Development repository for Chef's Splunk cookbook
Clone or download
tas50 Merge pull request #105 from mskdenigma/master
Added more inputs.conf flexibility
Latest commit c37d72c Jul 24, 2018
Permalink
Failed to load latest commit information.
.delivery Test with Local Delivery and not Rake Mar 1, 2017
.github Update Github PR template Oct 23, 2016
attributes enable amazon platform support for splunk forwarder Aug 28, 2017
definitions enable amazon platform support for splunk forwarder Aug 28, 2017
files/default A polar bear ice fishing for data. Jan 11, 2014
libraries Disable FC113 for now May 16, 2018
recipes Set ownership of web.conf file. Feb 8, 2018
spec Update specs to the latest platform versions May 16, 2018
templates [GH-36] - Removed restriction on input types Jan 24, 2018
test Merge remote-tracking branch 'upstream/master' into multisite-cluster May 23, 2017
.foodcritic add foodcritic file to ignore definition rule Oct 24, 2014
.gitignore Test with Local Delivery and not Rake Mar 1, 2017
.kitchen.dokken.yml Update kitchen configs Sep 25, 2017
.kitchen.yml Update kitchen configs Sep 25, 2017
.rubocop.yml Update rubocop exclusion Apr 11, 2017
.travis.yml Simplify Travis config and fix ChefDK 2.0 failures Jun 28, 2017
Berksfile Update platforms to test with test kitchen Jan 26, 2017
CHANGELOG.md Release 1.7.3 Apr 27, 2018
CONTRIBUTING.md Update docs Sep 18, 2015
Gemfile Testing updates and document Chef 12.1+ requirement Mar 7, 2017
LICENSE Update license file Mar 14, 2016
README.md [GH-36] - Removed restriction on input types Jan 24, 2018
TESTING.md Update docs Sep 18, 2015
chefignore Update ignore files Mar 14, 2016
metadata.rb Release 1.7.3 Apr 27, 2018

README.md

chef-splunk Cookbook

Build Status Cookbook Version

This cookbook manages a Splunk Universal Forwarder (client) or a Splunk Enterprise (server) installation, including a Splunk clustered environment.

The Splunk default user is admin and the password is changeme. See the setup_auth recipe below for more information about how to manage changing the password with Chef and Chef Vault.

This recipe downloads packages from Splunk directly. There are attributes to set a URL to retrieve the packages, so if the packages are mirrored locally, supply the local URL instead. At this time the cookbook doesn't support installing from networked package managers (like apt or yum), since Splunk doesn't provide package repositories.

Requirements

Chef 12.1+

Platforms

This cookbook uses Test Kitchen to do cross-platform convergence and post-convergence tests. The tested platforms are considered supported. This cookbook may work on other platforms or platform versions with or without modification.

  • Debian 7
  • Ubuntu 12.04, 14.04
  • CentOS 6
  • OmniOS r151008

Cookbooks

Used for managing secrets, see Usage:

  • chef-vault

Attributes

Attributes have default values set in attributes/default.rb. Where possible or appropriate, the default values from Splunk Enterprise are used.

General attributes:

  • node['splunk']['accept_license']: Whether to accept the Splunk EULA. Default is false. This must be set to true for Splunk to be functional with this cookbook, which means end users must read the EULA and agree to the terms.
  • node['splunk']['is_server']: Set this to true if the node is a splunk server, for example in a role. Default is false.
  • node['splunk']['disabled']: Disable the splunk agent by setting this to true. Default is false.
  • node['splunk']['receiver_port']: The port that the receiver (server) listens to. This is set to the Splunk Enterprise default, 9997.
  • node['splunk']['mgmt_port']: The port that splunkd service listens to, aka the management port. This is set to the Splunk Enterprise default, 8089.
  • node['splunk']['web_port']: The port that the splunkweb service listens to. This is set to the default for HTTPS, 443, as it is configured by the setup_ssl recipe.
  • node['splunk']['ratelimit_kilobytessec']: The default splunk rate limiting rate can now easily be changed with an attribute. Default is 2048KBytes/sec.

The two URL attributes below are selected by platform and architecture by default.

