Red Hat, Inc.

Steven John Whitehouse

David Alan Gilbert

Vivek Goyal

Stefan Hajnoczi

Miklos Szeredi

Secure and efficient access to host memory for guests is disclosed. For example, a system includes (i) a processor, (ii) a host memory, (iii) a hypervisor including a filesystem daemon, (iv) a guest including a storage controller, and (v) a first filesystem queue. The storage controller is configured to receive a file request, translate the file request, and add the translated file request to a filesystem queue. The filesystem daemon is configured to retrieve the translated file request from the filesystem queue and provide access to a file in the host memory identified in the file request to the guest.

Secure and efficient access to host memory for guests

Direct access to host memory for guests is disclosed. For example, a system includes a processor, a host memory, a filesystem daemon, a guest including a storage controller, and a filesystem queue accessible to the filesystem daemon and the storage controller. The storage controller receives a file retrieval request associated with a file stored in the host memory and forwards the file retrieval request to the filesystem daemon by adding the file retrieval request to the filesystem queue. The filesystem daemon retrieves the file retrieval request from the filesystem queue, determines a host memory address (HMA) associated with the file, and causes the HMA to be mapped to a guest memory address (GMA). The guest accesses the file in the host memory with the GMA, and later terminates access to the file, where the filesystem daemon is then configured cause the GMA to be unmapped.

Direct access to host memory for guests

A system and method for reading file blocks includes reading an inode associated with the file from the file system, the inode including one or more first block pointers, determining a height of a file tree associated with the file, and determining whether a value of a second block pointer selected from the one or more first block pointers is consistent with the file having been stored using a block allocation pattern. When the value of the second block pointer is consistent with the file having been stored using the block allocation pattern the method further includes pre-fetching a plurality of file blocks based on the block allocation pattern, verifying that the pre-fetched file blocks are consistent with the file tree, and retrieving one or more data blocks of the file. In some examples, the block allocation pattern corresponds to the file being stored in streaming order to consecutively and contiguously located blocks.

System and method for reading file blocks

A system and method are disclosed for detecting and handling resource contention in a cluster file system. In one implementation, a processing device determines a measure of congestion for a resource that is shared by a first node of a cluster file system and a second node of the cluster file system, where the first node has a first local queue for lock requests for the resource, and wherein the second node has a second local queue for lock requests for the resource. The processing device adjusts a parameter for a node of the cluster file system in view of the measure of congestion.

Distributed resource contention detection and handling

Methods for encoding file system metadata are described herein. According to one embodiment, a file system cache is maintained including information representing relationships between inodes and disk blocks of a disk having disk sections. Each disk section includes a data segment and a header encoding metadata for describing the data section of each disk section. The metadata is encoded using a set partitioning algorithm and each set represents a set of disk blocks. In response to a file system request for reading a disk section, metadata associated with the disk section is retrieved and decoded to extract information representing a relationship between inodes and disk blocks associated with the requested disk section. The file system request is then serviced using the decoded metadata and the associated data segment and one or more entries of the file system cache are updated using the decoded metadata.

Set partitioning for encoding file system allocation metadata

Some embodiments of a system and a method to detect contention for resource in a cluster file system have been presented. For instance, a processing device executable on a node in a cluster file system may put lock requests for a resource from the node into a local lock request queue, where multiple nodes in the cluster file system share the resource. Then the lock requests for the resource from this node and lock requests for the resource from other nodes may be tracked to measure congestion on the resource.

Distributed resource contention detection

A filesystem checker identifies a metadata block in a filesystem and determines a number of pointers pointing to the metadata block and a number of pointers embedded in the metadata block. The filesystem checker records the number of pointers pointing to the metadata block and the number of pointers embedded in the metadata block in a filesystem checker array. The filesystem checker verifies a consistency of the filesystem using data recorded in the filesystem checker array.

Linear sweep filesystem checking

A mechanism for inode event notification for cluster file systems is disclosed. A method of the invention includes implementing a lock scheme for a cache manager associated with an inode of a cluster file system, wherein the lock scheme includes a null lock (NL) used as a reference count for a lock value block (LVB) and an exclusive (EX) lock used for setting up inode event notifications, and associating a first global counter and a second global counter with the cache manager for the inode, wherein the first global counter is configured to track one or more processes that have the inode open in read-only mode and the second global counter is configured to track processes that have the inode open in read-write mode.

Inode event notification for cluster file systems

Some embodiments of a system and a method to detect contention for resource in a cluster file system have been presented. For instance, a processing device may measure time spent performing actual operations by each of a set of nodes in a cluster file system when a respective node holds a lock on a resource and time spent performing overhead operations by the set of nodes without synchronization of clocks across the cluster file system. Then the processing device can determine latency and throughput of the cluster file system based on the time spent performing actual operations and the time spent performing overhead operations.

Management of latency and throughput in a cluster file system

A method of and system for managing storage resources in a distributed file system is described. A lock for a storage resource is maintained on a lock-holding node. A master node that controls the lock-holding node receives a lock request from a requesting node. The lock request includes a request to obtain a lock for the storage resource, and a request to perform an action by the lock-holding node on the storage resource if the request to obtain the lock is not granted immediately.

Growing community of inventors

Swansea, United Kingdom

Steven John Whitehouse