Optimizing placement plans for host failure in cloud computing environments
Embodiments of the present invention provide systems, methods, and computer program products for optimizing a placement plan. In one embodiment, a method is disclosed in which a request for registration with an external advisor is received. A time to live is received from each external advisor and used to determine an overall timeout period value for a placement engine. After receiving a predictive failure alert, internal and external advisors are ranked according to criteria and advice is received from the qualified advisors. A placement plan is generated based on the advice received from the advisors.
1. A method for optimizing a placement plan, the method comprising:
receiving, by one or more processors, a request for registration from an advisor;
determining, by one or more processors, an initial timeout period value associated with the advisor;
determining, by one or more processors, an overall timeout period value for a placement engine based, at least in part, on the initial timeout period value associated with the advisor and a time to live value received from the advisor;
responsive to receiving a predictive failure alert for a first host in a first pool, qualifying the advisor and receiving advice from the advisor, wherein qualifying the advisor comprises: prioritizing the advisor based, at least in part, on the initial timeout period value associated with the advisor, a policy type of the advisor, and whether the advisor is an internal or external advisor; and
generating a placement plan for migrating a virtual machine from the first host based, at least in part, on the received advice from the advisor.
2. The method of claim 1 , wherein generating a placement plan for migrating a virtual machine from the first host based, at least in part, on the received advice from the advisor comprises:
determining, by one or more processors, whether there is an available host in the first pool to which to migrate the virtual machine;
responsive to determining that there is an available host in the first pool to which to migrate the virtual machine, generating a placement plan that instructs a placement engine to migrate the virtual machine to the available host in the first pool; and
responsive to determining that there is not an available host in the first pool to which to migrate the virtual machine, generating a placement plan that instructs the placement engine to migrate the virtual machine to an available host in a second pool.
3. The method of claim 2 , further comprising:
responsive to migrating the virtual machine to the available host in the second pool, updating, by one or more processors, an optimization goal of the second pool, based on a first optimization goal of the first pool and a second optimization goal of the second pool.
4. The method of claim 1 , further comprising:
storing, by one or more processors, advice data, wherein the advice data comprises the received advice from the advisor and the time to live value received from the advisor; and
responsive to identifying advice with an expired time to live value at a predetermined time interval, removing, by one or more processors, the advice with the expired time to live value at the predetermined time interval.
5. The method of claim 1 , where the step of determining, by the one or more processors, the overall timeout period value for the placement engine based, at least in part, on the initial timeout period value associated with the advisor and the time to live value received from the advisor occurs prior to receiving the predictive failure alert.