  • node['splunk']['forwarder']['url']: The URL to the Splunk Universal Forwarder package file.
  • node['splunk']['server']['url']: The URL to the Splunk Enterprise package file.

Special attributes for managing the Splunk user:

  • node['splunk']['user']: A hash of attributes to set for the splunk user resource in the user recipe. It's unlikely that someone would need to change these, other than the UID, but just in case...
  • username: the username
  • comment: gecos field
  • home: the home directory, defaults to /opt/splunkforwarder, will be set to /opt/splunk if node['splunk']['is_server'] is true.
  • shell: the shell to use
  • uid: the numeric UID. The default, 396 is an integer arbitrarily chosen and doesn't conflict with anything on the supported platforms (see list above). It is within the system UID range on Linux systems.
  • node['splunk']['server']['runasroot']: if runasroot is true (which is the splunk upstream package default) then the splunk server runs as root. If runasroot is false modify the init script to run as the node['splunk']['user']. This does not apply to the splunk client as they may need root permissions to read logfiles. NOTE1: you may also need to change node['splunk']['web_port'] on a splunk server to run on a port >1024 if you don't run as root (splunk user cannot bind to privelaged ports). NOTE2: If you want to switch from root to the splunk user or vice versa on an existing install, please stop the splunk service first before changing the runasroot boolean value.

The following attributes are related to setting up splunkweb with SSL in the setup_ssl recipe.

  • node['splunk']['ssl_options']: A hash of SSL options used in the setup_ssl recipe
  • node['splunk']['ssl_options']['enable_ssl']: Whether to enable SSL, must be set to true to use the setup_ssl recipe. Defaults to false, must be set using a boolean literal true or false.
  • node['splunk']['ssl_options']['data_bag']: The data bag name to load, defaults to vault (as chef-vault is used).
  • node['splunk']['ssl_options']['data_bag_item']: The data bag item name that contains the keyfile and crtfile, defaults to splunk_certificates.
  • node['splunk']['ssl_options']['keyfile']: The name of the SSL key file, and the content will be written to etc/auth/splunkweb/KEYFILE. Must be an element under data in the data bag item. See Usage for instructions. Defaults to 'self-signed.example.com.key', and should be changed to something relevant for the local site before use, in a role or wrapper cookbook.
  • node['splunk']['ssl_options']['crtfile']: The name of the SSL cert (crt) file, and the content will be written to /etc/auth/splunkweb/CRTFILE. Must be an element under data in the data bag item. See Usage for instructions. Defaults to 'self-signed.example.com.crt', and should be changed to something relevant for the local site before use, in a role or wrapper cookbook.

The following attributes are related to setting up a Splunk server with indexer clustering in the setup_clustering recipe:

  • node['splunk']['clustering']: A hash of indexer clustering configurations used in the setup_clustering recipe

  • node['splunk']['clustering']['enabled']: Whether to enable indexer clustering, must be set to true to use the setup_clustering recipe. Defaults to false, must be a boolean literal true or false.

  • node['splunk']['clustering']['num_sites']: The number of sites in the cluster. Multisite is enabled automatically if num_sites > 1. Defaults to 1, must be a positive integer.

  • node['splunk']['clustering']['mode']: The clustering mode of the node within the indexer cluster. Must be set using string literal 'master', 'slave', or 'searchhead'.

  • node['splunk']['clustering']['replication_port']: The replication port of the cluster peer member. Only valid when node['splunk']['clustering']['mode']='slave'. Defaults to 9887.

  • For single-site clustering (node['splunk']['clustering']['num_sites'] = 1):

    • node['splunk']['clustering']['replication_factor']: The replication factor of the indexer cluster. Defaults to 3, must be a positive integer. Only valid when node['splunk']['clustering']['mode']='master' and node['splunk']['clustering']['num_sites']=1 (single-site clustering).
    • node['splunk']['clustering']['search_factor']: The search factor of the indexer cluster. Only valid when node['splunk']['clustering']['mode']='master' and node['splunk']['clustering']['num_sites']=1 (single-site clustering). Defaults to 2, must be a positive integer.
  • For multisite clustering (node['splunk']['clustering']['num_sites'] > 1):

