Installation

Hardware

As CLAIMS Direct SOLR is a pre-configured, bundled distribution of Apache SOLR, it can be deployed on any number of nodes (individual instances). A group of nodes function to expose a collection. Further, multiple collections could be searched across the distribution.


There are many scenarios for a CLAIMS Direct deployment that range from indexing the entire content of CLAIMS Direct XML to the sparse indexing of certain fields and ranges of publication dates for application-specific usage. There could also be specific QoS requirements: minimum supported queries per second, average response time et al. All of these factors play a role in planning for a CLAIMS Direct SOLR deployment. Generally speaking, a full index with the entire content of CLAIMS Direct XML requires, at a minimum:

The ZooKeeper quorum could be placed together on SOLR search servers or, optionally, you could break out the ZooKeeper configuration into an additional 3 separate servers.

The following diagram represents the minimum architecture required to support a full CLAIMS Direct index.

Currently, the delivery of a fully populated CLAIMS Direct index requires the above SOLR and ZooKeeper configuration (8 SOLR servers + 3 ZooKeepers). Load balancers and web servers are required only if CLAIMS Direct Web Services (CDWS) will be installed as well. A customized deployment with select data to index is currently not offered fully populated. With a custom configuration, complete indexing will need to be done at the installation site.

Software

The CLAIMS Direct SOLR installation is a self-contained package suitable for deployment on any Linux server running Java 8. The simple prerequisite tool list follows:

The configuration script setup.sh assumes that each node in the cluster will have the same directory structure. For example, if you download to a machine and unpack the archive into path /cdsolr, the full path to the package will be /cdsolr/alexandria-solr-v2.1.2-distribution. Each node must have the path /cdsolr available for deployment. You are free to choose any mount point or path as long as they are uniform across all nodes in the cluster and as long as the mount point or path for each SOLR node has at least 1TB of available disk space.

Create User asolr

It is recommended to create the user asolr.

useradd -m asolr
passwd asolr
  => <password>

Configuring SolrCloud

CLAIMS Direct SOLR uses the Apache SOLR distribution for indexing. The fully populated delivery of the index is an 8-node collection as described above. The collection name is alexandria and although configurable, it is recommended not to change this setting.

Configuration

The following table lists the applicable nodes and the function in the environment:

The following variables should be configured in solr-alexandria-vars:

Configuring ZooKeeper

ZooKeeper is a distributed, open-source coordination service for distributed applications. It is an essential component in a functioning CLAIMS Direct SolrCloud environment. The CLAIMS Direct SolrCloud distribution comes bundled with a preconfigured, 3-node ZooKeeper quorum.

Configuration

A CLAIMS Direct ZooKeeper deployment requires 3 individual servers. These servers could be dedicated machines or they could share duties with 3 nodes in the SolrCloud collection. We will assume the following SOLR/ZooKeeper environment:

The following variables should be configured in solr-alexandria-vars:

Preparing Deployment

After ZooKeeper and SolrCloud node configuration is complete, the following prerequisites need to be installed and configured on each node:

Deployment

Running setup.sh.

The steps taken by setup.sh are described in sections that can be managed manually if desired.

Configuration Creation

The first step is to create the configuration files for both SOLR as well as ZooKeeper based on the main variable configuration file solr-alexandria-vars. setup.sh uses the following variables to create ZooKeeper configuration files for each node of the quorum. Note: italic entries should not be modified unless absolutely necessary.

In addition, a valid solr.xml is generated from the template: solrxml.in. The following variables are used:

Code Deployment

The next section in setup.sh deploys the code and configuration to all nodes. If you haven't enabled public key ssh access, the following commands need to be issued manually for each IP address in ALEXANDRIA_SOLR_CLOUD_NODES and ALEXANDRIA_SOLR_ZK_NODES:

rsync -va * $ALEXANDRIA_SOLR_USER@$IPADDRESS:$ALEXANDRIA_SOLR_HOME
# e.g.
rsync -va * asolr@10.234.1.91:/home/asolr/alexandria-solr-v2.1.2-distribution

Example Walk-Through

Assumptions:

Create user on all machines

# as root on all machines 10.234.1.91-99
useradd -m asolr
passwd asolr # create your own password

Enable public key ssh access

# as user asolr on 10.234.1.99 (ap01)
cd ~/.ssh
ssh-keygen
# for each node from 10.234.1.91 -> .98
ssh-copy-id -i id_rsa.pub asolr@10.234.1.91
ssh-copy-id -i id_rsa.pub asolr@10.234.1.92
ssh-copy-id -i id_rsa.pub asolr@10.234.1.93
etc ...

