Our team recently deployed a new version of application.cfm to our clustered production environment (with sticky sessions). The changes were load tested on our staging server (not clustered) and found to be error free. When we deployed to Production, the website completely stopped responding giving only an "Internal Server Error - Read" message. When we attempted to hit each one of the servers in the cluster the connection timed out.
Our thoughts are that the issue is somehow related to the load balancer. The staging environment is not a clustered environment so is not a true duplicate of the production environment. The staging server issue is in the hands of the client who understandably isn't thrilled about the expense of setting up a dupliate hardware environment of the Production environment. Here are my thoughts... The application name on the new application.cfm is different from the old version. Would changing the application name mid session (for some users) cause this kind of problem? I noticed in the CF exception logs that there were a number of NullPointer exceptions referencing the old application name at the time we deployed the new application.cfm. Our thoughts at this point are 1) try keeping the same application name 2) making the application.cfm change while CF services are stopped and then restarting services after the file deployment is complete 3) attempt the deployment one server at a time instead of to the whole cluster. Anyone ever experienced this sort of issue before? Thoughts? Suggestions? ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~| ColdFusion MX7 by Adobe® Dyncamically transform webcontent into Adobe PDF with new ColdFusion MX7. Free Trial. http://www.adobe.com/products/coldfusion?sdid=RVJV Archive: http://www.houseoffusion.com/groups/CF-Server/message.cfm/messageid:6261 Subscription: http://www.houseoffusion.com/groups/CF-Server/subscribe.cfm Unsubscribe: http://www.houseoffusion.com/cf_lists/unsubscribe.cfm?user=11502.10531.10