    • node['splunk']['clustering']['site']: The site the node belongs to. Valid values include site1 to site63
    • node['splunk']['clustering']['site_replication_factor']: The per-site replication policy of any given bucket. This is represented as a comma-separated list of per-site entries. Only valid when node['splunk']['clustering']['mode']='master' and multisite is true. Defaults to 'origin:2,total:3'. Refer to Splunk Admin docs for exact syntax and more details.
    • node['splunk']['clustering']['site_search_factor']: The per-site search policy for searchable copies for any given bucket. This is represented as a comma-separated list of per-site entires. Only valid when node['splunk']['clustering']['mode']='master' and multisite is true. Defaults to 'origin:1,total:2'. Refer to Splunk Admin docs for exact syntax and more details.

The following attributes are related to setting up a Splunk server with search head clustering in the setup_shclustering recipe:

  • node['splunk']['shclustering']: A hash of search head clustering configurations used in the setup_shclustering recipe
  • node['splunk']['shclustering']['enabled']: Whether to enable search head clustering, must be set to true to use the setup_shclustering recipe. Defaults to false, must be a boolean literal true or false.
  • node['splunk']['shclustering']['mode']: The search head clustering mode of the node within the cluster. This is used to determine if the node needs to bootstrap the shcluster and initialize the node as the captain. Must be set using string literal 'member' or 'captain'.
  • node['splunk']['shclustering']['label']: The label for the shcluster. Used to differentiate from other shclusters in the environment. Must be a string. Defaults to shcluster1. captain election. Must be set using string literal 'member' or 'captain'.
  • node['splunk']['shclustering']['replication_factor']: The replication factor of the shcluster. Defaults to 3, must be a positive integer.
  • node['splunk']['shclustering']['replication_port']: The replication port of the shcluster members. Defaults to 9900.
  • node['splunk']['shclustering']['deployer_url']: The management url for the shcluster deployer server, must be set to a string such as: https://deployer.domain.tld:8089. This attribute is optional. Defaults to empty.
  • node['splunk']['shclustering']['mgmt_uri']: The management url for the shcluster member node, must be set to a string such as: https://shx.domain.tld:8089. You can use the node's IP address instead of the FQDN if desired. Defaults to https://#{node['fqdn']}:8089.
  • node['splunk']['shclustering']['shcluster_members']: An array of all search head cluster members referenced by their mgmt_uri. Currently this will do a Chef search for nodes that are in the same environment, with search head clustering enabled, and with the same cluster label. Alternatively, this can be hard-coded with a list of all shcluster members including the current node. Must be an array of strings. Defaults to an empty array.

The following attributes are related to setting up a splunk forwarder with the client recipe

node['splunk']['outputs_conf'] is a hash of configuration values that are used to dynamically populate the outputs.conf file's "tcpout:splunk_indexers_PORT" configuration section. Each key/value pair in the hash is used as configuration in the file. For example the attributes/default.rb has this:

default['splunk']['outputs_conf'] = {
  'forwardedindex.0.whitelist' => '.*',
  'forwardedindex.1.blacklist' => '_.*',
  'forwardedindex.2.whitelist' => '_audit',
  'forwardedindex.filter.disable' => 'false'
}

This will result in the following being rendered in outputs.conf:

[tcpout:splunk_indexers_9997]
server=10.0.2.47:9997
forwardedindex.0.whitelist = .*
forwardedindex.1.blacklist = _.*
forwardedindex.2.whitelist = _audit
forwardedindex.filter.disable = false

As an example of outputs_conf attribute usage, to add an sslCertPath directive, define the attribute in your role or wrapper cookbook as such:

node.default['splunk']['outputs_conf']['sslCertPath'] = '$SPLUNK_HOME/etc/certs/cert.pem'

The server attribute in tcpout:splunk_indexers_9997 stanza above is populated by default from Chef search results for Splunk servers, or, alternatively, is statically defined in node attribute node['splunk']['server_list'].

node['splunk']['server_list'] is an optional comma-separated listed of server IPs and the ports. It's only applicable when there are no Splunk servers managed by Chef, e.g. sending data to Splunk Cloud which has managed indexers.

For example:

node.default['splunk']['server_list'] = '10.0.2.47:9997, 10.0.2.49:9997'

node['splunk']['inputs_conf'] is a hash of configuration values that are used to populate the inputs.conf file.

