We are upgrading certain applications in our environment to Coldfusion 10 and Server 2012 R2.
Everything in the QUA environment worked great, no issues. However, when I put the new servers in production, I started seeing random 503 issues from the servers. (the application pools were stopped) - Restarting the application pool revived the site without a need to restart coldfusion. In the interim of finding a fix, I had a powershell script checking the status of my app pool hourly and if it is down, I have it restart it and email me.
When I started looking into this, I did some tuning of my IIS connectors based on this article: Tuning ColdFusion 10 IIS Connector configuration — Adobe ColdFusion Blog - These particular servers are configured with one instance of coldfusion but separate connectors for each of the sites. Since implementing those changes last week, the 503 errors have only occurred one other time over the past weekend. Looking in the isapi_redirect.log file, I see this error that corresponds to each time the application pools crashed:
[error] init_jk::jk_isapi_plugin.c (2795): Initializing shm:(null) errno=-1. Load balancing workers will not function properly.
Just wondering what the above is actually referencing and if my issue is truly related to load?