Scheduling VMs to servers taking into account the health of the platform.

Registered by murali sundar on 2015-02-12

This blueprint is about paying attention to the health of the server and taking that into account at various layers in OpenStack, such as when the scheduler selects a server as a target for running a VM. Servers may be in various states of disrepair such as permanent subsystem failure (fan failure), temporarily degraded (too hot)etc. For example, if a server has a voltage-sensor or thermal-sensor anomaly, it is probably not the best target for more workloads until the problem can be assessed. Similarly, if a server has suffered a HDD failure and is consequently performing a RAID rebuild, its available IOPs for production workload may be degraded for a few hours until it gets back into a healthy state. This blueprint is about creating a general framework to model these states, detect these conditions, account for these conditions in a systematic way, and to expose it to upper level SW such as the scheduler.

Blueprint information

Status:
Not started
Approver:
None
Priority:
Undefined
Drafter:
murali sundar
Direction:
Needs approval
Assignee:
Andrzej Kuriata
Definition:
New
Series goal:
None
Implementation:
Unknown
Milestone target:
None

Related branches

Sprints

Whiteboard

Gerrit topic: https://review.openstack.org/#q,topic:bp/is,n,z

Addressed by: https://review.openstack.org/174764
    Scheduling VMs to servers taking into account the health of the platform.

(?)

Work Items

This blueprint contains Public information 
Everyone can see this information.