In the course of doing the experiment, a phenomenon was found:
Both A and B are working properly.
User C made two requests Q1 and Q2,Q1 request was processed by a, Q2 request was processed by B
A stopped, c initiated the request Q3, was handled by B normally,
A restart Startup
C Initiate request Q4,q5, ... These requests are handled by B
User D initiates a request to be processed by a
After that, the C's request can be processed by a and B normally.
I found some information on the Internet.
650) this.width=650; "src=" Http://s3.51cto.com/wyfs02/M02/59/F3/wKiom1TwNzqzGKZZAAEx-WN3tQ4373.jpg "title=" Qq20150227172104.png "alt=" Wkiom1twnzqzgkzzaaex-wn3tq4373.jpg "/>
Guess that might be a version of the problem
It is possible that the newly added Tomcat has not completed the session copy from the other nodes in the cluster, resulting in a temporary inability to process the request, just a few more ....
Http://localhost:9080/docs/cluster-howto.html
See the most section how it works
This article is from the "ten-li Rice flower" blog, please be sure to keep this source http://5880861.blog.51cto.com/5870861/1615783
Problems restarting Tomcat during Tomcat load balancing