Community Tip - Stay updated on what is happening on the PTC Community by subscribing to PTC Community Announcements. X
Hi,
I have configured the master server with 2 slaves with all revelant properties as per the windchill advance guide. I have the dedicated load balancer but do not know the prerequesties for configuring the load balancer. May I know the load balancer type required and any specific settings to be followed.
Web server works if I try with load balancer URL "It Works" but we type the full name http://xyz.domainname.com/Windchill, the page loads just with PTC and if I click on that, it turns to message "Temporary server is unavailable."
Can we access / login to windchill via slave server / master server URL in cluster environment without load balancer for testing, just to confirm that windchill works fine and no issues. So we can concentrate on load balancer part.
You can directly connect to a slave node in the cluster to verify that the cluster is working correctly. But, you can't just point the browser at the slave's hostname. A Windchill cluster will re-direct client requests back to the cluster alias. What you need to do is temporarily map the cluster alias to the slave node's IP address. The easiest way to do this is to just edit the hosts file on your client machine.
After mapping the cluster alias to one of your slave nodes, open up a browser and connect to Windchill using the cluster alias. This will bypass the load balancer and connect you directly to the slave node. Make sure you can log in and use the Windchill UI. If so, then re-map the cluster alias to the other slave node and test that one as well. If you can use Windchill normally from both slave nodes, then the issue must be with the load balancer configuration.
May I know which entry to be changed. Please find the attachement of the slave method server which shows servlet engine not responding.
Below are the entries in Site.xconf.
Master:
java.rmi.server.hostname=physical server name
wt.server.hostname=$(wt.rmi.server.hostname)
wt.rmi.server.hostname=$(java.rmi.server.hostname)
wt.httpgw.hostname=load balancer name
wt.cache.master.slaveHosts=slave names
wt.server.codebase=http://load balancer name/Windchill
wt.cache.master.codebase=http://master server name/Windchill
wt.cache.master.hostname=master server name
ie.ldap.serviceName=master server name
Slave:
java.rmi.server.hostname=physical server name
wt.server.hostname=null
wt.rmi.server.hostname=load balancer name
wt.httpgw.hostname=load balancer name
wt.cache.master.slaveHosts=slave names
wt.server.codebase=http://load balancer name/Windchill
wt.cache.master.codebase=http://master server name/Windchill
wt.cache.master.hostname=master server name
ie.ldap.serviceName=master server name
The properties look correct to me. Were there any errors in the Slave's MS log during startup? Have you mapped the cluster alias to 127.0.0.1 on the slave node?
No errors while starting up in slave server log but after method server ready, I can see the message like servlet engine not responding in the MS console.
Yes, I tried with and without mapping to 127.0.0.1. I want to confirm on windchill functionality before looking into load balancer part.
May I know, what the mapping to loopback address will do....