Description of problem: One of my compute nodes is never receiving new VMs. The running_vms field has not been updated since several VMs have failed on this node. This should be updated regularly, not when VMs are started and terminated. Version-Release number of selected component (if applicable): Essex How reproducible: always Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
I am not sure I understand what this bug is about. Resource updates (updatest to the in Essex are done through notifications and the capacity_notifier needs to be set in the notification_driver flag in the nova.conf as it is not used by default. Folsom does not have this problem as the resource tracker has been implemented at that time and the tables are updated correctly on every action. I will close it as not a bug.
What's the correct value to set notification_driver to? The default is nova.notifier.no_op_notifier. Thanks
The value of the notification_driver if you want the compute_nodes db table to be updated should be: nova.notifier.capacity_notifier This makes sense only in Essex - this does not exist in Folsom anymore as the updates are done automatically.