We use non sticky LB topology and it works well...however.... Shiro sessions are never passed from one instance of Shiro to another unless you have a clustered session that spans both instances and both Shiro instances are configured to use/read that common store.
Each Shiro instance has no knowledge of the number of instances in your cluster. As a guess I would say your LB is favouring one instance and every-so-often picking the other instance, which is your rare case when you lose your session. Can you please provide your shiro configuration? On 9 April 2014 12:52, hipjiveguy <[email protected]> wrote: > Hi there; > > We've been running Shiro fine under load balance using Apache, but have > noticed that in rare cases, it seems that the session isn't being passed > from one node to another. We're not using sticky sessions and the nodes in > the balancer are being selected in a round robin fashion. > > Are there any docs detailing usage in this situation? I assume the default > Shiro session manager wouldn't work under load balance? I'm new to Shiro > (helping maintain an existing project) and this just recently came up. > > Thanks, > > Brad > > PS: I originally posted this question the developer forum by mistake, and > posted it here once I realized where it should really go! Thanks! > > > > -- > View this message in context: > http://shiro-user.582556.n2.nabble.com/Any-special-consideration-needed-for-use-in-Load-Balancer-tp7579867.html > Sent from the Shiro User mailing list archive at Nabble.com. >
