Red Hat Israel, Ltd.

Red Hat, Inc.

Maor Lipchuk

Arik Hadas

Daniel Erez

Omer Frenkel

Livnat Peer

An apparatus includes a storage device of a host computing device. The storage device is to store a virtualization manager. The apparatus also includes a processing device of the host computing device and operatively coupled to the storage device. The processing device is to determine that a first instance of a virtual machine on a first host computing device is paused based on an error associated with a connection to a storage device of the first host computing device, determine whether the second host computing device has access to the storage device of the first host computing device, instantiate a second instance of the virtual machine on the second host computing device when the second host computing device is determined to have access to the storage device of the first host computing device, and to stop the first instance of the virtual machine on the first host computing device.

Automated restart of paused virtual machines due to input/output errors

Virtual machine synchronization and recovery is disclosed. At a first instance in time, a first snapshot request to generate a first snapshot of a virtual machine (VM) is received. The VM utilizes a storage unit. A first snapshot layer is generated for the storage unit. The first snapshot layer preserves a state of the storage unit at the first instance in time. A first VM record that corresponds to the first snapshot request and that includes a first VM record identifier (ID) is generated. The first VM record ID is stored in metadata associated with the first snapshot layer, and the first VM record is stored in a first location.

Virtual machine synchronization and recovery

A method includes identifying, by a host computing device hosting a virtual machine, a plurality of disk volumes of a virtual disk stored on a first storage device, wherein the disk volumes are arranged in a hierarchical order. The method further includes determining, by the host computing device, a block that is present in a first disk volume and a second disk volume of the plurality of disk volumes, wherein the second disk volume is at a higher level in the hierarchical order than the first disk volume. The method also includes copying, by the host computing device, data from the block stored on the first storage device to a backup disk volume residing on a second storage device and removing the block on the first storage device.

Moving outdated data from a multi-volume virtual disk to a backup storage device

In an example, a method includes copying a base state of a virtual machine from a source volume to a destination volume, the destination volume having a physical capacity and a virtual capacity, the virtual capacity being larger than the physical capacity, and including a plurality of storage blocks. The method further includes, in response to an instruction to write data to the destination volume, writing the data to at least a first storage block in the plurality of storage blocks and setting a primary flag included in metadata stored on the destination volume from a first state to a second state, where the primary flag is associated with the first storage block. The method further includes subsequently reading a state of the primary flag to determine that data in the first storage block no longer form part of the base state of the virtual machine.

Base state for thin-provisioned volumes

Mechanisms for detaching a thin-provisioned virtual machine (VM) from a VM template are disclosed. A request to detach a thin-provisioned VM from a VM template is received. The thin-provisioned VM is linked to the VM template and has at least one delta volume that is linked to a volume of the VM template. For each delta volume of the thin-provisioned VM, a corresponding merged volume is caused to be generated that is not linked to the VM template and that comprises contents of the delta volume and contents of the volume of the VM template to which the delta volume is linked. VM configuration data is altered to indicate that the thin-provisioned VM is an independent VM that is no longer linked to the VM template.

Detach virtual machine from virtual machine template

An example method for the gradual creation process of server virtual machines includes a virtualization manager locking a virtual machine template, saving a configuration of a virtual machine, locking the virtual machine, and directing a worker host to create a volume. The worker host creates the volume, and the virtualization manager unlocks the virtual machine. A destination host executes the virtual machine, and the worker host merges the volume with a disk of the virtual machine template.

Gradual creation process of server virtual machines

A method for fast evacuation of a cloned disk to a storage device includes a source storage device generating a first write volume, where new written data is stored in the first write volume. A destination storage device generates a second write volume, where the new written data is also stored. The destination storage device generates a delta volume in a cloned disk, which is also located in the destination storage device. Differences between a template volume and a modified template volume are stored to the delta volume. The destination storage device merges the template volume and the delta volume, creating a second modified template volume.

Fast evacuation of a cloned disk to a storage device

Mechanisms for snapshot management are provided. A request that requests that a first storage entity of type volume entity be generated is sent to an external storage service via an external interface. An entity chain associated with a virtual machine is generated. A first entry contains a first identifier (ID) that identifies the first storage entity, an entity type ID that identifies the first storage entity as a volume entity, and an active storage entity indicator. A first snapshot command is received. A request is sent to the external storage service via the external interface that requests that a second storage entity of type snapshot-volume entity be generated. A second entry is added to the entity chain that contains a second ID that identifies the second storage entity, an entity type ID that identifies the second storage entity as a snapshot-volume entity, and a reference to the first entry.

Snapshot management with an external storage service

Implementations of the disclosure provide for live storage domain decommissioning in a virtual environment. A method of the disclosure includes identifying, by a processing device, a storage domain to decommission from a storage system. The storage domain comprises a plurality of disk images. It is determined whether the disk images are associated with a virtual machine. Thereupon, the disk images are hot-unplugged from the virtual machine. Responsive to detecting that the disk images are unplugged from the virtual machine, the disk image is deleted from the storage domain and the storage domain is removed from the storage system.

Live storage domain decommissioning in a virtual environment

Systems and methods for managing virtual machine templates. An example method may comprise: identifying, by a computer system, a virtual machine snapshot represented by a volume chain comprising one or more copy-on-write (COW) volumes; and creating, based on the virtual machine snapshot, a virtual machine template represented by a new COW volume, wherein the new COW volume comprises data from the COW volumes of the volume chain.

Virtual machine template management

Clients access compute resources in a data center organized in logical clusters. A cluster level quota governs access to the compute resources, regardless of the location of the resources which can be part of multiple computers or part of multiple data centers. The cluster level quota can manage allocation and usage of storage, memory, and CPU resources in multi-tenant data center environments. A user requests a computing resource from a logical compute cluster in a data center of host machines hosting virtual machines. According to permissions associated with the user for the logical compute cluster and a cluster quota, the system grants access to the computing resource to the user in response to the request when the cluster quota permits an expected use of the computing resource according to the request.

Virtual resource allocation and resource and consumption management

The subject matter of this specification can be implemented in, among other things, a computer-implemented method including sending, from a virtual desktop server manager at a data center and over a network, at least one request to a virtual machine storage domain for virtual machine installation images. The virtual machine storage domain stores the virtual machine installation images separate from the data center. The method further includes receiving, from the virtual machine storage domain over the network, the virtual machine installation images. The method further includes caching the virtual machine installation images in a data storage domain within the data center. The method further includes receiving a request to present a list of the virtual machine installation images. The method further includes in response to receiving the request to present the list, presenting the list of the cached virtual machine installation images.

Growing community of inventors

Rishon Letzion, Israel

Maor Lipchuk