Pages

Sunday, 13 April 2014

Application Server Crash or Hung




Symptom:

An application is not responding to incoming requests. The application server process is no longer running.

When we encounter application server crash or hang.

In both cases, the application server is unresponsive. The difference is that for a crash the application server process dies unexpectedly and for a hang the process is still running.

You should check to see if the application server process is running to determine if you are experiencing a crash or a hang. To do this, you need to know the process ID of the application server. You can then use the appropriate operating system command to check to see if the process is running. If it is not running, the problem is a crash. If it is running, the problem is a hang.

 Note: For problem determination strategies for application server crashes, see WebSphere Application Server V6: Application Server Crash Problem Determination at:


 

If a problem does occur that might negatively impact your business, you need to be able to respond quickly and effectively.

To quickly identify the root cause of problems that does occur to minimize the business impact.

Most problems are caused by configuration issues, environment issues, application code defects, or a misunderstanding of WebSphere Application Server. Many of these problems can be resolved easily without calling IBM Support to open a problem management record (PMR).

àTypes of Problem Symptoms:

Here are the common types of symptoms that you might see. Almost every symptom falls into one of these categories:

 

a. We cannot install or migrate WebSphere Application Server or install an application into WebSphere Application Server.

b. We experience difficulties in WebSphere Application Server system management or configuration.

c. An application or WebSphere Application Server process (for example, an application server, node agent, or deployment manager) is unable to start.

d. An application does not respond to incoming requests.

 

This could occur due to

·        Web server, Edge component, or Web server plug-in problem,

·        Application Server Crash, Hang,

·        Out of memory condition or 100% CPU utilization condition.

 

Determine the Root cause of the Problem:-

e. An application produces unexpected results (possibly errors or exceptions).

f. An application cannot connect to an external system or resource.

g. An application performs slowly or its performance degrades over time.

 

à Applying WebSphere maintenance:

Solution Terminology Changes

V5 terms                   V6 terms

Fix pack                    à Refresh pack

Cumulative Fix                    à Fix Pack

Fix packs for Version 6 are delivered on a regular basis, about once every three months. They only contain fixes for APARs (authorized program analysis report), and they do not introduce any new features or functionality to the product. You can think of a fix pack as preventive maintenance. Fix packs do not contain upgrades to the Java Software Development Kit (SDK).

Refresh packs for Version 6 are considerably larger upgrades than fix packs. In addition to containing fixes for APARs, all refresh packs contain some new features and functionality. They also upgrade the Java SDK.  However, you should never install a refresh pack immediately on a production system. All applications should be fully regression tested with the new refresh pack. Upgrading to a new refresh pack level should be planned when there is sufficient time in the development cycle to test it.

 

à Checking the Prerequisites:

Another strategy for preventing problems is to ensure that all software and hardware in your environment meets the prerequisites of WebSphere Application Server V6.

I.e. processor, Minimum disk Space for installation, Physical memory (free Command in UNIX) 1 GB recommended.

 

à Monitoring:

You can monitor for configuration problems and runtime events within the administrative console. To do this:

1. Log onto the administrative console.
2. Expand Troubleshooting.
3. Expand Configuration Problems or Runtime Messages.


You can view errors, warnings, and informational messages directly in the administrative console.

You can select each message that you see to get more details about the message, the source of the message, and the reason why it occurred.

Another way to monitor your application is by using the Tivoli Performance Viewer; you can observe performance metrics, such as

·        Average response time,

·        Number of requests,

·        Thread pool sizes,

·        Connection pool sizes,

·        JVM memory, CPU, I/O,

·        And system paging,  

to monitor the health of WebSphere Application Server and your applications.

 

à Many of the most severe problems, such as application server crashes, hangs, and out of memory conditions, require the largest amount of diagnostic data.

àConfigure the thread monitor for hang detection. WebSphere Application Server V6 includes a thread monitor feature. By default, the thread monitor checks the status of all active threads every three minutes. If it finds a thread that has been active for more than ten minutes, it outputs a warning to the SystemOut log, similar to the following:

WSVR0605W: Thread threadname has been active for hangtime and may be hung.
There are totalthreads threads in total in the server that may be hung.

The thread monitor makes it easier to determine that a problem has occurred. If you see a WSVR0605W warning in your SystemOut log, you out know that a thread has stopped responding. You can then perform further diagnostic steps to determine the cause of the hung thread.

No comments:

Post a Comment