How to avoid network downtime?
Since then, no surprises have occurred with Novell products to solve the high availability of network resources. Naturally, Novell ceased to support its once-very popular NetWare 3.x product, since over the nearly ten-year history of its existence in the world of information technology, much has flowed. The high availability solutions we reviewed, such as SFT III and NHAS, were not developed in future versions of the operating system. If SFT III was not planned for version NetWare 5.x, then NHAS was released for this version, but it quickly found a worthy alternative in the form of Novell® Cluster Services ™ (NCS), which was familiar to us under the working name Orion 2. StandbyServer Products for NetWare and StandbyServer Many-to-One, which reduce server downtime, have found their niche and do a great job with tasks ranging from NetWare 3.x to NetWare 5.x. In NetWare 6.0, for high-availability solutions, it is recommended to use NCS, which is included in the default package with a license for two nodes. This product most fully meets the needs of today.
Novell ® Claster Services ™ is a server clustering system that provides high availability and manageability of important network resources, including data (volumes), applications, server licenses and services. This is a multi-node cluster product for NetWare 5.x, 6, supporting eDirectory ™, failover and failback modes, as well as migration (load balancing) of separately managed resources.
Novell Cluster Services (NCS), included with NetWare 6, is licensed to use two servers in a cluster; additional server licenses (totaling up to 32) can be purchased additionally. For NetWare 5.x NCS is offered as a standalone product. NCS provides failover protection and load balancing for critical network resources, including data (volumes), applications, server licenses, and services such as iPrint, iFolder, and client access. Organizations have the ability to develop a cluster without backup servers; this allows, instead of backing up equipment in case of failures, to use all the cluster servers as active, for example, to optimize the performance of the Web site.
A single NetWare cluster can include servers with different hardware. Network administrators can accurately indicate which of the cluster servers will take over the functions of the failed server, thereby distributing certain applications and data to the most suitable cluster servers.
Storage for clusters (including SAN data) is managed by Novell Storage Services (NSS). Cluster administration is carried out using standard consoles used to manage any NetWare resources – ConsoleOne or Remote Manager. No other tools are needed.
The following example will help you understand the benefits of Novell Cluster Services.
Suppose you have three server clusters configured, and a Web server is installed on each server in the cluster. Each cluster server is the host of two Web sites. All data, graphics, and e-mail messages from each Web site are stored in a shared disk subsystem connected to each server in the cluster. The figure below shows how it looks.
During normal cluster operation, each server maintains a constant connection with other servers in the cluster and periodically polls all registered resources for failure detection. Assume that there is a problem with hardware or software on Web server 1, and users whose Internet access, email, and information depends on Web server 1 lose their connections. The figure below shows how resources are moved when Web server 1 crashes.
Web site A is being moved to Web server 2, and Web site B is being moved to Web server 3. IP addresses and corresponding licenses will also be transferred to Web server 2 and Web server 3. The pattern for moving the Web sites is because we defined it beforehand when setting up the cluster to get a uniform load on the servers.
If Web Server 1 crashes, Novell Cluster Services software performs the following actions:
Re-mount the volumes of shared data (which were previously mounted on Web server 1) on Web server 2 and Web server 3, as defined.
Re-launching applications (which functioned on Web server 1) on Web server 2 and Web server 3, as defined.
Moving IP addresses to Web server 2 and Web server 3, as defined.