PowerScale OneFS 9.5.0.0 API Reference Guide

Data Mover (SmartSync) overview

Data Mover (SmartSync) enables you to push, pull and proxy data movement operations between S3/Object and NFS interface from public cloud with AWS, GCP, Azure and ECS.

It also provides backup capabilities to provide incremental and full fidelity backup and restore operations. Data Mover provides fast, accurate and low-risk data replication for Dell Technologies file and object deployment on-premises or in the cloud. Protect your business critical data from ransom, corruption, and human error (accidental deletion). You can keep a pristine copy of your unstructured data anywhere you decide. You can get an end-to-end, automated, heterogeneous, multiprotocol solution designed to safely, accurately and easily replicate and restore your data.

Data Mover (SmartSync) enables you to prioritize data copy jobs for OneFS. You can set the High Priority, Normal, Low to control the amount of cycles a job gets in respect to with other jobs. This only applies to prior to the job start, not changing priority on the fly. If any attribute of a policy is updated while a job for the policy is running, the next job will exhibit the updated attribute.

If a file cannot be completely copied, the Data Mover logs the file and continue with the job. When the job reaches the end of work it will AUTO-PAUSE for admin action to either RETRY or mark the job PARTIAL-COMPLETE. If PARTIAL-COMPLETE the dataset will be annotated and the files will be retried on the next job if it is a REPEAT-COPY policy. If the number of failures exceeds 100 or a critical error is hit, the Data Mover shall AUTO-PAUSE and provide the Admin the option to RETRY or CANCEL the job. If cancelled the data will be invalid.

You can pause, resume or cancel a job in progress. A cancel aborts the running job and remove the movement hence forcing a re-baseline.

Through Data Mover (SmartSync) you can copy a source dataset to multiple targets. You can also create a copy policy from a target dataset. A chain is when source is A for a dataset and that is copied to B, which in turn is copied to C. A chain shall at a minimum go 4 copies deep, so A > B > C > D > E.