Jump to: navigation, search

Difference between revisions of "SMILA/Documentation/HowTo/How to setup SMILA in a cluster"

m (Objectstore configuration)
(Objectstore configuration)
 
(17 intermediate revisions by one other user not shown)
Line 1: Line 1:
 
=== Introduction  ===
 
=== Introduction  ===
  
SMILA is primarily thought of as a framework where you can plug in your own high-performant/high-scalable components (e.g. for storage). Nevertheless, it is also possible to set up SMILA out-of-the-box on a cluster by using its default implementations. This permits ''horizontal scaling'' having the effect that importing and processing jobs/tasks will be shared across the cluster nodes. (Remark: We also have a ''vertical scaling'' on each cluster machine, but this is not new, because you also have this with a single-node SMILA.)  
+
SMILA is primarily thought of as a framework where you can plug in your own or third-party high-performant/high-scalable components (e.g. for data storage). Nevertheless, it is also possible to set up SMILA out-of-the-box on a cluster by using its default implementations. This permits ''horizontal scaling'' having the effect that importing and processing jobs/tasks will be shared across the cluster nodes. (Remark: We also have a ''vertical scaling'' on each cluster machine, but this is not new, because you also have this with a single-node SMILA.)  
  
 
The following steps describe how to set up SMILA on multiple cluster nodes.
 
The following steps describe how to set up SMILA on multiple cluster nodes.
Line 9: Line 9:
 
If you want to use Solr for indexing, you need to set up a separate Solr server, because the Solr instances embedded in SMILA cannot be shared with the other SMILA instances.  
 
If you want to use Solr for indexing, you need to set up a separate Solr server, because the Solr instances embedded in SMILA cannot be shared with the other SMILA instances.  
  
==== Single-Node Server ====
+
==== Single node server ====
  
*Download a Solr 3.x archive from [http://lucene.apache.org/solr/ http://lucene.apache.org/solr/]. This HowTo was tested with Solr v. 3.6.1.  
+
*Download a Solr 3.x archive from http://lucene.apache.org/solr/. This HowTo was tested with Solr v. 3.6.1.  
 
*Unpack the archive to a local directory; you will get a directory like <tt>/home/smila/solr/apache-solr-3.6.1</tt>.  
 
*Unpack the archive to a local directory; you will get a directory like <tt>/home/smila/solr/apache-solr-3.6.1</tt>.  
 
*Copy the files from <tt>SMILA/configuration/org.eclipse.smila.solr</tt> to the Solr machine (<tt>solr.properties</tt> isn't needed here), e.g. to <tt>/home/smila/solr/smila-cores</tt>.
 
*Copy the files from <tt>SMILA/configuration/org.eclipse.smila.solr</tt> to the Solr machine (<tt>solr.properties</tt> isn't needed here), e.g. to <tt>/home/smila/solr/smila-cores</tt>.
Line 17: Line 17:
 
<pre>java -Dsolr.solr.home=/home/smila/solr/smila-cores -jar start.jar
 
<pre>java -Dsolr.solr.home=/home/smila/solr/smila-cores -jar start.jar
 
</pre>  
 
</pre>  
*Check if Solr is running at [http://localhost:8983/solr/DefaultCore/admin/ http://localhost:8983/solr/DefaultCore/admin/] (replace <tt>localhost</tt> with the name of your Solr server, if necessary).
+
*Check if Solr is running at http://localhost:8983/solr/DefaultCore/admin/ (replace <tt>localhost</tt> with the name of your Solr server, if necessary).
  
==== Distributed Server ====
+
==== Distributed server ====
  
 
For larger data volumes you will need to set up Solr in a distributed way, too. However, using a distributed Solr setup is not yet fully supported by the SMILA integration (especially during indexing).
 
For larger data volumes you will need to set up Solr in a distributed way, too. However, using a distributed Solr setup is not yet fully supported by the SMILA integration (especially during indexing).
  
=== Configuring SMILA on cluster node ===
+
=== Configuring SMILA on cluster nodes ===
  
 
On each cluster node, you have to do the following SMILA configuration changes.
 
On each cluster node, you have to do the following SMILA configuration changes.
Line 31: Line 31:
 
You have to define which nodes belong to the cluster.
 
You have to define which nodes belong to the cluster.
  
Configuration file: <pre>configuration/org.eclipse.smila.clusterconfig.simple/clusterconfig.json</pre>
+
Configuration file: <pre>SMILA/configuration/org.eclipse.smila.clusterconfig.simple/clusterconfig.json</pre>
  
 
Enter new section "clusterNodes" stating the host names of the individual cluster nodes:
 
Enter new section "clusterNodes" stating the host names of the individual cluster nodes:
Line 47: Line 47:
 
You have to define a '''shared''' data directory for all nodes ("shared" means that the selected directory must be accessbile from every machine in your cluster under the same path).
 
You have to define a '''shared''' data directory for all nodes ("shared" means that the selected directory must be accessbile from every machine in your cluster under the same path).
  