  • node['splunk']['inputs_conf']['host']: A string that specifies the default host name used in the inputs.conf file. The inputs.conf file is not overwritten if this is not set or is an empty string.
  • node['splunk']['inputs_conf']['ports']: An array of hashes that contain the input port configuration necessary to generate the inputs.conf file.
  • node['splunk']['inputs_conf']['inputs']: An array of hashes that contain the input configuration necessary to generate the inputs.conf file. This attribute supports all input types.

For example:

node.default['splunk']['inputs_conf']['ports'] = [
  {
    port_num => 123123,
    config => {
      'sourcetype' => 'syslog'
    }
  }
]

node.default['splunk']['inputs_conf']['inputs'] = [
  {
    input_path => 'monitor:///var/log/syslog',
    config => {
      'sourcetype' => 'syslog'
    }
  }
]

The following attributes are related to upgrades in the upgrade recipe. Note The version is set to 4.3.7 and should be modified to suit in a role or wrapper, since we don't know what upgrade versions may be relevant. Enabling the upgrade and blindly using the default URLs may have undesirable consequences, hence this is not enabled, and must be set explicitly elsewhere on the node(s).

  • node['splunk']['upgrade_enabled']: Controls whether the upgrade is enabled and the attributes/upgrade.rb file should be loaded. Set this in a role or wrapper cookbook to perform an upgrade.
  • node['splunk']['upgrade']: Sets server_url and forwarder_url attributes based on platform and architecture. These are only loaded if upgrade_enabled is set.

Definitions

splunk_installer

The Splunk Enterprise and Splunk Universal Forwarder package installation is the same save the name of the package and the URL to download. This definition abstracts the package installation to a common baseline. Any new platform installation support should be added by modifying the definition as appropriate. One goal of this definition is to have a single occurance of a package resource, using the appropriate "local package file" provider per platform. For example, on RHEL, we use rpm and on Debian we use dpkg.

Package files will be downloaded to Chef's file cache path (e.g., file_cache_path in /etc/chef/client.rb, /var/chef/cache by default).

The definition has two parameters.

  • name: The name of the package (e.g., splunk).
  • url: The URL to the package file.

Examples

For example, if the nodes in the environment are all Debian-family, and the desired splunkforwarder package is provided locally as splunkforwarder.deb on an internal HTTP server:

splunk_installer 'splunkforwarder' do
  url 'https://www-int.example.com/splunk/splunkforwarder.deb'
end

The install_forwarder and install_server recipes use the definition with the appropriate url attribute.

Recipes

This cookbook has several composable recipes that can be used in a role, or a local "wrapper" cookbook. The default, client, and server recipes are intended to be used wholesale with all the assumptions they contain.

The general default assumption is that a node including the default recipe will be a Splunk Universal Forwarder (client).

client

This recipe encapsulates a completely configured "client" - a Splunk Universal Forwarder configured to talk to a node that is the splunk server (with node['splunk']['is_server'] true). The recipes can be used on their own composed in a wrapper cookbook or role. This recipe will include the user, install_forwarder, service, and setup_auth recipes.

It will also search a Chef Server for a Splunk Enterprise (server) node with splunk_is_server:true in the same chef_environment and write out etc/system/local/outputs.conf with the server's IP and the receiver_port attribute in the Splunk install directory (/opt/splunkforwarder).

Setting node['splunk']['outputs_conf'] with key value pairs updates the outputs.conf server configuration with those key value pairs. These key value pairs can be used to setup SSL encryption on messages forwarded through this client:

# Note that the ssl CA and certs must exist on the server.
node['splunk']['outputs_conf'] = {
  'sslCommonNameToCheck' => 'sslCommonName',
  'sslCertPath' => '$SPLUNK_HOME/etc/certs/cert.pem',
  'sslPassword' => 'password'
  'sslRootCAPath' => '$SPLUNK_HOME/etc/certs/cacert.pem'
  'sslVerifyServerCert' => false
}

The inputs.conf file can also be managed through this recipe if you want to setup a splunk forwarder just set the default host:

node['splunk']['inputs_conf']['host'] = 'myhost'

Then set up the port configuration for each input port:

node['splunk']['inputs_conf']['ports'] =
[
  {
    port_num => 123123,
    config => {
      'sourcetype' => 'syslog',
      ...
    }
  },
  ...
]

default

The default recipe will include the disabled recipe if node['splunk']['disabled'] is true.

