Shulou(Shulou.com)05/31 Report--

In this issue, the editor will bring you an example analysis of the detection of cpu hunger in the was log newspaper HMGR0152W. The article is rich in content and analyzes and narrates it from a professional point of view. I hope you can get something after reading this article.

The WebSphere Application Server console shows that some applications are unavailable. After checking the log, it is found that HMGR0152W: CPU hunger has been detected. The current thread scheduling delay is 102 seconds. The HMGR0152W message indicates that there is a JVM thread scheduling delay in the current process. The performance is that during the peak period of system operation, the system appears that the WEB page opens slowly, the login application system is slow, and the application runs slowly, sometimes even can not be accessed, which will seriously cause WAS downtime, and there are many reasons that this scheduling thread may not be able to obtain CPU resources in a timely manner.

Principle: the WebSphere Application Server high availability manager component contains thread scheduling delay detection logic, which periodically starts a thread to track whether the thread has been dispatched and whether it is running as planned. By default, the delay detection thread runs every 30 seconds, and if it does not run for 5 seconds beyond the schedule, a HMGR0152W message is logged in the log. This message shows the delay or the difference between the expected CPU and the actual CPU of the thread.

Here are some common causes of this problem:

1. Physical memory is overused and paging is generated.

two。 The heap size of the process is too small, resulting in frequent garbage collection or too long blocking the execution of other threads.

3. There are too many threads running in the system, and too much load may lead to overuse of CPU.

This warning message may lead to WAS instability if not corrected in time. You need to analyze why there is a delay in the scheduling thread and what measures need to be taken. Here are some common solutions:

1. Increase physical memory to prevent paging.

two。 Optimize JVM heap memory configuration to optimize garbage collection. (detailed garbage collection can be set in the WAS console, upper and lower limits of JVM)

3. Reduce the overall load of the system to an acceptable value.

If the HMGR0152W message is not reported frequently, and the message delay is very short (for example, less than 20 seconds), no other problems usually occur, and the message can be ignored at this time.

Of course, in this case, you can also do a javacore to see which threads are waiting (1, what resources are waiting, 2, how many requests are waiting, 3, how many threads are executing); in addition, check the configuration of the operating system, which is also related to operating system parameters. (this can be further confirmed by the results in javacore); check the data for resource conflicts or table locking.

For the analysis of JAVACORE, please refer to my "was process occupies a high CPU rate for an example of deep resolution" http://xjsunjie.blog.51cto.com/999372/1136156

Highly available management thread scheduling delay detection can be achieved by configuring one of the following two custom properties:

1. The IBM_CS_THREAD_SCHED_DETECT_PERIOD parameter determines how often the delay detection thread runs. The default value is 30 seconds.

2. The IBM_CS_THREAD_SCHED_DETECT_ERROR parameter determines how long the delay needs to be logged with this warning.

These properties are in the WAS console, and you can configure them as follows:

1. In the administrative console, click Server > Core Group > Core Group Settings, and then select the core group name.

two。 Under other properties, click Custom Properties > New.

3. Enter the property name and value.

4. Save the configuration.

5. If it is a network deployment environment, please pay attention to synchronize all nodes.

6. Restart all servers so that the configuration takes effect on that server.

Although we can use the above two properties to increase the thread scheduling detection time, it is not recommended to use them to prevent HMGR0152W warning messages from being reported completely. This method is only used to optimize the system to eliminate thread scheduling delays.

The above is the example of the was log report HMGR0152W detected cpu hunger that Xiaobian shared for you. If you happen to have similar doubts, please refer to the above analysis to understand. If you want to know more about it, you are welcome to follow the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.