To use HDFS rather than a local filesystem, you must be using Hadoop 2.x and you will need to instruct Solr to use the `HdfsDirectoryFactory`. There are also several additional parameters to define. These can be set in one of three ways:
* Pass JVM arguments to the `bin/solr` script. These would need to be passed every time you start Solr with `bin/solr`.
* Modify `solr.in.sh` (or `solr.in.cmd` on Windows) to pass the JVM arguments automatically when using `bin/solr` without having to set them manually.
* Define the properties in `solrconfig.xml`. These configuration changes would need to be repeated for every collection, so is a good option if you only want some of your collections stored in HDFS.
For standalone Solr instances, there are a few parameters you should modify before starting Solr. These can be set in `solrconfig.xml` (more on that <<HdfsDirectoryFactory Parameters,below>>), or passed to the `bin/solr` script at startup.
This example will start Solr in standalone mode, using the defined JVM properties (explained in more detail <<HdfsDirectoryFactory Parameters,below>>).
In SolrCloud mode, it's best to leave the data and update log directories as the defaults Solr comes with and simply specify the `solr.hdfs.home`. All dynamically created collections will create the appropriate directories automatically under the `solr.hdfs.home` root directory.
* Set `solr.hdfs.home` in the form `hdfs://host:port/path`
This command starts Solr in SolrCloud mode, using the defined JVM properties.
=== Modifying solr.in.sh (*nix) or solr.in.cmd (Windows)
The examples above assume you will pass JVM arguments as part of the start command every time you use `bin/solr` to start Solr. However, `bin/solr` looks for an include file named `solr.in.sh` (`solr.in.cmd` on Windows) to set environment variables. By default, this file is found in the `bin` directory, and you can modify it to permanently add the `HdfsDirectoryFactory` settings and ensure they are used every time Solr is started.
For example, to set JVM arguments to always use HDFS when running in SolrCloud mode (as shown above), you would add a section such as this:
For performance, the `HdfsDirectoryFactory` uses a Directory that will cache HDFS blocks. This caching mechanism replaces the standard file system cache that Solr utilizes. By default, this cache is allocated off-heap. This cache will often need to be quite large and you may need to raise the off-heap memory limit for the specific JVM you are running Solr in. For the Oracle/OpenJDK JVMs, the following is an example command-line parameter that you can use to raise the limit when starting Solr:
A root location in HDFS for Solr to write collection data to. Rather than specifying an HDFS location for the data directory or update log directory, use this to specify one root location and have everything automatically created within this HDFS location. The structure of this parameter is `hdfs://host:port/path/solr`.
Hadoop can be configured to use the Kerberos protocol to verify user identity when trying to access core services like HDFS. If your HDFS directories are protected using Kerberos, then you need to configure Solr's HdfsDirectoryFactory to authenticate using Kerberos in order to read and write to HDFS. To enable Kerberos authentication from Solr, you need to set the following parameters:
A keytab file contains pairs of Kerberos principals and encrypted keys which allows for password-less authentication when Solr attempts to authenticate with secure Hadoop.
The Kerberos principal that Solr should use to authenticate to secure Hadoop; the format of a typical Kerberos V5 principal is: `primary/instance@realm`.
The ability to automatically add new replicas when the Overseer notices that a shard has gone down was previously only available to users running Solr in HDFS, but it is now available to all users via Solr's autoscaling framework. See the section <<solrcloud-autoscaling-auto-add-replicas.adoc#the-autoaddreplicas-parameter,SolrCloud Autoscaling Automatically Adding Replicas>> for details on how to enable and disable this feature.
The ability to enable or disable the autoAddReplicas feature with cluster properties has been deprecated and will be removed in a future version. All users of this feature who have previously used that approach are encouraged to change their configurations to use the autoscaling framework to ensure continued operation of this feature in their Solr installations.
For users using this feature with the deprecated configuration, you can temporarily disable it cluster-wide by setting the cluster property `autoAddReplicas` to `false`, as in these examples: