| What are some of the ways to implement high availability for your WebSphere Application Servers (WASs)? IBM Global Services has implemented high-availability WAS servers in different ways for customers at our commercial server farms. In some cases, IBM uses AIX Clustering with HACMP or Microsoft OS Clustering (MSCS). In other cases, IP sprayers (e.g., Network Dispatchers, Cisco Local Directors, Cisco Distributed Directors, etc.) provide failover and load balancing capability. There is also the possibility of using a single server that has very high availability with redundancy built in (e.g., fault-tolerant systems such as IBM z Series mainframes or Tandem computers). This article gives "lessons learned" and recommendations on the use of different techniques for implementing high-availability WebSphere servers. IBM's WebSphere Web Application server (WAS) product provides a powerful base for building sophisticated Web (Internet or Intranet) applications. For most Web servers, high availability is of paramount importance. Availability is not an all-or-nothing proposition. Availability can be thought of as a continuum. Reliable hardware and software provide the base level of availability. Advanced features such as RAID devices provide an enhanced level of availability. High-availability software provides near continuous access to data and applications (e.g., OS clustering where a hot standby server might take a minute or so to come online in a failover situation). Fault-tolerant systems ensure the constant availability of the entire system, but at a higher cost. WAS V4 "server groups" and WAS V5 "clusters" can be used with high-availability techniques to provide load balancing. So a good starting point for high availability would be to use WebSphere's built-in failover capability using the Work Load Management (WLM) facility. OS clustering can be used along with the failover capability in WebSphere to enhance the overall system failover. OS clustering provides high availability but often not load balancing. The primary goal of high-availability clustering software is to minimize, or ideally , eliminate, the need to take your resources out of service during maintenance and reconfiguration activities. Two popular examples of OS clustering are AIX HACMP and Microsoft OS Clustering (MSCS). OS clustering supports hardware failover for multiple servers, or nodes, in a cluster sharing the same common disk device (thus, OS clustering is also referred to as hardware clustering). OS clustering often provides failover protection so that if two servers running WAS are set up for failover, and one server becomes unavailable, the second server will come up, use the shared disk subsystem, restart the respective applications, and continue with the processes. It will use the same disk where the WAS data resides and the same IP address that the WebSphere server is using. Therefore, the failover function is transparent to clients . However, if the shared disk device OS clustering is using goes down, the system is down. Thus, it is very important to have a mirrored (RAID 1) shared disk for OS clustering. Often OS clustering provides only failover protection ”not load balancing ”but newer versions of products such as HACMP will also provide some load balancing. Often IP sprayers such as Network Dispatcher are used in conjunction with OS clustering to provide very high availability and load balancing. High-availability systems are an excellent solution for applications that can withstand a short interruption should a failure occur but which must be restored quickly. Some industries have applications so time-critical that they cannot withstand even a few seconds of down-time. Many other industries, however, can withstand small periods of time when their database is unavailable. For those industries, basic OS clustering can provide the necessary continuity of service without total redundancy. |