Download package onto ap01 (10.234.1.99) and unpack into /home/asolr

cd /home/asolr
wget http://alexandria.fairviewresearch.com/software/alexandria-solr-v2.1.2-distribution.tar.gz
tar zxvf alexandria-solr-v2.1.2-distribution.tar.gz
cd /home/asolr/alexandria-solr-v2.1.2-distribution

Adjust solr-alexandria-vars 

The following variables need to be confirmed or set:

Run setup.sh

$ ./setup.sh 
Initializing ZooKeeper 3-node quorum configuration ...
  * -> zoo-1.cfg
  * -> zoo-2.cfg
  * -> zoo-3.cfg
Initializing solr.xml ... done
Testing ssh access to solr:
  * -> node: asolr@10.234.1.91 ... ok
  * -> node: asolr@10.234.1.92 ... ok
  * -> node: asolr@10.234.1.93 ... ok
  * -> node: asolr@10.234.1.94 ... ok
  * -> node: asolr@10.234.1.95 ... ok
  * -> node: asolr@10.234.1.96 ... ok
  * -> node: asolr@10.234.1.97 ... ok
  * -> node: asolr@10.234.1.98 ... ok
Testing ssh access to zookeeper:
  * -> node: asolr@10.234.1.91 ... ok
  * -> node: asolr@10.234.1.92 ... ok
  * -> node: asolr@10.234.1.93 ... ok
Deploying SOLR ...
  * -> asolr@10.234.1.91:/home/asolr ... ok
  * -> asolr@10.234.1.92:/home/asolr ... ok
  * -> asolr@10.234.1.93:/home/asolr ... ok
  * -> asolr@10.234.1.94:/home/asolr ... ok
  * -> asolr@10.234.1.95:/home/asolr ... ok
  * -> asolr@10.234.1.96:/home/asolr ... ok
  * -> asolr@10.234.1.97:/home/asolr ... ok
  * -> asolr@10.234.1.98:/home/asolr ... ok
Deploying ZooKeeper ...
  * -> asolr@10.234.1.91:/home/asolr ... ok
  * -> asolr@10.234.1.92:/home/asolr ... ok
  * -> asolr@10.234.1.93:/home/asolr ... ok

Starting and Bootstrapping the ZooKeeper Quorum

./zookeeperctl start
Starting zookeeper ... STARTED
Starting zookeeper ... STARTED
Starting zookeeper ... STARTED


# check the quorum is running:
./zookeeperctl status
Mode: follower
Mode: follower
Mode: leader


# install SOLR configuration files
./bootstrap-zookeeper.sh 
Bootstrapping 10.234.1.91 ...
  -> name: cdidx-v2.1.2
  -> directory: /home/asolr/alexandria-solr-v2.1.2-distribution/conf-2.1.2

Starting and Bootstrapping the SOLR Collection

 $ ./solrctl start
Initializing solr.xml ... done
asolr@10.234.1.91: SOLR ( start ) ...Waiting up to 30 seconds to see Solr running on port 8080 [/]  
Started Solr server on port 8080 (pid=10520). Happy searching!
### etc ...
 
# check cluster status
./solrctl status
asolr@10.234.1.91: SOLR ( status ) ...
Found 1 Solr nodes: 
Solr process 11273 running on port 8080
{
  "solr_home":"/home/asolr/alexandria-solr-v2.1.2distribution",
  "version":"5.3.1 1703449 - noble - 2015-09-17 01:48:15",
  "startTime":"2016-02-23T08:18:36.059Z",
  "uptime":"0 days, 0 hours, 1 minutes, 9 seconds",
  "memory":"63.6 MB (%6.5) of 981.4 MB",
  "cloud":{
    "ZooKeeper":"10.234.1.91:2181,10.234.1.92:2181,10.234.1.93:2181",
    "liveNodes":"8",
    "collections":"0"}}
### etc.

# create the collection
./bootstrap-alexandria.sh
 
Creating (action=CREATE) on http://10.234.1.91:8080/alexandria-v2.1/admin/collections ...
  -> name=alexandria
  -> numShards=8
  -> replicationFactor=1
  -> maxShardsPerNode=1
  -> collection.config=cdidx-v2.1.2
  -> property.config=solrconfig-alexandria.xml
  HTTP/1.1 200 OK
  Content-Type: application/xml; charset=UTF-8
  Transfer-Encoding: chunked
<?xml version="1.0" encoding="UTF-8"?>
<response>
 <lst name="responseHeader">
  <int name="status">0</int>
  <int name="QTime">3136</int>
 </lst>
  # etc ..
</response> 

SOLR Interface

At this point, the ZooKeeper quorum and all SOLR nodes should be running. You can now visit http://10.234.1.91:8080/alexandria-v2.1/old.html

Stopping SOLR Cluster and ZooKeeper Quorum

./solrctl stop
asolr@10.234.1.91: SOLR ( stop ) ...
# etc.
 
./zookeeperctl stop