Instance remains in scheduling state if RabbitMQ is down during instance creation

Bug #961237 reported by Unmesh Gurjar
12
This bug affects 1 person
Affects Status Importance Assigned to Milestone
OpenStack Compute (nova)
Invalid
Low
Unassigned

Bug Description

Scenario: Stop the RabbitMQ server and try launching an instance.

Branch: master

Expected Result: The instance must go into error state.
Actual Result: Instance goes into vm_state=building, task_state=scheduling, power_state=0.

The instance state must be updated to "error" since launching the instance failed.

Tags: ntt
Changed in nova:
assignee: nobody → dhanalaxmi (baby-adabala)
Revision history for this message
dhanalaxmi (baby-adabala) wrote :

Hi,
If we stop RabbitMQ server and create a new instance, the instance is in "BUILD" state.
After starting the RabbitMQ server, instance state is changed from 'BUILD' to 'ACTIVE'.

This is tested in Essex version.

Changed in nova:
assignee: dhanalaxmi (baby-adabala) → nobody
Revision history for this message
Mandar Vaze (mandarvaze) wrote :
Sean Dague (sdague)
Changed in nova:
status: New → Confirmed
Thierry Carrez (ttx)
Changed in nova:
importance: Undecided → Low
Revision history for this message
Rohit Karajgi (rohitk) wrote :

This bug was observed in Essex. After the state management changes in Folsom, periodic tasks in Compute
manage this scenario by setting the instance state to Error.

Moreover, after RabbitMQ comes back up, the message casts onto the Scheduler and Compute, the instances gets
spawned and changes state to Active, hence this does not occur anymore.

Marking invalid.

Changed in nova:
status: Confirmed → Invalid
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Duplicates of this bug

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.