[ILM] Add "wait for snapshot" policy field to Delete phase

[yuliacech]

Summary

This PR adds a text field for "wait for snapshot policy" in Delete phase of an index lifecycle policy.

How to test

1. Create a repository for snapshots
2. Create a snapshot policy that backs up an index
3. Create an index lifecycle policy that deletes the managed index with "wait for snapshot" option
4. Create an index template with an associated lifecycle policy and a write index for rollover
5. Add 1 document to the index for rollover to happen (for testing we are setting max_docs: 1).
6. After conditions for Delete phase are met (for testing Delete phase immediately follows rollover) and snapshot policy is executed (for testing run the snapshot policy manually), the managed index is deleted.

Expand for console commands

(when running yarn es you need to add -E path.repo=./my_repo_test to be able to create a repository)

PUT /_snapshot/my_repo
{
  "type": "fs",
  "settings": {
    "location": "./my_repo_test",
    "compress": true
  }
}

(index to be backed up test-*, snapshots created once a year and deleted after 10 min)

PUT _slm/policy/my_snapshot_policy
{
  "name": "<snapshot-{now}>",
  "schedule": "0 0 0 1 1 ?",
  "repository": "my_repo",
  "config": {
    "indices": "test-*"
  },
  "retention": {
    "expire_after": "10m",
    "min_count": 1,
    "max_count": 3
  }
}

(rollover after 1 doc, delete after my_snapshot_policy created a snapshot).

PUT _ilm/policy/my_policy
{
  "policy": {
    "phases": {
      "hot": {
        "min_age": "0ms",
        "actions": {
          "rollover": {
            "max_size": "50gb",
            "max_age": "30d",
            "max_docs": 1
          },
          "set_priority": {
            "priority": 100
          }
        }
      },
      "delete": {
        "min_age": "0d",
        "actions": {
          "wait_for_snapshot": {
            "policy": "my_snapshot_policy"
          },
          "delete": {}
        }
      }
    }
  }
}

(update lifecycle poll interval)

PUT /_cluster/settings
{
  "transient": {
    "logger.org.elasticsearch.xpack.core.indexlifecycle": "TRACE",
    "logger.org.elasticsearch.xpack.indexlifecycle": "TRACE",
    "logger.org.elasticsearch.xpack.core.ilm": "TRACE",
    "logger.org.elasticsearch.xpack.ilm": "TRACE",
    "indices.lifecycle.poll_interval": "10s"
  }
}

PUT _template/test
{
  "index_patterns": [
    "test-*"
  ],
  "settings": {
    "number_of_shards": 3,
    "number_of_replicas": 0,
    "index.lifecycle.name": "my_policy",
    "index.lifecycle.rollover_alias": "test-alias"
  }
}
PUT /test-000001
{
  "aliases": {
    "test-alias": {
      "is_write_index": true
    }
  }
}

PUT test-alias/_doc/1
{
  "a": "a"
}

Release Note

We added a snapshot policy name field to Delete phase of index lifecycle policy. This option ensures that the snapshot policy is executed before the managed index is deleted.

[elasticmachine]

Pinging @elastic/es-ui (Team:Elasticsearch UI)

[yuliacech]

As suggested by Alison, I'll be adding some tests to verify payload data.

[cjcenizal]

Tested locally, code LGTM! Great work Yulia. I'll review again after you've added tests.

I tested this by creating an ILM policy via the UI that resulted in this configuration:

{
  "policy": {
    "phases": {
      "hot": {
        "min_age": "0ms",
        "actions": {
          "rollover": {
            "max_age": "100s",
            "max_size": "50gb",
            "max_docs": 1
          },
          "set_priority": {
            "priority": 100
          }
        }
      },
      "delete": {
        "min_age": "0d",
        "actions": {
          "wait_for_snapshot": {
            "policy": "test"
          },
          "delete": {
            "delete_searchable_snapshot": true
          }
        }
      }
    }
  }
}

Then I created a policy that would snapshot everything and set it to execute in a couple of minutes (a little tricky because it shows timing according to GMT instead of my local timezone).

Then I followed the other quick-test steps from the ILM README to associate my policy with an index.

In a couple minutes, the snapshot was taken and then index was deleted as expected.

[yuliacech]

@elasticmachine merge upstream

[yuliacech]

Hey team,
I added some jest integration tests: for the scope of this PR, tests are only for Delete phase. But the setup for jest client integration is now in place to add more tests later.
Please let me know what you think!

[cjcenizal]

Tests LGTM! Thanks for adding those!

[cjcenizal]

The failing tests look unrelated to these changes. I don't see master failing CI so I'd try merging in the last version of master to see if a fix got merged.

[yuliacech]

@elasticmachine merge upstream

[kibanamachine]

💚 Build Succeeded

• continuous-integration/kibana-ci/pull-request
• Commit: fc255ec

History

• 💔 Build #53092 failed b0bcf85
• 💚 Build #52371 succeeded dbdf8e9

To update your PR or re-run it, just comment with:
@elasticmachine merge upstream