What steps will reproduce the problem?
1.3 host,1 tomcat instance and 1 memcached instance on each host.
2.use sticky session and set failoverNodes like this.
------10.100.10.21:host1:tomcat1-------
<Context>
<Manager className="de.javakaffee.web.msm.MemcachedBackupSessionManager"
memcachedNodes="n1:10.100.10.21:11211,n2:10.100.10.22:11211,n3:10.100.10.23:11211"
failoverNodes="n1"//In the node of IP 22 is n2 and In the node of IP 23 is n3
requestUriIgnorePattern=".*\.(ico|png|gif|jpg|css|js)$"
transcoderFactoryClass="de.javakaffee.web.msm.serializer.kryo.KryoTranscoderFactory"
/>
</Context>
3.nginx use ip_hash strategy for these 3 tomcat instances.
4.open a page and send a request to nginx, assuming that the request has been
processed by tomcat1.Then the session will be backup in memcached2 or
memcached3.Now,assuming that the session is backup in memcached2.Next, to kill
tomcat1,then flush the page.
At this time if the request fails over to tomcat2, tomcat2 will obtain session
from memcached2 and the result will be successful. At this time, session will
only exist in tomcat2 and memcached2.
What is the expected output? What do you see instead?
My expectation is that, if the session has been obtained from memcached2
successfully,the session should be transferred to memcached1 or
memcached3,rather than still exists only in memcached2 and tomcat2.
What version of the product are you using? On what operating system?
tomcat 6,msm 1.6.5
I use ubuntu 10.04 os on a VMware,and install nginx and 3 memcached instances
with port 11211,11212,11213.
I start 3 tomcat instances on my local machine with win7 os.
Original issue reported on code.google.com by 150349...@qq.com on 3 Jan 2014 at 5:29
Original issue reported on code.google.com by
150349...@qq.com
on 3 Jan 2014 at 5:29