摘要:
Systems and methods for replicating input-output (I/O) events of a cluster shared volume (CSV) are presented. Such systems may include a cluster device or node (in a cluster of devices), a coordinator node, and a replica server. Replication techniques disclosed herein may include capturing, using a processor operatively associated with a cluster device, a direct I/O event generated at the cluster device (e.g., by a virtual machine). A legacy file system filter driver of the cluster device may be used and enabled to perform the capturing of the direct I/O event. The legacy file system filter driver may be different than a CSV filter driver associated with the CSV, and may be enabled after the CSV filter driver is enabled. The captured direct I/O event may then be transmitted to the replica server using, e.g., a replication filter, to maintain a backup copy of the event at the replica server.
摘要:
A resource management node is provided including a processor and a memory. The memory is coupled to the processor and comprises computer readable program code that when executed by the processor causes the processor to perform operations including creating a recovery virtual machine (VM) for a primary machine for each of a primary machines. The recovery VM is stored in a cache file in a kernel of an operating system. The data in the cache file is stored in a format readable by a virtual hypervisor. The processor is further caused to perform operations including providing the recovery VM from the cache file in the kernel of the operating system upon request when the primary machine fails without conversion of the data in the cache file.
摘要:
A system, method, and computer-readable storage medium having a computer-readable instruction thereon to replicate transactions in a Zettabyte File System (ZFS). The method includes operations including determining, using a tracing module, a commitment of a current file transaction to a disk, obtaining when a probe function associated with the current file transaction is triggered, the current file transaction, recording a data change contained in the current file transaction, registering a callback function and a reference to the data change contained in the current file transaction after the commitment of the current file transaction, and sending the data change to a remote server.
摘要:
A resource management node is provided including a processor and a memory. The memory is coupled to the processor and comprises computer readable program code that when executed by the processor causes the processor to perform operations including creating a recovery virtual machine (VM) for a primary machine for each of a primary machines. The recovery VM is stored in a cache file in a kernel of an operating system. The data in the cache file is stored in a format readable by a virtual hypervisor. The processor is further caused to perform operations including providing the recovery VM from the cache file in the kernel of the operating system upon request when the primary machine fails without conversion of the data in the cache file.
摘要:
The system and method described herein may support multiple-layer data replication using a block input/output (“BIO”) structure that an operating system kernel employs to provide an intermediate mechanism between file systems and block devices. For example, the system and method described herein may capture events that occur when the kernel uses the BIO structure to encapsulate input/output requests that will change data on a master block device. As such, a replication mechanism may send the changes encapsulated within the BIO structure to a replay mechanism, which may then replicate the changes on a standby block device to ensure that the master block device and the standby block device have an identical state. Alternatively, a filter may be used to only replicate changes to certain partitions or file system objects on the master block device.
摘要:
Systems and methods for replicating input-output (I/O) events of a cluster shared volume (CSV) are presented. Such systems may include a cluster device or node (in a cluster of devices), a coordinator node, and a replica server. Replication techniques disclosed herein may include capturing, using a processor operatively associated with a cluster device, a direct I/O event generated at the cluster device (e.g., by a virtual machine). A legacy file system filter driver of the cluster device may be used and enabled to perform the capturing of the direct I/O event. The legacy file system filter driver may be different than a CSV filter driver associated with the CSV, and may be enabled after the CSV filter driver is enabled. The captured direct I/O event may then be transmitted to the replica server using, e.g., a replication filter, to maintain a backup copy of the event at the replica server.