Start and stop servers

Starting the servers actually means to start several server processes. The Node Managers should be configured as services on all platforms. On Windows this is configured automatically during the installation process, on Unix it is a manual step done after the installation. It is required to know about the different components in IFS Middleware Server and understand the difference between a Node Manager, an Admin Server and a Managed Server described here.
NOTE: Compared to earlier releases of the middle tier server it is not enough to simply stop the windows services in order to stop the whole cluster! It will only stop the Node Managers but the Admin Server, the Managed Servers and the HTTP Server will still be running.

Overview

IFS Middleware Server can be controlled using scripts located under <ifs_home>/instance/<instance>/bin. All the scripts referred to in this guide resides under this folder.

Contents

• Starting Up the Cluster
• Taking Down the Cluster
• Controlling the Node Managers
• Controlling the Servers
• Check the Server Status
• Start the Batch- and Connect Server manually on Windows
• Start the Batch- and Connect Server manually on Unix types of OS
• Template for boot/shutdown on Unix types of OS
• Unix Script Usage Reference

Starting Up the Cluster

When starting the servers, the order each component is started in is important. They should preferably be started in the following order:

AS Node Manager -> Admin Server -> Managed Server(s)

The order of which the Managed Servers are started does not matter, i.e. ManagedServer3 can be started before ManagedServer1.

The HTTP Server follows the same logic:

HTTP Node Manager -> HTTP Server

However, the HTTP Server does not depend on the managed servers, so it is possible to start up the HTTP Server and then the Managed Servers.

Taking Down the Cluster

The order is also important when bringing down the servers. Without the Node Manager or the Admin Server, the Managed Servers cannot be stopped. It is also important to stop all server instances correctly using the scripts because if a server is killed it will be started again thanks to the crash recovery feature. Shutting down the computer without properly stopping the servers will result in the servers getting started as soon as the computer is booted up again.

The order of which the Managed Servers should be shut down

Managed Server(s) -> Admin Server -> AS Node Manager

The order of which the HTTP server should be stopped.

HTTP Server -> HTTP Node Manager

Controlling the Node Managers

There are two Node Managers that work independently of each other, so stopping one of them will have no effect on the other one. The scripts are a bit different depending on the platform since on windows they are controlled as services.

Command	Windows	Linux
Start Application Server Nodemanager	start_as_nodemanager.cmd	./as_nodemanager.sh start
Stop Application Server Nodemanager	stop_as_nodemanager.cmd	./as_nodemanager.sh stop
Start HTTP Nodemanager	start_http_nodemanager.cmd	./http_nodemanager.sh start
Stop HTTP Nodemanager	stop_http_nodemanager.cmd	./http_nodemanager.sh stop

Controlling the Servers

There are different types of servers. Admin Server, Managed Server and HTTP Server. The HTTP Server is controlled separately from the other servers. It is recommended to make sure that the Admin Server is running before attempting to start or stop a Managed Server, but the HTTP Server can be controlled without the Admin Server. Always make sure that the Node Managers are running.
The start and stop scripts for the application servers takes one argument; the name of the server to start or stop. The argument is case sensitive except for the Admin Server. The name of the Managed Servers was given in the Cluster member configuration step. If no argument is given the script will prompt for the name of the server.
The start and stop scripts will always prompt for the password.

Example:

Starting admin server: start_as_server AdminServer
Stopping admin server: stop_as_server AdminServer
Starting a managed server with name ManagedServer1: start_as_server ManagedServer1
Stopping a managed server with name ManagedServer1: stop_as_server ManagedServer1

It is also possible to start or stop all Managed Servers, the Admin Server and the HTTP Server at once except for the Node Managers. When using the start_all_servers script it is mandatory to start the Node Managers first and when using the stop_all_servers script the Node Managers will not be shut down automatically. The scripts take no arguments.

This script might run into some problems in case some servers are in an incompatible state.

Check the Server Status

It is possible to check the status of the cluster by using the check_server_status script. However, it will only check the status of the application server Node Manager, the admin server and the managed servers.
The script will if possible give the current status of all servers, but if the Node Manager and the Admin Server is down, it is not possible for the script to get the status of the Managed Servers.

Example:
Check status of the cluster: check_server_status
Result:

Nodemanager is RUNNING
AdminServer is RUNNING

There are 3 server(s) in cluster: Cluster1
States of the servers are
ManagedServer3---RUNNING
ManagedServer2---SHUTDOWN
ManagedServer1---RUNNING

Start the Batch- and Connect Server manually on Windows

If you do not want to use Windows services, then you have to start the batch and connect servers manually. The program/scripts for Batch Server and Connect Server are located in the <ifs_home>\instance\<Instance>\bin catalog.

1. Start the Batch Server.
   
   The script is located in <ifs_home>\instance\<Instance>\bin
   Run command start_batchserver1.cmd
2. Start the Connect Server.
   
   The script is located in <ifs_home>\instance\<Instance>\bin
   Run command start_connectserver1.cmd

Start the Batch- and Connect Server manually on Unix types of OS

The start scripts for the Batch Server and Connect Server are located in the <ifs_home>/instance/<Instance>/bin catalog.

1. Start the Batch Server.
   
   The script is located in <ifs_home>/instance/<Instance>/bin
   Run command ./batchserver1.sh start as user ifs
2. Start the Connect Server.
   
   The script is located in <ifs_home>/instance/<Instance>/bin
   Run command ./connectserver1.sh start as user ifs

Template for boot/shutdown on Unix type of OS

A template for boot/shutdown is included as <ifs_home>/bin/ifsctl.sh.
This script should always be run as root and must be edited for your specific installation, e.g. if more Batch/Connect servers are added.

Unix Script Usage Reference

This section is a reference of the capabilities, location, and names of the Unix version of the server scripts.

Boot/shutdown template script

Batch Server Script

Connect Server Script