You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm thinking of a new feature for IM: the possibility to request the virtual machines of a certain type of node are running in different physical nodes.
Suppose the following scenario: We have a distributed storage system well-configured with redundancy and deployed in HA. If all the VMs that compose the storage infrastructure are running in the same physical node and this node goes down, the storage will be lost. Thus, although the administrator configured the storage with redundancy, the storage can be lost.
Maybe a solution can be implemented as a new field in system specification at the RADL to indicate that the VMs of that type of node must be allocated in different (if it is possible) physical nodes. From the technical point of view, we can check the node affinity policies (https://docs.openstack.org/senlin/pike/scenarios/affinity.html).
The text was updated successfully, but these errors were encountered:
I'm thinking of a new feature for IM: the possibility to request the virtual machines of a certain type of node are running in different physical nodes.
Suppose the following scenario: We have a distributed storage system well-configured with redundancy and deployed in HA. If all the VMs that compose the storage infrastructure are running in the same physical node and this node goes down, the storage will be lost. Thus, although the administrator configured the storage with redundancy, the storage can be lost.
Maybe a solution can be implemented as a new field in system specification at the RADL to indicate that the VMs of that type of node must be allocated in different (if it is possible) physical nodes. From the technical point of view, we can check the node affinity policies (https://docs.openstack.org/senlin/pike/scenarios/affinity.html).
The text was updated successfully, but these errors were encountered: