Qumulo, Inc.

Michael Anthony Chmiel

Kevin David Jamieson

Aaron Passey

Thomas Gregory Rothschilds

Yuxi Bai

Jonathan Michael MacLaren

Sihang Su

Matthew Christopher McMullan

Tyler Morrison Moody

Pathirat Kosakanchit

Rowan Arthur Phipps

Kevin Ross O'Neill

Duncan Robert Fairbanks

Junjie Liang

Stephen Craig Fleischman

Nicholas Graeme Williams

Neal Thomas Fachan

Steven Henry Haber

Ying Fairweather

Graham Edwin Ellis

Thorne Davis Garvin

Ezra Elias Kilty Cooper

Daniel Marcos Motles

Embodiments are directed to a file system that include object stores. An object store for write requests may be provided. Write ahead log (WAL) entries that include data blocks may be generated. A WAL object may be generated based on the WAL entries and stored in the object store. An in-memory overlay may be updated to associate the data blocks with the WAL object. A checkpoint operation may be executed to: generate an index object that includes index entries that associate other data blocks with data objects stored in the object store; update the index object to include index entries that associate the data blocks with the WAL object; store the updated index object in the object store; update the in-memory overlay to remove the association of the data blocks and the WAL object and update the in memory WAL to remove records of successfully checkpointed WAL objects.

Object-based storage with garbage collection and data consolidation

Embodiments are directed to tiered data store with persistent layers. A write tier in the file system for storing in a file system. A value for a performance metric that corresponds to write requests to the file system may be predicted based on characteristics of the write requests such that the performance metric may be determined based on a plurality of interactions with the write tier. The predicted value that exceeds a threshold value of the performance metric may be employed to cause performance of further actions, including: queuing a portion of the write requests in a memory buffer based on the predicted value and the threshold value; combining the queued portion of the write requests into s; storing the data segments in the write tier such that a measured value of the performance metric may be less than the threshold value.

Tiered data storage with ephemeral and persistent tiers

Embodiments are directed to managing data in a file system that includes a plurality of storage nodes and a plurality of storage volumes in a cloud computing environment. Metrics associated with each storage volume may be monitored. In response to the metrics exceeding a threshold value, performing further actions, including: determining storage volumes that are unhealthy based on the metrics that exceed the threshold value; updating metadata associated with the storage volumes to indicate that the storage volumes are unhealthy; decoupling the unhealthy storage volumes from storage nodes coupled to the unhealthy storage volumes; determining replacement storage volumes based on the metadata associated with the unhealthy storage volumes; updating other metadata associated with the replacement storage volumes to indicate that the replacement storage volumes are healthy storage volumes; and coupling the healthy storage volumes with the storage nodes that were coupled to the unhealthy storage volumes.

Managing cloud storage for distributed file systems

Embodiments are directed to managing data in a file system that includes storage nodes and storage volumes in a cloud computing environment (CCE). A backup job may be executed to perform further actions, including: providing a backup identifier that is associated with the backup job; determining the storage nodes based on the file system and the CCE; disabling the file system by disabling the storage nodes, wherein each disabled storage node is associated with a shutdown timestamp; determining storage volumes associated with each disabled storage node; and generating storage images that archive a copy of information from the storage volumes; associating the storage images with the backup identifier; confirming that the file system is consistent based on an absence of file system activity occurring subsequent to each shutdown timestamp; and enabling the file system by enabling the storage nodes.

Backup services for distributed file systems in cloud computing environments

Embodiments are directed to managing data in a file system over a network. A source file system that includes a plurality of objects may be provided. A replication job that copies each object associated with a source replication snapshot to a target file system may be executed. The replication job may be associated with a job identifier. Recovery point information that includes the job identifier, a source snapshot number that corresponds to the source replication snapshot, a target snapshot number that corresponds to a target replication snapshot may be generated. The recovery point information may be stored on the source file system and a copy of the recovery point information may be stored on the target file system. The recovery point information or the copy of the recovery point information may be employed to recover from errors detected during execution of a next replication job.

Recovery checkpoints for distributed file systems

Embodiments are directed to a replication engine that provides a root node for a file that may be associated with a replication snapshot where the file is comprised of the root node, internal nodes, and data nodes. The replication engine may start at the root node and traverse the file system. If a visited internal node has an epoch value that matches the epoch, the replication engine may continue the traversal by visiting a next unvisited adjacent internal node. If the visited internal node has an epoch value that mismatches the epoch, the replication engine may backtrack to a nearest adjacent internal node that matches the epoch. And, in response to visiting a data node the replication engine may be arranged to determine blocks that are associated with the data node and communicate a copy of the blocks to a target file system.

Data replication in distributed file systems

Embodiments are directed to managing data in a file system. A source file system may be mirrored on a target file system by executing one or more replication jobs. A root object of the source file system may be associated with a current read-only identifier corresponds to a current update epoch of the source file system. A replication job that copies objects from the source file system to the target file system may be executed. A write request directed to the source file system may be received. Write objects associated with the write request may be determined The source file system may be traversed upwards from the write objects. A read-only identifier associated with each visited object may be determined based on the read-only identifier matching the current read-only identifier. The write request may be enabled based on the mismatch of the read-only identifier and current read-only identifier.

Managing file system state during replication jobs

Embodiments are directed to a file system engine that provides a file system with parent objects associated with child objects. The file system engine provides a replication snapshot associated with an epoch of the file system such that each child object modified during the epoch and each associated parent object is associated with the replication snapshot. A root object of a portion of the file system included in the replication snapshot may be provided. A replication engine may traverse the portion of file system starting from the root object such that the parent objects that are not associated with the replication snapshot are omitted from the traversal. The replication engine may determine replication objects based on the traversal such that each replication object was modified during the epoch. Then the replication engine may execute a replication job that copies replication objects to a target file system.

Replicating file system objects in distributed file systems

Embodiments are directed to file systems. A replication engine may establish a secure communication channel between a source file system and a target file system. The replication engine may: instantiate a replication job associated with rules; determine changes in the source file system; determine characteristics of the replication job that may be based on the changes; compare the to the characteristics and a black-out schedule; execute the replication job to communicate the changes in the source file system to the target file system based on a result of the comparison. Upon completion of the replication job, the replication engine may automatically instantiating a next replication job to copy subsequent changes in the source file system to the target file system.

Growing community of inventors

Seattle, WA, United States of America

Michael Anthony Chmiel