Abstract: The technologies described herein are generally directed toward increasing a storage capacity of a storage cluster. In an embodiment, a method can include identifying a cluster for a change in storage capacity, the cluster having been deployed using a set of node devices that support the cluster. The method can further include, based on a capacity specification, deploying a virtual machine on a node device that is not part of the set of node devices. Further, the method can include, based on a capacity specification, merging the node device into the set of node devices, resulting in a merged set of node devices to support the cluster and achieve the change in storage capacity.

Abstract: The technologies described herein are generally directed toward migrating a function between virtual machines. In an embodiment, a method can include identifying a cluster for scaling, deployed using a first node device of a set of node devices supporting the cluster, wherein the first node device has a first virtual machine deployed thereon to support a function of functions of the cluster. The method can further include, based on a scaling specification, deploying, by the system, a second virtual machine on a second node device. Finally, the method can include, based on the scaling specification, migrating, by the system, the function, from the first virtual machine to the second virtual machine.

Abstract: Systems, methods, and computer readable medium for optimizing storage allocations based on system resources are disclosed. According to one aspect, the subject matter described herein includes a method for dynamic, policy-based allocation of system resources. The method includes specifying a plurality of system resource collections for representing sets of like system resources and their dependencies. An order in which the collections are to be evaluated during selection of a resource is also specified. A policy for determining the best resource within each collection is also specified. Statistics about the system resources in the collections are maintained. A request for a resource is received. In response to receiving the request, the collections are evaluated in the specified order. The best resource is selected from each collection based on the dependencies, prior selections, and statistics to implement the policy.

Abstract: In accordance with a method and system described herein, in order to map the physical location and connections between all of the components of a network from the perspective of a particular host, i.e., discover the fabric of the network, and the hierarchical relationship between storage disk array components including chassis IDs, storage processors, and storage processor ports, queries are generated and sent to the different operating systems or software that are implemented on the components, i.e., hardware devices of the network. These queries ask the device what it sees. Starting with identifiable starting points and ending points in the fabric, the host selectively queries components within the fabric. Using the query data the host applies a series of comparison and process of eliminations steps, in order to complete a table of the components and connections that comprise the network fabric.

Abstract: The method and system provides for continuing issuance of input/output requests on a fabric having a plurality of paths connected to adapters on a computer managing the input/output requests. If a change in a path within the fabric occurs, a registered state change notification is typically issued on the fabric which is detected. The fabric on which the input/output request was issued is searched for a path failure, and if a failure is detected, a path is recovered.

Abstract: A method and system for recovering nodes in a hierarchical computer network is disclosed. The operating system of the network detects and exception condition, recovers only the nodes within the scope of the exception condition, and issues I/O requests to nodes during recovery that are not within the scope of the exception condition. The method may involve a computer network system with a plurality of paths connected to adapters on at least one host computer for managing input/output (I/O) requests between the host computer and fibre channel devices, which have a plurality of logical units (LUNs) associated therewith.

Abstract: In accordance with a method and system described herein, in order to map the physical location and connections between all of the components of a network from the perspective of a particular host, i.e., discover the fabric of the network, and the hierarchical relationship between storage disk array components including chassis IDs, storage processors, and storage processor ports, queries are generated and sent to the different operating systems or software that are implemented on the components, i.e., hardware devices of the network. These queries ask the device what it sees. Starting with identifiable starting points and ending points in the fabric, the host selectively queries components within the fabric. Using the query data the host applies a series of comparison and process of eliminations steps, in order to complete a table of the components and connections that comprise the network fabric.

Abstract: A host system, such as a server, controls a plurality of different class storage devices on a network. The operating system on the host assigns a unique identification to each storage device using a standard protocol and maintains the identification in the operating system. This allows control of the storage devices without complications even though different classes of devices, using differing identification protocols, are assembled on the same network.