At present, most of me is to update the system in Monday morning, and then go to the headquarters. However, it seems to have formed a tradition, clearly running good things. Once I upgraded, I started to make a pleasing meeting, it will strike, until the urgency Call, although the system has always been tested by a certain strength - Of course, it is not a gravity test every time. The upgrade is too frequent, it is impossible to test complete each time. The last time was upgraded for two hours, it began to slow down. After pressure testing and debug, it was not too sure some code. It seems to have been tested by strict ten times, spent a week of work. time. This time, it is nothing more than the URL overwritten in the periphery, basically no procedure for Tomcat, but two hours in the middle of the night, Tomcat is turned on. However, in the daytime load, it is necessary to run a day, but the fart is not; this can know what is the reason. Take a closer look in the evening, discover the cause of the machine, should be because Tomcat appears out of Memory anomalies at runtime, and the reason for this exception is estimated because two errors repeatedly appear: one is due to deletion of a normal include The login interface caused by JSP is abnormal (meaningful all day no one to log in), and the other is a colleague who has not been detailed with the old-fashioned JSP / PARAM mode, and the data type error is repeatedly output; There is an error that causes garbage to be cleared, and finally abnormally exits. However, this explanation is also very savvy, because the pressure is much better during the day, don't you do it? In any case, you still have a little bigger in the HVM's processing memory stack. References http://java.sun.com/docs/Hotspot/gc1.4.2/#3.1. total HEAP|Outline, plus two error clears, probably no longer appear again.
Full ID: http://zwwwxy.blogchina.com/blog/Article_81038.1378434.html
