The i2 Analyze configuration specifies
the structure of your Solr cluster. To provide
high availability, configure a Solr cluster over
at least two Solr servers.
About this task
In the i2 Analyze configuration, provide the
details about the Solr servers in your
environment. To provide high availability of the
content of the Solr indexes, specify the number of
replicas, the minimum replication factor, and the
location of the replicas.
For more information about SolrCloud, see How SolrCloud
Works.
Procedure
- Specifying the Solr
nodes.
To deploy Solr for high
availability, you must have at least two Solr
nodes on separate hosts.
- Add the Solr nodes to your
topology.xml
file.
For
example:
<solr-nodes>
<solr-node
memory="2g"
data-dir="C:/i2/i2analyze/data/solr"
host-name="solr_server1_host_name"
id="node1"
port-number="8983"
/>
<solr-node
memory="2g"
data-dir="C:/i2/i2analyze/data/solr"
host-name="solr_server2_host_name"
id="node2"
port-number="8983"
/>
</solr-nodes>
Where
solr_serverx_host_name is the
hostname of a Solr server.
For more information
about the possible values for each attribute, see
Solr.
- Configuring Solr
replicas.
In Solr, the data is stored
as documents in shards.
Every shard consists of at least one
replica. For a highly available
solution, you must have more than one replica of
each shard and these replicas must be distributed
across the servers that host the Solr
nodes.
- Specifying the replication
factor in your
topology.xml
.The
replication factor is the number of replicas to be
created for each shard. For high availability,
this must be 2 or more. You specify the
replication factor in the
num-replicas
attribute of the
<solr-collection>
element.
- Specifying the minimum
replication factor in your
topology.xml
.The
minimum replication factor defines when data is
successfully replicated in Solr. If you have three
replicas for a shard and a minimum replication
factor of 2, a write operation is deemed
successful if the data is written to at least two
replicas. You specify the minimum replication
factor in the
min-replication-factor
attribute
of the <solr-collection>
element.
The following extract from a
topology.xml
file shows an
example of the
num-replicas
and
min-replication-factor
attributes:
<solr-collections>
<solr-collection
num-replicas="2"
min-replication-factor="1"
id="main_index"
type="main"
max-shards-per-node="4"
num-shards="1"
/>
...
</solr-collections>
- Specify the Solr autoscaling
policies.
To ensure that your system
can still operate if a Solr server or node fails,
you must ensure that there is at least on replica
of each shard still available. To do this, you can
create a replica placement policy to specify that
each replica of a shard must be placed on a
distinct host.
What to do next
Continue configuring the i2 Analyze
configuration. For more information, see Deploying i2 Analyze with high availability.