BonFIRE logo and link to main BonFIRE site

Table Of Contents

Previous topic

Experiment Lifecycle

Next topic

BonFIRE OVF Schema Definition

This Page

States of OCCI Resources

In BonFIRE OCCI, the resource descriptions for the COMPUTE, NETWORK (only at the Virtual Wall) and STORAGE resources include a state attached. The content of this is determined by the virtualization manager and so will differ between testbeds. Where log files are available, the same states will be used. The following is a list of the states on the different virtualization managers used in BonFIRE:

OpenNebula

The COMPUTE resource can be in the following states:

INIT         --> This state is no longer used, but may appear in some old log files.
PENDING      --> The VM is waiting for the scheduler
HOLD         --> The VM needs to be released
ACTIVE (*)   --> the VM just entered the VM life cycle manager
STOPPED      --> The VM is stopped, waiting for resume
SUSPENDED    --> The VM is suspended, waiting for resume in the same host
DONE         --> The VM has ended the lifecycle succesfully
FAILED       --> The VM has ended the lifecycle due to a problem

(*) where the ACTIVE state get unfolded in

PROLOG           --> The images of the VM are being staged to the node
BOOT             --> The VM is booting
RUNNING          --> The VM is running
MIGRATE          --> The VM is being migrated
SAVE_STOP        --> The VM is in process of being stopped
SAVE_SUSPEND     --> The VM is in process of being suspended
SAVE_MIGRATE     --> The VM is in process of saving becayse of it being migrated
PROLOG_MIGRATE   --> The VM is having its files staged in the target host due to migration
PROLOG_RESUME    --> The VM is having its files staged in the target host due to resuming
EPILOG_STOP      --> The VM is having its files staged into the front-end due to stopping
EPILOG           --> The VM is having its files staged into the front-end due to ending lifecycle
SHUTDOWN         --> The VM is in process of shutting down
CANCEL           --> The VM is in process of being cancel
FAILURE          --> The VM has reached a fail state in its VM lifecycle manager
CLEANUP          --> Files of the VM are being cleaned
UNKNOWN          --> The VM has disappeared from the hypervisor

The STORAGE resource can be in the following states:

INIT       --> This state is no longer used, but may appear in some old log files.
LOCKED     --> The storage is being copied or created.
READY      --> The storage is ready to be used.
USED       --> The storage is used by at least one VM. If may still be used by other VMs.
USED_PERS  --> The persistent storage is used by at a VM. It cannot be used by other VMs.
DISABLED   --> The storage is disabled at the site level and cannot be used.
ERROR      --> An error occurred. The storage cannot be used.
DELETE     --> The image is being deleted.

VirtualWall

A quick remark. The Virtual Wall deploys all resources involved with a particular EXPERIMENT atomically. The states that are involved in this atomic action are indicated by an asterisk (*). The italic part between brackets - in the state’s description - points to how this state is represented in the Virtual Wall log files (http://ssh.be-ibbt.bonfire-project.eu/logs). The COMPUTE resource can be in the following states:

PENDING      --> The COMPUTE is described by the experimenter and waiting for the scheduler or the EXPERIMENT GO
SWAPIN [1]   --> The Virtual Wall is preparing the EXPERIMENT
BOOTING [2]  --> The physical nodes that will harbor the COMPUTEs are prepared (*os_setup started ... os_setup finished*)
ISUP         --> The COMPUTE is fully booted and accessible via the BonFIRE WAN or via the EXPERIMENT's emulated NETWORK
BFSETUP      --> BonFIRE specific init scripts and the experimenter's postinstall scripts are executed
ACTIVE       --> Everything is set and running, the real EXPERIMENT can start
TERMINATING  --> The COMPUTE is shutting down as part of the EXPERIMENT shutdown, the COMPUTE is still accessible
TERMINATED   --> The COMPUTE is removed from the system and not accessible anymore
[1]where the SWAPIN state gets unfolded in the following substates.
*Pre run  --> The EXPERIMENT description is validated and prepared for 'swap-in'. (*Beginning pre run ... Pre run finished*)
*Swap-in  --> The EXPERIMENT is being deployed to the Virtual Wall. (*Beginning swap-in*)
*Assign   --> The different resources are mapped to physical reality. (*assign_wrapper started ... assign_wrapper finished*)
[2]where the TERMINATING state gets unfolded in the following substates.
*Save-as  --> When necessary the persistent STORAGES are saved and new images are created for the SAVE_AS STORAGES.
*Swap-out --> All physical resources (COMPUTES, NETWORKS and STORAGES) are freed. (*tbswap out started ... tbswap out finished*)

The NETWORK resource can be in the following states:

PENDING       --> The NETWORK is described by the experimenter and waiting for the scheduler or the EXPERIMENT GO
*SWAPIN [1]      --> The Virtual Wall is preparing the EXPERIMENT
UP            --> The NETWORK is available for packages to be sent by the connected COMPUTES
DOWN          --> The NETWORK is brought down, i.e. every communication using this NETWORK will fail
*TERMINATING [2]  --> The NETWORK is shutting down as part of the EXPERIMENT shutdown, communication over the NETWORK is still possible
*TERMINATED    --> The NETWORK is removed from the Virtual Wall gigabit switch

The STORAGE resource can be in the following states:

PENDING       --> The STORAGE is described by the experimenter. STORAGES created outside the context of an EXPERIMENT are initialized immediately and other STORAGES are waiting for the scheduler or the EXPERIMENT GO
LINKED        --> The STORAGE is linked to a COMPUTE
*SWAPIN [1]       --> The Virtual Wall is preparing the EXPERIMENT
READY         --> The STORAGE is mounted on its linked COMPUTE
*TERMINATING [2]  --> The EXPERIMENT is shutting down. STORAGES with ``persistent=YES`` are finalized
*TERMINATED    --> The STORAGE is removed from the system and not accessible anymore