The installation of elasticsearch (which is distributed RESTful, search engine built on top of Apache Lucene) on Centos, will be easier if it is in repositories. But it isn’t. And since the best way to install application on Centos is to use rpm packages we will need to build it. At the very beginning you need to install Java:
yum install java-1.6.0-openjdk.x86_64
To not write our own SPEC files we will use the one kindly written by Tavis Aitken. It could be downloaded from his github page. After unpacking copy SPEC file to your ~/rpmbuild/SPECS directory. Besides you need to copy config files to ~/rpmbuild/SOURCES.
cp tavisto-elasticsearch-rpms-c78de68/SPECS elasticsearch.spec rpmbuild/SPECS/ cp tavisto-elasticsearch-rpms-c78de68/SOURCES/* rpmbuild/SOURCES/
If you don’t have ~/rpmbuild directory you may need to install the packages rpm-build and rpmdevtools, and run rpmdev-setuptree command.
yum install rpm-build rpmdevtools rpmdev-setuptree
I would recommend to use last version of elasticsearch (at the time of writing it’s 0.19.1) since there are a lot of bug fixes (for instance, this one). To do this you need to modify SPEC file to point to the last version:
Version: 0.19.1
Now we are ready to download the tarball:
spectool -g rpmbuild/SPECS/elasticsearch.spec mv elasticsearch-0.19.1.tar.gz rpmbuild/SOURCES/elasticsearch-0.19.1.tar.gz
And finally build rpm and install it:
rpmbuild -bb rpmbuild/SPECS/elasticsearch.spec rpm -ivh rpmbuild/RPMS/x86_64/elasticsearch-0.19.1-1.el6.x86_64.rpm
At this point we need to configure the application. Comment out all lines in /etc/sysconfig/elasticsearch. After that configure /etc/elasticsearch/elasticsearch.yml to fit your requirements. Here is a simple example:
cluster.name: elasticsearch node.name: "Franz Kafka" path.conf: /etc/elasticsearch path.data: /var/lib/elasticsearch path.logs: /var/log/elasticsearch network.bind_host: _eth0:ipv4_ network.publish_host: _eth0:ipv4_ network.host: _eth0:ipv4_ transport.tcp.port: 9300 http.port: 9200 discovery.zen.ping.multicast.enabled: false
Start elasticsearch and make sure it’s running:
service elasticsearch start netstat -lpn | grep java
Enable auto-start:
chkconfig elasticsearch on
Note that for security reasons it would be a good idea to restrict access to elasticsearch with iptables.
Now you can test if it’s working. You may PUT a data:
curl -XPUT http://localhost:9200/twitter/tweet/2 -d '{
"user": "kimchy",
"post_date": "2009-11-15T14:12:12",
"message": "You know, for Search"
}'
{"ok":true,"_index":"twitter","_type":"tweet","_id":"2","_version":1}
And then try to GET it:
curl -XGET http://localhost:9200/twitter/tweet/2
{"_index":"twitter","_type":"tweet","_id":"2","_version":1,"exists":true, "_source" : {
"user": "kimchy",
"post_date": "2009-11-15T14:12:12",
"message": "You know, for Search"
}}
Seems it’s working. Let’s delete test data:
curl -XDELETE 'http://localhost:9200/twitter/tweet/1'
{"ok":true,"found":false,"_index":"twitter","_type":"tweet","_id":"1","_version":1}