Curator¶

From https://www.elastic.co/guide/en/elasticsearch/client/curator/current/about.html#about:

Elasticsearch Curator helps you curate, or manage, your Elasticsearch indices and snapshots by:

Obtaining the full list of indices (or snapshots) from the cluster, as the actionable list

Iterate through a list of user-defined filters to progressively remove indices (or snapshots) from this actionable list as needed.

Perform various actions on the items which remain in the actionable list.

Configuration¶

Curator actions are stored in /opt/so/conf/curator/action/. These actions are run by cron jobs managed by Salt.

Curator defaults to closing indices older than 30 days. To modify this, edit /opt/so/saltstack/local/pillar/global.sls and change the close setting for each index under the elasticsearch:index_settings section.

Note

/opt/so/saltstack/local/pillar/global.sls only lists default indices. If you add any new indices, you will need to manually add them to that file.

Curator also deletes old indices to prevent your disk from filling up. The mechanism for this depends on how Elasticsearch is configured, so you can learn more in the Elasticsearch section.

For more information about the Curator close and delete settings, please see the Elasticsearch section.

Creating Actions¶

If you would like to create a custom Curator action, you will need to create a Curator action file and corresponding script file and then update Curator’s state file.

Curator action file¶

You can add the action file to /opt/so/saltstack/local/salt/curator/files/action/. Make sure the file ownership is curator:socore. The file will automatically get copied into /opt/so/conf/curator/action/.

Script file¶

The script file is what actually executes Curator and specifies the action file. This script file must be placed in /opt/so/saltstack/local/salt/curator/files/bin/. See /opt/so/saltstack/default/salt/curator/files/bin/ for examples of script files and copy one over to modify if needed.

State file¶

Next, Curator’s state file (init.sls) must be modified. This will be located at /opt/so/saltstack/local/salt/curator/ and will copy files and create the cron job.

If /opt/so/saltstack/local/salt/curator/init.sls does not already exist, you can copy /opt/so/saltstack/default/salt/curator/init.sls to /opt/so/saltstack/local/salt/curator/init.sls and modify as shown below.

If /opt/so/saltstack/local/salt/curator/init.sls does already exist, create a backup of the file by copying it to a safe directory. Then, copy the default file located at /opt/so/saltstack/default/salt/curator/init.sls to the location of the current file and run a diff against the two init.sls files. Inside this file that was just copied over, the “cur” and “cron” sections must be added for your Curator job along with anything included in the diff output. Do not edit the original file in the directory.

To add the new Curator job, copy and modify one of the existing sections or use these examples:

For the “cur” section:

cur<custom-name>:
  file.managed:
    - name: /usr/sbin/so-curator-<custom-name>
    - source: salt://curator/files/bin/<your_script_file_name>
    - user: 934
    - group: 939
    - mode: 755

For the “cron” section:

so-curator<custom-name>:
 cron.present:
   - name: /usr/sbin/so-curator-<custom-name> > /opt/so/log/curator/<your_logfile>.log 2>&1
   - user: root
   - minute: '*'
   - hour: '*'
   - daymonth: '*'
   - month: '*'
   - dayweek: '*'

This particular cron section will run the task every minute. After this, restart Curator with sudo so-curator-restart and note any errors (changes are not errors).

To confirm that the job was added correctly, run crontab -l and look for the new task’s cron job.

If the task’s cron job does not show up, then there may have been errors during the restart process. You must fix those errors for the cron job to be created.

Logs¶

When Curator completes an action, it logs its activity in a log file found in /opt/so/log/curator/.

Curator vs Index Lifecycle Management (ILM)¶

The goal of Security Onion is to allow you to concentrate on finding evil rather than spending time managing infrastructure. The default mode that Security Onion deploys allows each node to be independent and removes the complexity of shard migration across multiple nodes. These nodes will use Curator to trim indices as needed ensuring that they never run out of disk space. This is especially important when running in standalone mode. Finally, it should also be noted that ILM does not support deletion based on disk space.