Table of Contents
Introduction
This note details some of the operational aspects of the Access server. Ultimately this can not cover every aspect of the server and it's not within the capabilities of AGTEK R&D to document things very deeply. The best reference of the server operational behavior is the code base.
Code
The source code for the Access server is in the R&D Bitbucket (GIT) server. Project “Server”, Repository “Agteknet”.
As of this writing the code is designed and structured to be developed using Intellij (community) IDE for Java. The needed packages are checked in to the repository. A separate page detailing the layout of the repository speaks to the layout. AGTEK Access Server Development
Updating the server
After a server build has completed, and it's ready to roll out, the build user performs a “snap” to capture the built server. The key file is “storage.tar”. This tar contains everything to update the server, except for the properties files. Typically this file is copied to a server machine's /tmp directory. The update is applied thusly:
ssh -i accessops.pem ec2-user@IP-OF-SERVER
# After completing log in
# Escape from the captive menu.
cd /mnt/access
./server.sh update /tmp/storage.tar
The server.sh script will copy the server property files to /tmp, stop the server, untar the archive, restore the property files, and restart the server.
Please note that this only updates a single server. The current (2018-October) configuration has two servers running behind a load balancer. Interactions with load balancer and server are details in Access AWS Operations
Administration Operations
Most admin functions (wrt to data in the databas) can be done using the AccessSupport tool. NOTE: This tool has some server start/stop functions that probably don't work correctly. They were somewhat invalidated when the server was rearchitected to run multiple instances.
Starting and stopping servers is still best done using direct control with the server.sh script.
Monitor Behavior
A tool, AccessMonitor exists to show running status of both servers. This tool is intended for DevOps (not AGTEK Support). Using this tool you are able to “keep an eye” on both servers and take action if something becomes amiss. You'd be looking for a server disappearing from the graph “stats” or exceedingly long api method execution times. Both conditions are indicative of something extremely wrong and corrective action must take place.
