HubStor Inc.

Overview of Stubbing Options


Stubbing is the practice of moving the original data from the source and replacing it with a smaller shortcut or placeholder which acts as a pointer to assist with recalling the original item.

Stubbing is controlled with policies within each connector that is setup in the HubStor Connector Service. It is common that older and larger items will be stubbed. For example, a stubbing policy might specify that video and audio file types that have not been accessed in over 12 months are in scope of stubbing.

HubStor provides three stubbing methods:
  1. Seamless stubs (aka. invisible stubbing/ghosting) -- Seamless stubbing is best suited to Windows Server environments and provides a transparent experience to users and applications. Items that are stubbed become offline files where their size on disk is near zero. The seamless stubs have the same metadata and file type icon as the original, and they automatically retrieve on any file open command whether that be coming from a user or application. When items retrieve, there is an optional local caching feature that optimizes subsequent recalls (i.e. the retrieval agent always checks the local cache instead and only fetches from the cloud if necessary). Seamless stubbing requires the 'Retrieval Service', a Windows Server agent, to be installed on each instance of Windows Server that may handle requests on seamless stubs. Thus seamless stubbing is available only to storage that is front-ended with Windows Server file servers and SharePoint Documents(Lists not supported). See OVERVIEW OF SEAMLESS STUBBING AND THE RETRIEVAL SERVICE.
  1. Link-based stubs (aka. seamless stubbing for CIFS/NFS) -- Link-based stubs work with any storage (Windows, CIFS, NFS) to provide a near-seamless stubbing method for CIFS and NFS-based storage. CIFS-based storage, such as NetApp or EMC Celerra, is highly scalable but presents a lock-in challenge to customers. Traditional file-level storage tiering methods for these storage appliances has involved either complex network-layer intercepts, client software, or vendor specific agent integrations. With HubStor's link-based stubbing, you can perform simple and effective policy-based storage tiering without the headaches of older methods. To achieve this implementation, a central Windows machine needs to host the same 'Retrieval Service' that is used for seamless stubs. This service will act as the controller for servicing recalls. A stub share directory must be created which will contain information necessary to recall from the cloud. Next, you create stubbing policies in connectors within the HubStor Connector Service as you normally would. In the CIFS/NFS directory, stubbed items appears as shortcuts. When a user or application opens the file, the request is handled by the central Retrieval Service. See OVERVIEW OF LINK-BASED STUBBING.
  1. HTML stubs (aka. universal stubs) -- HTML stubs provide an agentless approach to stubbing that can work with any storage (Windows, CIFS, NFS). However, HTML stubs are not transparent to users and they do not automatically retrieve for applications or processes. The HTML stub is an HTML file that replaces the original item. When a user clicks the HTML stub, the HTML content displays in the user's default browser with a message saying the item has been archived and offering links to download it or view it in the HubStor User Portal. See HOW HTML-BASED FILE SYSTEM STUBS WORK.

    • Related Articles

    • Overview of Link-based Stubbing

      HubStor's link-based stubbing method is intended for use with CIFS/NFS-based storage, such as NetApp or EMC Celerra. It can also work with Windows-based file shares. With HubStor's link-based stubbing, you can perform simple and effective ...
    • Overview of Seamless Stubbing and the Retrieval Service

      The Retrieval Service is a Windows Server agent that must be deployed on all file servers front-ending the storage you plan to tier to the cloud with the 'seamless stubbing' option. With the Retrieval Service installed, connectors can be configured ...
    • How to Install and Configure Link-based Stubbing

      This article describes the steps necessary to configure link-based stubbing. Below are the high-level steps involved: Select a Windows Server to act as the Retrieval Server.  This server can an existing Windows Server or a provisioned virtual ...
    • Security of Link-based Stubbing and Stub Share Permission

      This article discusses the security aspects of the link-based stubbing feature. The 'Everyone' Permission on the Stub Share You will need to manually add the 'Everyone' group to the stub share with read/write permission, but then remove the NTFS ...
    • What permissions does the service account of the Connector Service require for seamless stubbing?

      If your Connector Service is configured with file connectors that will create seamless stubs, the service account of the Connector Service will need to have local machine admin rights on the target file servers where seamless stubs will be created.