Skip to content

Latest commit

 

History

History
164 lines (110 loc) · 5.7 KB

readme.adoc

File metadata and controls

164 lines (110 loc) · 5.7 KB
Raw

Server Configuration

In this lab, you will gain hands-on experience with the basic configuration of a GemFire server cluster.

Concepts you will gain experience with:

  • Configuring and starting a locator for member discovery

  • Configuring and starting GemFire cache servers

  • Loading data into the GemFire distributed system

Configuring GemFire properties file

Generally, locators and servers use the same distributed system properties file, which lists locators as the membership coordinator.

Edit cluster/gemfire.properties as follows:

  1. Set the locators property to have port 10334 (see the gemfire properties file reference)

  2. Set the cache-xml-file property to point to the server cache xml file. Note that this will be a relative path to where the servers run. Consequently, you will need to set the path to ../cluster.xml.

Starting the Locator

From a terminal prompt, navigate to the cluster/ folder and start the gemfire shell:

$ cd cluster
$ gfsh

Invoke the following command:

gfsh> help start locator

Take the time to learn about the command start locator and its many parameters. The locator name is a required parameter, all other parameters are optional.

Let’s start a locator:

gfsh> start locator --name=locator

Here’s the console output:

Starting a Geode Locator in .../server-configuration/cluster/locator...
...
Locator in .../server-configuration/cluster/locator on 172.20.1.118[10334] as locator is currently online.
Process ID: 5199
Uptime: 2 seconds
GemFire Version: 9.0.2
Java Version: 1.8.0_121
Log File: .../server-configuration/cluster/locator/locator.log
JVM Arguments: -Dgemfire.enable-cluster-configuration=true -Dgemfire.load-cluster-configuration-from-dir=false -Dgemfire.launcher.registerSignalHandlers=true -Djava.awt.headless=true -Dsun.rmi.dgc.server.gcInterval=9223372036854775806
Class-Path: /usr/local/Cellar/gemfire/9.0.2/libexec/lib/geode-core-9.0.2.jar:/usr/local/Cellar/gemfire/9.0.2/libexec/lib/geode-dependencies.jar

Successfully connected to: JMX Manager [host=172.20.1.118, port=1099]

Cluster configuration service is up and running.

Notice how this command has the side effect of creating a subfolder (named after the locator name) and that the locator process is logging to the locator.log file within that folder.

Starting the Servers

Let’s have two servers join the locator to form a cluster.

Invoke the following command:

gfsh> help start server

Take the time to learn about the command start server and its many parameters. There are two aspects to the start server command that are significant:

  1. Since we’re starting multiple server processes on the same machine, each daemon process must listen on a unique port. A simple way to guarantee that each new server listens on a different port is with the option --server-port=0.

  2. The servers must join the locator, and so will need to be given the locator’s address. There are two ways to do this: using the option --locators=$hostname[$port], or indirectly by referencing the gemfire.properties file you prepared earlier. We’ll use this latter option.

Invoke the start server command:

gfsh> start server --name=server1 --server-port=0 --properties-file=gemfire.properties

Here’s the console output:

Starting a Geode Server in .../server-configuration/cluster/server1...
...
Server in .../server-configuration/cluster/server1 on 172.20.1.118[50719] as server1 is currently online.
Process ID: 5293
Uptime: 2 seconds
GemFire Version: 9.0.2
Java Version: 1.8.0_121
Log File: .../server-configuration/cluster/server1/server1.log
JVM Arguments: -Dgemfire.default.locators=172.20.1.118[10334] -DgemfirePropertyFile=.../server-configuration/cluster/gemfire.properties -Dgemfire.use-cluster-configuration=true -Dgemfire.start-dev-rest-api=false -XX:OnOutOfMemoryError=kill -KILL %p -Dgemfire.launcher.registerSignalHandlers=true -Djava.awt.headless=true -Dsun.rmi.dgc.server.gcInterval=9223372036854775806
Class-Path: /usr/local/Cellar/gemfire/9.0.2/libexec/lib/geode-core-9.0.2.jar:/usr/local/Cellar/gemfire/9.0.2/libexec/lib/geode-dependencies.jar

Now start a second server in the same manner. Give the second server a unique name (e.g. server2).

List the members of your cluster with the list members command:

gfsh> list members
 Name   | Id
------- | -----------------------------------------------
locator | 172.20.1.118(locator:5199:locator)<ec><v0>:1024
server1 | 172.20.1.118(server1:5293)<v1>:1025
server2 | 172.20.1.118(server2:5346)<v2>:1026

Congratulations, you now have a three-member cluster up and running.

Data Operations to Verify Cluster

Let’s verify that this cluster is operational by running another test that will simply insert a few records into a region and validate the count of records inserted.

This can be done directly from the IDE by running the test named VerifyCluster, or from the command line:

cd server-configuration
gradle test

If the test passes, we’re good to go on to the next module.

But before we go on, let’s shut down our cluster.

If you’ve exited gfsh, you must first re-launch gfsh and re-connect to the cluster:

$ gfsh
gfsh> connect

You’ll see output similar to this:

Connecting to Locator at [host=localhost, port=10334] ..
Connecting to Manager at [host=172.20.1.118, port=1099] ..
Successfully connected to: [host=172.20.1.118, port=1099]

Invoke the following command:

gfsh> help shutdown

Proceed to issue the shutdown command, and make sure to also shutdown the locator:

gfsh> shutdown --include-locators=true

Congratulations!! You have completed this lab.