[BUG] Shrink and delete policy with single shard index deletes all data

[lockwooddev]

What is the bug?

A policy with a shrink and consecutive delete on a single sharded index will not shrink, but immediately run the delete action an therefore remove all original data without warning.

How can one reproduce the bug?

{
    "policy": {
        "policy_id": "my-policy",
        "description": "",
        "default_state": "rollover",
        "states": [
            {
                "name": "rollover",
                "actions": [
                    {
                        "rollover": {
                            "min_size": "20gb",
                            "min_index_age": "1d"
                        }
                    }
                ],
                "transitions": [
                    {
                        "state_name": "merge"
                    }
                ]
            },
            {
                "name": "merge",
                "actions": [
                    {
                        "shrink": {
                            "num_new_shards": 1
                        }
                    }
                ],
                "transitions": [
                    {
                        "state_name": "delete"
                    }
                ]
            },
            {
                "name": "delete",
                "actions": [
                    {
                        "delete": {}
                    }
                ]
            }
        ],
        "ism_template": [
            {
                "index_patterns": [
                    "my-index*"
                ],
                "priority": 100
            }
        ]
    }
}

What is the expected behavior?

I would expect it to block the delete as this combination would cause data loss

What is your host/environment?

OpenSearch 2.9.0

[bharath-techie]

[ Triage attendees - 1 2 3 4]

Would you like to open a PR for the same issue to handle the race condition ?