This article provides an overview of Grid DS-Client.

Data storage protection is a fast growing market. More and more data needs to be backed up and the backup jobs / sets are getting bigger and bigger. Meanwhile, backup time windows are shrinking and RTO (recovery time objective) is vital. This means multiple DS‑Clients for a single LAN are becoming commonplace. Automatic balancing between multiple DS‑Clients for higher performance becomes a new challenge. High availability becomes more important. The Agentless Grid DS‑Client is the solution to achieve these requirements and to meet customer Service Level Agreements.

Grid DS‑Client is designed to deliver better performance scalability of DS‑Client and to increase the availability of the Backup Service. It is designed so that DS‑Client can survive failures of some of its nodes without interrupting the backup service.

Configuring Grid DS‑Client means that several DS‑Clients will work together to provide backup and restore services as if they are a single DS‑Client (from the DS‑System perspective). Any node in the Grid DS‑Client can provide the same service (backup, restore, delete, synchronization, admin, etc.) as long as the nodes are configured properly (see “Hardware and software setup”).

All nodes in a Grid DS‑Client share the same DS‑Client database. They also share the same Account Number, DS‑Client number and private/account encryption keys.

Each node in the Grid DS‑Client has a separate network connection to DS‑System.

One of the nodes will be selected as the Main Node to perform common internal tasks and to synchronize and balance activities. It also interacts with DS‑User GUI. Any other node is called a “Leaf Node”. Leaf nodes will perform tasks assigned by the Main node.

Activities are controlled by the Main Node and assigned to all nodes in a balanced manner. The balance is measured by the load (determined by the activities assigned, compared to the “MaxSession” parameter configured for that node) and CPU usage.

If the Main Node goes down, the other (leaf) nodes will continue their current tasks. In the meantime, a new Main Node will be elected.

Even if only one node is alive and all other nodes are down, the Grid DS‑Client will still be available and will continue providing the same backup services.

We recommend that you run all nodes on the same platform and Windows version. For information on supported operating systems, see the Client Software Installation Guide.

For each node to be able to provide the same backup services, they must have the same access to the source data. This means that the nodes must be configured the same way with regards to network credentials:

All nodes must also be able to access that source data with that same network credential.

Moreover, if one node has software installed to access a specific data source type (e.g. “Oracle Client” to backup Oracle Server, “iSeries Access for Windows” to backup System i Servers, etc.) all other nodes must also be configured with the same capability.

Each node must be able to access the DS‑Client database using a trusted connection. This means that a central database must be configured for the Grid DS‑Client to access. It is recommended that the database be clustered to avoid single point of failures.

All nodes (including the DS‑Client database machine) must be in the same domain to achieve the above two requirements.

(Not applicable to Cloud-to-Cloud DS-Client) Grid DS‑Client is supported on VMware. Multiple nodes can be configured to run on virtual machines on the same physical machine. Since all these nodes share the same CPU/RAM on the physical machine, if some nodes are idle or down, other nodes will get more CPU/RAM resources and will become more powerful.

Database configuration is the same as for stand‑alone DS‑Clients:

Each node’s service account should have enough credentials to remotely start/stop the DS‑Client service on all other nodes. This is required for the “Auto upgrade” feature. In addition, perform the following:

All nodes (including the DS‑Client database machine) should be synchronized to the same time server and configured to the same time zone to ensure time‑consistency. Otherwise, confusing (or even failure) scenarios may occur regarding logs and backups.

To configure a Grid DS-Client, you must prepare to fulfill these requirements:

For information on advanced configuration for a Grid DS-Client, see “Configuring the Grid DS-Client nodes (node level configuration)”.

The following procedure assumes that all the existing DS-Clients with which you plan to configure a new Grid DS-Client are stand-alone DS-Clients. Each stand-alone DS-Client is connected to its own database instance, which is not shared by other DS-Clients.

For information on advanced configuration for a Grid DS-Client, see “Configuring the Grid DS-Client nodes (node level configuration)”.

You may add a new node, remove an existing node, upgrade a stand‑alone DS-Client to become a new node in an existing Grid DS-Client, and downgrade an existing node of a Grid DS-Client to become a stand‑alone DS-Client.

You can upgrade an existing stand-alone DS-Client or add a newly installed DS-Client to become a new node in an existing Grid DS-Client.

DS‑User can connect only to the Main Node.

Each node will connect to DS‑System using a separate network connection, using its own external IP.

The default path for Grid DS‑Client Configuration Tool is C:\Program Files\CloudBackup\DS-Client\DSConfig.exe.

User input and modification of nodes is controlled by the Add, Update and Delete buttons.

In general, the DS‑User functionalities are the same as for the stand‑alone DS‑Client, with a few additional functionalities for managing the Grid DS‑Client. For details, see “About Grid DS-Client”.

To manually upgrade a Grid DS‑Client to a new version:

Grid DS‑Clients can auto-upgrade if their DS-System is configured to allow it.

The first node that downloads a higher version auto‑upgrade package from DS‑System will stop the DS‑Client service on all other nodes. After the first node has finished upgrading, it will restart the DS‑Client services on all other nodes. Once those nodes restart, they will connect to DS‑System and download the auto‑upgrade package (this time auto‑upgrade will only stop the DS‑Client service on the local node).

Auto‑upgrade will be triggered if the following occurs:

When starting DS‑Client service, at least one other node is found to have a higher version (only the node performing upgrading will be stopped for upgrading and restarted automatically).

When connecting to DS‑System, a higher version DS‑Client auto‑upgrade package is available.

The Recover DS‑Client database process can be performed on any node (Classic DS-User > Setup > System Activities: Repair). Since the DS‑Client Database Recovery needs exclusive access to the database, all nodes are required to stop for recovery. After the node that performed recovery successfully downloads the database dump, it will notify all other nodes to stop. Once it detects all other nodes have stopped, it will notify the user to restart the DS‑Client service on that node. Before that node has successfully recovered the DS‑Client database (database load completed), other nodes will not be able to start.

For more information on this process, see the Knowledge Base article in “Recovering the DS-Client Database”.

If a node’s version is lower than the highest version in the Grid DS‑Client, it will perform auto‑upgrade right after starting the DS‑Client service.

If auto‑upgrade is not allowed or fails to upgrade, a version compatibility check will be performed and nodes that are not compatible will not be allowed to start.

The default configuration (created by installing and running the Grid DS‑Client Configuration Tool) is sufficient to run a Grid DS-Client. Advanced users may wish to configure individual nodes for optimum performance.

For different nodes in the same Grid, some DS‑Client parameters can be different. These local parameters are flagged Local Setting. All local parameters must be configured from the DS-Client configuration - Advanced Tab (see “Configuring the advanced settings”).

Every Local Setting parameter has a corresponding Global Setting that is the default value for each node.

A special parameter is “DSCBuffer”. This is the same as DS‑Client Buffer in the DS-Client Configuration - Parameters tab. However, it can be set to a different value for each DS‑Client node because each DS‑Client node can have its own buffer.