Configuration file: <pre>configuration/org.eclipse.smila.objectstore.filesystem/objectstoreservice.properties</pre>
+
Configuration file: <pre>SMILA/configuration/org.eclipse.smila.objectstore.filesystem/objectstoreservice.properties</pre>
 +
 
 +
(Directory/File will not exist in older SMILA versions - just create it).
  
 
Set a root path to the shared directory:
 
Set a root path to the shared directory:
Line 63: Line 65:
 
==== Solr configuration ====
 
==== Solr configuration ====
  
You have to point to the Solr server that you set up above.
+
You have to point to the Solr server that we installed above.
  
Configuration file: <pre>configuration/org.eclipse.smila.solr/solr.properties</pre>
+
Configuration file: <pre>SMILA/configuration/org.eclipse.smila.solr/solr.properties</pre>
  
 
<code>
 
<code>
 
   solr.embedded=false
 
   solr.embedded=false
 
   ...
 
   ...
   solr.serverUrl=http://<SOLR-HOST>:8983/solr
+
   solr.serverUrl=<nowiki>http://<SOLR-HOST>:8983/solr</nowiki>
 
</code>
 
</code>
  
 
==== Jetty configuration ====
 
==== Jetty configuration ====
  
To monitor the cluster node, you have to make SMILA HTTP server accessible from external.
+
To monitor the cluster node, you have to make the SMILA HTTP server accessible from external.
  
File: <pre>SMILA.ini</pre>
+
File: <pre>SMILA/SMILA.ini</pre>
  
 
<code>
 
<code>
Line 86: Line 88:
  
 
See also [[SMILA/Documentation/Enable Remote Access|Enabling Remote Access to SMILA]]
 
See also [[SMILA/Documentation/Enable Remote Access|Enabling Remote Access to SMILA]]
 
  
 
=== Monitoring ===
 
=== Monitoring ===
Line 94: Line 95:
 
==== Startup ====
 
==== Startup ====
  
After SMILA has been started, [http://CLUSTER-NODE:8080/smila http://CLUSTER-NODE:8080/smila] should show you the configured cluster nodes (SMILA 1.2):
+
After having started SMILA, accessing <tt><nowiki>http://<CLUSTER-NODE>:8080/smila</nowiki></tt> should return the configured cluster nodes in the response (SMILA 1.2):
 
<code>  
 
<code>  
 
     ...
 
     ...
Line 107: Line 108:
 
</code>
 
</code>
  
==== Job run ====
+
==== Running jobs ====
  
After a job run has been started you can check the number of tasks that are currently processed on each node in the zookeeper state: [http://CLUSTER-NODE:8080/zookeeper/smila/taskmanager/hosts/CLUSTER-NODE http://CLUSTER-NODE:8080/zookeeper/smila/taskmanager/hosts]
+
After having started a job run, you can check the number of tasks that are currently being processed on each node in ZooKeeper's state at <tt><nowiki>http://<CLUSTER-NODE>:8080/zookeeper/smila/taskmanager/hosts/</nowiki></tt>.
  
There you should see a list of the cluster nodes, and the following output for each of them: (The given sample output means that 6 tasks are currently processed on the given cluster node)
+
There, you should see a list of cluster nodes and the following output for each of them: (The given sample output means that 6 tasks are currently being processed on the given cluster node.)
 
<code>
 
<code>
 
   stat: ...
 
   stat: ...
Line 117: Line 118:
 
</code>
 
</code>
  
You can also count the <code>inprogress</code> tasks under [http://CLUSTER-NODE:8080/smila/tasks http://CLUSTER-NODE:8080/smila/tasks], which is the number of tasks currently processed in the whole cluster. This number can be compared with the <code>maxScaleUp</code> setting for a worker in the <code>clusterconfig.json</code> which is the max. number of tasks allowed to be processed on one node. (see also [[SMILA/Documentation/TaskManager#External_REST_API | Taskmanager REST API]])
+
You can also count the <code>inprogress</code> tasks under <tt><nowiki>http://<CLUSTER-NODE>:8080/smila/tasks</nowiki></tt>, which is the number of tasks currently processed in the whole cluster. This number can be compared with the <code>maxScaleUp</code> setting for a worker in the <code>clusterconfig.json</code> which is the max. number of tasks allowed to be processed on one node. (see also [[SMILA/Documentation/TaskManager#External_REST_API | Taskmanager REST API]]).
 
+
  
 
=== Some useful commands ===
 
=== Some useful commands ===
  
Removing all documents from a Solr core (unix-shell command):
+
Removing all documents from a Solr core (Unix-shell command):
  
 
<source lang="text">
 
<source lang="text">
Line 128: Line 128:
 
</source>
 
</source>
  
Optimize Solr index (unix-shell command):
+
Optimizing the Solr index (Unix-shell command):
  
 
<source lang="text">
 
<source lang="text">
Line 134: Line 134:
 
</source>
 
</source>
  
''Hint: When using Windows, replace all ' by "''
+
''Hint: When using Windows, replace all ' marks with "''.

