watcher

compute_model problem

Asked by suzhengwei

I create a continus audit used vm_workload_consolidation strategy, and find some problems.
1.First audit, self._compute_model is none, and it will build self._compute_model use watcher_cluster_data_model_collectors. The next audit, the self._compute_model is still none, which make me confused.
2.Once a compute node is set disabled, it will be removed from self._compute_model, because scoped_model only return enable node. There may be some problem with ServerConsolidation Strategy. If all the enabled compute nodes are overloaded, but the instance can't migrate to disabled nodes.

Question information

Language:
English Edit question
Status:
Open
For:
watcher Edit question
Assignee:
No assignee Edit question
Last query:
Last reply:

This question was originally filed as bug #1685787.

Revision history for this message
suzhengwei (sue.sam) said : 		#1

I create a continus audit used vm_workload_consolidation strategy, and find some problems.
1.First audit, self._compute_model is none, and it will build self._compute_model use watcher_cluster_data_model_collectors. The next audit, the self._compute_model is still none, which make me confused.
2.Once a compute node is set disabled, it will never be removed from self._compute_model, because scoped_model only return enable node. There may be some problem with ServerConsolidation Strategy. If all the enabled compute nodes are overloaded, but the instance can't migrate to disabled nodes.

Can you help with this problem?

Provide an answer of your own, or ask suzhengwei for more information if necessary.

To post a message you must log in.