Windows File System Clustering

High Availability Clustering for Windows File and Print Server

Many businesses depend upon simple facilities like file and print share for every day operation - for example to print invoices out to customers. With only a medium sized office with a few hundred staff, it is likely to make financial sense to protect the server(s) to remove the potential for lost productivity and revenue.

To illustrate, as an example, an office has 200 staff with a central file server upon which all business related documents are stored upon for a nightly backup by the IT administrator. Incremental backups occur at midnight in the weekdays, and each Saturday a full backup is taken.

If one of the hard disks fail, the worst case scenario is that it will take a few hours to replace (assuming a spare is available within the department), and restoration will need to be undertaken from the backup archive. This could result in having to read data from four or five different tapes, which is likely to be a time consuming process.

But what if there was an exact duplicate of the data on a hot backup server? Downtime would be only minutes, productivity would be unaffected, and the primary server could be repaired when convenient.

Contact Nordicmind for solution proposal.

---

Windows File Server failover

Windows file servers are protected by monitoring key processes and file systems. When a file server or dependent resources is detected not to respond, LifeKeeper will optionally attempt to undertake a local recovery on the server (depending upon configuration). If this recovery fails, then the file server is recovered on the remote server.

In order for failover to take place, all nodes within the cluster need to be able to access the same data files (i.e. the data the file server is sharing with the end users). There are two ways data can be available to all clients, either using a shared storage solution (for example SAN, NAS, SCSI) or LifeKeeper's Data Replication add on can be used to replicate data between servers over a LAN or WAN.

Shared storage allows for a larger cluster to be created, and can also scale further.

Data Replication on the other hand, allows the two nodes to be any distance apart and normally reduces the cost of the solution. It can therefore be used to produce an identical copy of the data on a remote site or different server, effectively enhancing availability and reliability by giving you redundant storage in seperate locations

As illustrated in the image below, LifeKeeper can run in an Active/Active configuration if necessary. Active/Active configurations allow all nodes in a cluster to be active, allowing maximum hardware utilisation while still providing redundancy.

Further down Active/Standby configurations are discussed with the Data Replication option

In the above illustration two nodes form the cluster. Each node runs an IIS webserver as well as sharing files to a network. Both nodes have access to the same shared storage unit. In the event of one node failing, the other node is capable of running all four applications, therefore ensuring continuity of service.

Data Replication - eliminating the need for Shared Storage

LifeKeeper data replication replicates data from a local disk located on an active server to a local disk located on a backup server. When recovery on the backup server takes place, data is up-to-date due to the replication that has already taken place.

Recovery of Windows File Servers to a remote site

Using the data replication feature, it is possible to undertake remote failover to a remote site. This normally involves the use of Asynchronous data replication for performance purposes. LifeKeeper is used to initiate and coordinate the failover to the remote site.

The data replication software provides numerous additional features, for instance is is possible to pause the mirror and undertake a backup on the remote copy of the data, and then unpause the mirror without suffering a full resynchronisation of the data.

Java GUI administration

A Java GUI simplifies administration, configuration and management of the failover and data replication process. This can be done via the console or through a web browser.

---

Original Source: www.openminds.co.uk