Thanks for the replies. We are going to use "F5" for load balancing the Tomcat instances. As mentioned all the machines will share the same VIP. No Apache HTTP server for now !!
I was able to use the session replication ( PersistenceManager) on tomcat along with Apache HTTP server. I have to check how I can achieve the same thing using F5. Appreciate if you can share quick information for using tomcat session replication along with F5. Durga Prasad Daniel Ruggeri wrote: > > On 4/6/2012 10:34 AM, Pratapani wrote: >> Thanks for the reply. >> I have used the mod_proxy . Changed the httpd.conf to have 2 tomcat >> instances in cluster. >> <Proxy balancer://testcluster stickysession=JSESSIONID> >> BalancerMember ajp://127.0.0.1:8009 min=10 max=100 route=node1 >> loadfactor=1 >> BalancerMember ajp://127.0.0.1:8019 min=20 max=200 route=node2 >> loadfactor=1 >> </Proxy> >> Tomcat clustering works perfect. If Tomcat instance 1 ( node1 ) is >> shutdown >> , the request is serviced by instance 2 . >> >> I would like to have Apache HTTP Server also in cluster and avoid the >> single >> point of failure of Apache HTTP server i.e if one instance of Apache HTTP >> server is down, the application should still work. >> >> I did not find any configuration setting to make 2 Apache HTTP servers >> instances into a cluster. >> >> Appreciate your help. >> >> Durga Prasad > > Durga; > What you are describing is outside the scope of httpd, unfortunately. > There are a few options - the two most popular are: > **Configure an HA VIP using OS-level clustering utilities* - this is the > windows managed integrated load balancer, or Linux-HA w/heartbeat. In > this configuration, you would bind both httpd instance to the same VIP > and allow the clustering utility to manage what node is live. With this > setup, your capacity at the web layer is limited to the capacity of a > single httpd server. > **An external load balancer* - this can be a hardware IP sprayer, > firewall, software service on another machine, etc. Anything operating > on layer 4 or up would do. In this sort of configuration you will bind > httpd to a distinct IP address bound to each machine in the load > balanced pair and let the upstream service manage the traffic. Capacity > in this configuration at the web layer scales linearly with each httpd > server. Beware, this option can be expensive. > > -- > Daniel Ruggeri > > > -- View this message in context: http://old.nabble.com/How-to-set-up-2-or-more-machines-running-HTTPD-instances-to-serve-a-same-request-%28-in-a-load-balanced-fashion%29.-tp33564835p33646045.html Sent from the Apache HTTP Server - Users mailing list archive at Nabble.com. --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@httpd.apache.org For additional commands, e-mail: users-h...@httpd.apache.org
