Deployment of MEC to the production environment is not fully automated. It involves a number of manual steps which must be followed exactly to ensure correct operation. In the past we've had instances when steps were overlooked, resulting in significant support effort and client pain that could have been avoided.
An incomplete deployment may seem fine during a cursory trial but can still have hidden bugs waiting to create problems. We will never have resources to do a thorough functional check after a new deployment.
We hope to fully automate deployment at some point, as Noesis has done. Until that is done the following steps must be followed:
|1.||Use MECMD to upgrade all the v2 DBs on SQLSERVER2, backing up each database first. This is a command line option within the MECMD tool. If you are unsure it's working with the current structure, test it first on a recent copy on your local machine to save time. In some cases having to restore a DB requires you restart the MSSQL service to release the connections. NOTE: The new version of MECMD that you do not yet have will do the update within a transaction. If you run into errors I find the easiest way to fix is to find the *.tmp file in the scripts folder that matches the new hash (see the console output) and run it within a transaction on the restored database. You will be able to quickly correct and test problems with the combined script.|
|2.||Use the Publish option in VS to produce a .zip file of MECWebApplication project (build config Release). Do the same thing to produce .zip files for MECWeb.REST and MECWeb.Authentication.|
|3.||Unzip all three into the application directory on appserver2|
|4.||Update config files for the application (Web.Config, Settings.Config, and Connections.Config) – including reducing cache lifetime (10 seconds). This is usually the hard part.|
|5.||Check IIS setting that the v2 app pool runs in .NET 4 , not in .NET 2.|
©2017 Managing Energy Inc.