Latest revision as of 10:26, 19 October 2012

Introduction

SMILA is primarily thought of as a framework where you can plug in your own or third-party high-performant/high-scalable components (e.g. for data storage). Nevertheless, it is also possible to set up SMILA out-of-the-box on a cluster by using its default implementations. This permits horizontal scaling having the effect that importing and processing jobs/tasks will be shared across the cluster nodes. (Remark: We also have a vertical scaling on each cluster machine, but this is not new, because you also have this with a single-node SMILA.)

The following steps describe how to set up SMILA on multiple cluster nodes.

Install external Solr server

If you want to use Solr for indexing, you need to set up a separate Solr server, because the Solr instances embedded in SMILA cannot be shared with the other SMILA instances.

Single node server

  • Download a Solr 3.x archive from http://lucene.apache.org/solr/. This HowTo was tested with Solr v. 3.6.1.
  • Unpack the archive to a local directory; you will get a directory like /home/smila/solr/apache-solr-3.6.1.
  • Copy the files from SMILA/configuration/org.eclipse.smila.solr to the Solr machine (solr.properties isn't needed here), e.g. to /home/smila/solr/smila-cores.
  • Go to /home/smila/solr/apache-solr-3.6.1/example and run:
java -Dsolr.solr.home=/home/smila/solr/smila-cores -jar start.jar

Distributed server

For larger data volumes you will need to set up Solr in a distributed way, too. However, using a distributed Solr setup is not yet fully supported by the SMILA integration (especially during indexing).

Configuring SMILA on cluster nodes

On each cluster node, you have to do the following SMILA configuration changes.

Cluster configuration

You have to define which nodes belong to the cluster.

Configuration file:
SMILA/configuration/org.eclipse.smila.clusterconfig.simple/clusterconfig.json

Enter new section "clusterNodes" stating the host names of the individual cluster nodes:

{
  "clusterNodes": ["PC-1", "PC-2", "PC-3"],    
  "taskmanager":{
  ...
}

Objectstore configuration

You have to define a shared data directory for all nodes ("shared" means that the selected directory must be accessbile from every machine in your cluster under the same path).

Configuration file:
SMILA/configuration/org.eclipse.smila.objectstore.filesystem/objectstoreservice.properties

(Directory/File will not exist in older SMILA versions - just create it).

Set a root path to the shared directory:

 root.path=/data/smila/shared
 ...

Idea.png
NFS or SMB/CIFS?

When running on Linux, you can use either an NFS or an SMB/CIFS directory (mounted via Samba) for the objectstore. First tests seem to indicate that using a SMB/CIFS directory is much faster, especially if lots of small files are written (as is the case during crawling processes by the Delta or Visited Links service). Also, we had stability issues with an NFS mount, where a lot of "state NFS file handle" errors occurred.

Of course, the results may largely depend on your environment and could be completely different in your network.


Solr configuration

You have to point to the Solr server that we installed above.

Configuration file:
SMILA/configuration/org.eclipse.smila.solr/solr.properties

 solr.embedded=false
 ...
 solr.serverUrl=http://<SOLR-HOST>:8983/solr

Jetty configuration

To monitor the cluster node, you have to make the SMILA HTTP server accessible from external.

File:
SMILA/SMILA.ini

 ...
 -Djetty.host=0.0.0.0
 ...

See also Enabling Remote Access to SMILA

Monitoring

You can use the REST API to monitor SMILA cluster activities.

Startup

After having started SMILA, accessing http://<CLUSTER-NODE>:8080/smila should return the configured cluster nodes in the response (SMILA 1.2):

   ...
   cluster: {
     nodes: [
       "PC-1",
       "PC-2",
       "PC-3"
     ]
   }
   ...

Running jobs

After having started a job run, you can check the number of tasks that are currently being processed on each node in ZooKeeper's state at http://<CLUSTER-NODE>:8080/zookeeper/smila/taskmanager/hosts/.

There, you should see a list of cluster nodes and the following output for each of them: (The given sample output means that 6 tasks are currently being processed on the given cluster node.)

  stat: ...
  data: "6"

You can also count the inprogress tasks under http://<CLUSTER-NODE>:8080/smila/tasks, which is the number of tasks currently processed in the whole cluster. This number can be compared with the maxScaleUp setting for a worker in the clusterconfig.json which is the max. number of tasks allowed to be processed on one node. (see also Taskmanager REST API).

Some useful commands

Removing all documents from a Solr core (Unix-shell command):

curl http://localhost:8983/solr/DefaultCore/update?commit=true -H "Content-Type: text/xml" --data-binary '<delete><query>*:*</query></delete>'

Optimizing the Solr index (Unix-shell command):

curl http://localhost:8983/solr/DefaultCore/update?commit=true -H "Content-Type: text/xml" --data-binary '<optimize/>'

Hint: When using Windows, replace all ' marks with ".