It will include the client or server recipe depending on whether the is_server attribute is set.

The attribute use allows users to control the included recipes by easily manipulating the attributes of a node, or a node's roles, or through a wrapper cookbook.

disabled

In some cases it may be required to disable Splunk on a particular node. For example, it may be sending too much data to Splunk and exceed the local license capacity. To use the disabled recipe, set the node['splunk']['disabled'] attribute to true, and include the recipe on the required node, or just use the default recipe.

install_forwarder

This recipe uses the splunk_installer definition to install the splunkforwarder package from the specified URL (via the node['splunk']['forwarder']['url'] attribute).

install_server

This recipe uses the splunk_installer definition to install the splunk (Enterprise server) package from the specified URL (via the node['splunk']['server']['url'] attribute).

server

This recipe encapsulates a completely configured "server" - Splunk Enterprise configured to receive data from Splunk Universal Forwarder clients. The recipe sets the attribute node['splunk']['is_server'] to true, and is included from the default recipe if the attribute is true as well. The recipes can be used on their own composed in a wrapper cookbook or role, too. This recipe will include the user, install_server, service, and setup_auth recipes. It will also conditionally include the setup_ssl and setup_clustering recipes if enabled via the corresponding node attributes, as defined in Attributes above.

It will also enable Splunk Enterprise as an indexer, listening on the node['splunk']['receiver_port'].

service

This recipe sets up the splunk service, and applies to both client and server use, since splunk is the same service for both deployments of Splunk.

The attribute node['splunk']['accept_license'] must be true in order to set up the boot script. If it's true, then the boot script gets put into place (/etc/init.d/splunk on Linux/Unix systems), with the license accepted. The service is managed using the Chef init service provider, which operates by using the /etc/init.d/splunk script for start, stop, restart, etc commands.

setup_auth

This recipe loads an encrypted data bag with the Splunk user credentials as an -auth string, 'user:password', using the chef-vault cookbook helper method, chef_vault_item. See Usage for how to set this up. The recipe will edit the specified user (assuming admin), and then write a state file to etc/.setup_admin_password to indicate in future Chef runs that it has set the password. If the password should be changed, then that file should be removed.

setup_clustering

This recipe sets up Splunk indexer clustering based on the node's clustering mode or node['splunk']['clustering']['mode']. The attribute node['splunk']['clustering']['enabled'] must be set to true in order to run this recipe. Similar to setup_auth, this recipes loads the same encrypted data bag with the Splunk secret key (to be shared among cluster members), using the chef-vault cookbook helper method, chef_vault_item. See Usage for how to set this up. The recipe will edit the cluster configuration, and then write a state file to etc/.setup_cluster_{master|slave|searchhead} to indicate in future Chef runs that it has set the node's indexer clustering configuration. If cluster configuration should be changed, then that file should be removed.

It will also search a Chef Server for a Splunk Enterprise (server) node of type cluster master, that is with splunk_clustering_enable:true and splunk_clustering_mode:master in the same chef_environment and use that server's IP when configuring a cluster search head or a cluster peer node to communicate with the cluster master (Refer to master_uri attribute of clustering stanza in etc/system/local/server.conf).

Indexer clustering is used to achieve some data availability & recovery. To learn more about Splunk indexer clustering, refer to Splunk Docs.

setup_shclustering

This recipe sets up Splunk search head clustering. The attribute node['splunk']['shclustering']['enabled'] must be set to true in order to run this recipe. Similar to setup_auth, this recipes loads the same encrypted data bag with the Splunk secret key (to be shared among cluster members), using the chef-vault cookbook helper method, chef_vault_item. See Usage for how to set this up. The recipe will edit the cluster configuration, and then write a state file to etc/.setup_shcluster to indicate in future Chef runs that it has set the node's search head clustering configuration. If cluster configuration should be changed, then that file should be removed.

