LISTENER监听器日志service_update自动更新频率问题？

liuxbgiant

在以下DB版本下，
Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - Production
PL/SQL Release 11.2.0.1.0 - Production
CORE    11.2.0.1.0      Production
TNS for 32-bit Windows: Version 11.2.0.1.0 - Production
NLSRTL Version 11.2.0.1.0 - Production

listener日志文件内容为：
11-7月 -2012 16:04:26 * service_register * orcl * 0
11-7月 -2012 16:04:35 * service_update * orcl * 0
11-7月 -2012 16:04:41 * ping * 0
11-7月 -2012 16:04:44 * service_update * orcl * 0
11-7月 -2012 16:05:14 * service_update * orcl * 0
11-7月 -2012 16:05:44 * service_update * orcl * 0
11-7月 -2012 16:06:14 * service_update * orcl * 0
11-7月 -2012 16:06:44 * service_update * orcl * 0
11-7月 -2012 16:07:14 * service_update * orcl * 0
11-7月 -2012 16:08:11 * service_update * orcl * 0
11-7月 -2012 16:08:14 * service_update * orcl * 0
11-7月 -2012 16:08:44 * service_update * orcl * 0
11-7月 -2012 16:09:14 * service_update * orcl * 0
11-7月 -2012 16:09:44 * service_update * orcl * 0
11-7月 -2012 16:10:14 * service_update * orcl * 0
11-7月 -2012 16:10:44 * service_update * orcl * 0
11-7月 -2012 16:11:14 * service_update * orcl * 0
11-7月 -2012 16:11:17 * service_update * orcl * 0
11-7月 -2012 16:12:14 * service_update * orcl * 0
11-7月 -2012 16:12:44 * service_update * orcl * 0
11-7月 -2012 16:13:14 * service_update * orcl * 0
11-7月 -2012 16:13:44 * service_update * orcl * 0
11-7月 -2012 16:14:14 * service_update * orcl * 0
在启动监听服务后，service_update的频率是否有一定的规律，如有，则请maclean liu解释下它的相关性。图中显示有时为30秒，有时较大，为什么？

Maclean Liu(刘相兵

ODM FINDING:


Understanding Instance Load Balancing
With server-side connection load balancing, the listener routes incoming client connections to instances on the least loaded nodes and then to the least loaded instance on that node. The listener determines connection distribution based on profile statistics that are dynamically updated by PMON. The higher the node load the more frequent PMON updates the load profile. Updates occur in as little as one minute on heavily loaded nodes and may take up to ten minutes on lightly loaded nodes.

Listener logs and PMON traces both offer valuable information when evaluating instance load balancing. We will review both areas in detail in this document.

Please note that the examples that follow utilize the Unix "grep" command to locate strings in the listener logs, Windows systems have a similar facility, "find", which can be used in the same manner.



pmon定期会给listener 更新 goodness 实例负载信息，(The higher the node load the more frequent PMON updates the load profile)系统负载越高这种更新越频繁

更多信息可以参考
Understanding and Troubleshooting Instance Load Balancing [ID 263599.1]

你也可以通过 10257 trace name context forever, level 16 事件来 TRACE PMON的行为