service_call:   Exception: 'ServiceException'
FATAL service_call: ------------
/usr/share/omf-expctl-5.3/omf-expctl/nodeHandler.rb:273:in `service_call': ServiceException (ServiceException)
    from /usr/share/omf-expctl-5.3/omf-expctl/node/nodeSet.rb:449:in `setPxeEnvMulti'
    from /usr/share/omf-expctl-5.3/omf-expctl/node/nodeSet.rb:414:in `pxeImage'
    from /usr/share/omf-expctl-5.3/omf-expctl/node/rootNodeSetPath.rb:100:in `pxeImage'
    from /usr/share/omf-expctl-5.3/omf-expctl/nodeHandler.rb:815:in `shutdown'
    from /usr/share/omf-expctl-5.3/omf-expctl.rb:99

History

#1
Updated by Christoph Dwertmann over 1 year ago

  • Target version changed from OMF 5.3 to OMF 5.3.1

This is broken in 5.3. As soon as we have a way to query which nodes are in a slice, we can implement system:topo:all again.

#2
Updated by Thierry Rakotoarivelo 12 months ago

  • Target version changed from OMF 5.3.1 to OMF 6.0

In the new Resource Handling scheme, we are considering everything as a resource. Thus the slice is also a resource, and the EC can send a request to it to find out the list of all resources it includes. This could be implemented for example by browsing the PubSub nodes under the "OMF/SLICEX/resources".

Fixing this in the current 5.3.x branch would require additional ad-hoc code on the EC-side, as the current OMF does not have any mechanism to allow an entity to find out the resource in a slice. Developing that code for 5.3.x might take some of our time out of working on 5.4... so I propose to move this to 5.4 for now. If we have a strong push from the users to include this again, then we will reconsider moving it to 5.3.x.

In the meantime, testbed operators can implement 'all' by explicitly listing all their resources in the all.rb topology file in the EC repository.

#3
Updated by Christoph Dwertmann 2 months ago

  • Status changed from Assigned to Closed
  • Target version changed from OMF 6.0 to OMF 5.4
  • % Done changed from 0 to 100

Fixed in c845477

The system:topo:all topology contains all nodes in a testbed.

Also available in: Atom PDF