It will also search a Chef Server for a Splunk Enterprise (server) node of type cluster master, that is with splunk_shclustering_enable:true and the same splunk_shclustering_label in the same chef_environment and use that server's IP when building the list of shcluster_members.

The search head cluster configuration is deployed as a custom Splunk app that is written to etc/apps/0_autogen_shcluster_config to take advantage of Splunk's built in config layering. All nodes with splunk_shclustering_enable:true will receive this app.

On the first Chef run on a node with splunk_shclustering_mode:captain, this recipe will build and execute the Splunk command to bootstrap the search head cluster and initiate the captain election process.

In addition to using this recipe for configuring the search head cluster members, you will also have to manually configure a search head instance to serve as the search head cluster's deployer. This is done by adding a [shclustering] stanza to that instance's etc/system/local/server.conf with the same pass4SymmKey = <secret> and the same shcluster_label = <splunk_shclustering_label>. This deployer is optional, but should be configured prior to running the bootstrap on the captain and then the search head cluster member nodes configured with this deployer node's mgmt_uri set in the member node's splunk_shclustering_deployer_url

Search head clustering is used to achieve high availability & scaling. To learn more about Splunk search head clustering, refer to Splunk Docs.

upgrade

Important Read the upgrade documentation and release notes for any particular Splunk version upgrades before performing an upgrade. Also back up the Splunk directory, configuration, etc.

This recipe can be used to upgrade a splunk installation, for example from an existing 4.2.1 to 4.3.7. The default recipe can be used for 6.0.1 after upgrading earlier versions is completed. Note that the attributes file is only loaded w/ the URLs to the splunk packages to upgrade if the node['splunk']['upgrade_enabled'] attribute is set to true. We recommend setting the actual URL attributes needed in a wrapper cookbook or role.

user

This recipe manages the splunk user and group. On Linux systems, the user and group will be created with the system attribute; other platforms may not be aware of system users/groups (e.g., illumos/solaris). Both resources will be created with the UID or GID of the node['splunk']['user']['uid'] attribute. The default value is 396, arbitrarily chosen to fall under the system UID/GID set by /etc/login.defs on both RHEL and Debian family Linux systems. If this is a conflicting UID/GID, then modify the attribute as required.

Usage

Data Bag Items

Splunk Secrets & Admin User Authentication

Splunk secret key and admin user authentication information should be stored in a data bag item that is encrypted using Chef Vault. Create a data bag named vault, with an item splunk_CHEF-ENVIRONMENT, where CHEF-ENVIRONMENT is the node.chef_environment that the Splunk Enterprise server will be assigned. If environments are not used, use _default. For example in a Chef Repository (not in a cookbook):

% cat data_bags/vault/splunk__default.json
{
  "id": "splunk__default",
  "auth": "admin:notarealpassword",
  "secret": "notarealsecret"
}

Or with an environment, 'production':

% cat data_bags/vault/splunk_production.json
{
  "id": "splunk_production",
  "auth": "admin:notarealpassword",
  "secret": "notarealsecret"
}

Then, upload the data bag item to the Chef Server using the chef-vault knife encrypt plugin (first example, _default environment):

knife encrypt create vault splunk__default \
    --json data_bags/vault/splunk__default.json \
    --search 'splunk:*' --admins 'yourusername' \
    --mode client

More information about Chef Vault is available on the GitHub Project Page.

Web UI SSL

A Splunk server should have the Web UI available via HTTPS. This can be set up using self-signed SSL certificates, or "real" SSL certificates. This loaded via a data bag item with chef-vault. Using the defaults from the attributes:

% cat data_bags/vault/splunk_certificates.json
{
  "id": "splunk_certificates",
  "data": {
    "self-signed.example.com.crt": "-----BEGIN CERTIFICATE-----\n...SNIP",
    "self-signed.example.com.key": "-----BEGIN RSA PRIVATE KEY-----\n...SNIP"
  }
}

Like the authentication credentials above, run the knife encrypt command. Note the search here is for the splunk server only:

knife encrypt create vault splunk_certificates \
    --json data_bags/vault/splunk_certificates.json \
    --search 'splunk_is_server:true' --admins 'yourusername' \
    --mode client

License and Authors

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.