salt.pillar.mongo#

Read Pillar data from a mongodb collection

depends:

pymongo (for salt-master)

This module will load a node-specific pillar dictionary from a mongo collection. It uses the node's id for lookups and can load either the whole document, or just a specific field from that document as the pillar dictionary.

Salt Master Mongo Configuration#

The module shares the same base mongo connection variables as salt.returners.mongo_future_return. These variables go in your master config file.

mongo.db: <database name>
mongo.host: <server ip address>
mongo.user: <MongoDB username>
mongo.password: <MongoDB user password>
mongo.port: 27017

Or single URI:

mongo.uri: URI

where uri is in the format:

mongodb://[username:password@]host1[:port1][,host2[:port2],...[,hostN[:portN]]][/[database][?options]]

Example:

mongodb://db1.example.net:27017/mydatabase
mongodb://db1.example.net:27017,db2.example.net:2500/?replicaSet=test
mongodb://db1.example.net:27017,db2.example.net:2500/?replicaSet=test&connectTimeoutMS=300000

More information on URI format can be found in https://docs.mongodb.com/manual/reference/connection-string/

Configuring the Mongo ext_pillar#

The Mongo ext_pillar takes advantage of the fact that the Salt Master configuration file is yaml. It uses a sub-dictionary of values to adjust specific features of the pillar. This is the explicit single-line dictionary notation for yaml. One may be able to get the easier-to-read multi-line dict to work correctly with some experimentation.

ext_pillar:
  - mongo: {collection: vm, id_field: name, re_pattern: \.example\.com, fields: [customer_id, software, apache_vhosts]}

In the example above, we've decided to use the vm collection in the database to store the data. Minion ids are stored in the name field on documents in that collection. And, since minion ids are FQDNs in most cases, we'll need to trim the domain name in order to find the minion by hostname in the collection. When we find a minion, return only the customer_id, software, and apache_vhosts fields, as that will contain the data we want for a given node. They will be available directly inside the pillar dict in your SLS templates.

Module Documentation#

salt.pillar.mongo.ext_pillar(minion_id, pillar, collection='pillar', id_field='_id', re_pattern=None, re_replace='', fields=None)#

Connect to a mongo database and read per-node pillar information.

Parameters:

  • collection (*) -- The mongodb collection to read data from. Defaults to 'pillar'.

  • id_field (*) -- The field in the collection that represents an individual minion id. Defaults to '_id'.

  • re_pattern (*) -- If your naming convention in the collection is shorter than the minion id, you can use this to trim the name. re_pattern will be used to match the name, and re_replace will be used to replace it. Backrefs are supported as they are in the Python standard library. If None, no mangling of the name will be performed - the collection will be searched with the entire minion id. Defaults to None.

  • re_replace (*) -- Use as the replacement value in node ids matched with re_pattern. Defaults to ''. Feel free to use backreferences here.

  • fields (*) -- The specific fields in the document to use for the pillar data. If None, will use the entire document. If using the entire document, the _id field will be converted to string. Be careful with other fields in the document as they must be string serializable. Defaults to None.