5 Tips for WebLogic Server Administrators
Part of the BEA WebLogic Server 8 For Dummies Cheat Sheet
A WebLogic Server administrator's job has many facets. And as you administer systems, you gain experience of what works and what doesn't work. Here are five useful tips for WebLogic Server administration.
As a WebLogic Server administrator, you'll follow many procedures, including tasks such as these:
Restarting the server
Shutting down the server for routine maintenance
Deploying new versions of WebLogic Server
Backing up the server
Installing the latest patches
Creating WebLogic Server resources such as data sources
You should have written instructions for each of these procedures, which will enable you to follow the same procedure each time, ensuring consistency.
Written procedures also enable your company to perform these operations when you're away. In addition, if you take a new position in the company or with a new firm, having written procedures enables you to fulfill your responsibility to transfer knowledge to the new administrator.
Define a service level agreement
A service level agreement (SLA) helps to define what end users expect from your server in terms of reliability. Most users expect that a system will be up and running 24 hours a day, 7 days a week. Such a schedule is simply not possible. Many events will cause your system to be down for a period of time. For example, dealing with hardware failures, routine updates, or rebooting your server to name a few.
The SLA is the contract between you and the users that your system supports. This contract should specify the amount of time that your system will be allowed to be down through the year.
In addition to defining maintenance periods, a properly written service level agreement should also specify the following:
When maintenance will be performed
How many minutes of unexpected outage are allowed per year
How soon the system must return after an unexpected outage
How often backups will be performed
The overall percent of time that the server should be up
Set up on-call procedures
At some point, the system will go down unexpectedly. When an unexpected outage occurs, you and your staff must be ready to deal with it. The outage may be something that the administrator can handle or something related to the software. If the outage is caused by a software error, a developer will need to get involved in the solution. Additionally, these outages may occur outside regular business hours. This is especially true if you work for a multinational corporation.
Plan for growth
When your system is first deployed, you may not be thinking about growth. But you should have a plan when your current system is outgrown. In general, you have two choices when your system can no longer handle the amount of processing required:
Upgrade your server to a faster machine. Perhaps one of the simplest ways to handle more requests is to upgrade to a faster machine. This may mean the purchase of a new server or simply adding another processor to your current server. When you upgrade to a faster machine, you must make sure that your server is properly copied across the network to the new machine. All configuration settings and installed packages should be copied to the new machine.
Add additional servers to your cluster. If you're running a cluster of servers, you can simply add another server. If you're not running a cluster of servers and your request volume is becoming too high, you should consider using a cluster of servers. Adding another server to the cluster causes WebLogic Server to have another server that can share some of the workload. This allows the application as a whole to be able to accept more connections.
Back up your servers
Backing up data is an important part of any administrator's job. For backing up WebLogic, you'll need to back up the part of your web application that changes — the SQL database. If this data is already being backed up by a database administrator, you don't need to worry about backing up application data.
If you lose the hard drive on your WebLogic server, you'll be expected to reinstall everything and get the server running again. If your application was packaged as a web application archive (WAR) file, you can quickly get your application back up by redeploying the WAR file.