Bug when calling Document.init() more than once

Using

elasticsearch==6.3.1
elasticsearch-dsl==6.2.1

Against an Elastic Search cluster: 6.3.2

Using this Document to persist data:

class Data(Document):
    some_field = Text()
    
    class Meta:
        doc_type = 'data_doc_type'
    
    class Index:
        name = 'data'
        doc_type = 'data_doc_type'
        settings = {
            "number_of_shards": 60,
            "number_of_replicas": 1
        }

When i call init Data.init() for the first time it actually works ok and creates the index and mapping correctly within ElasticSearch, but when i need to call it again to apply new mapping to that index i come accross this exception:

File "data.py", line ***, in init
    Data.init()
  File "elasticsearch_dsl/document.py", line 128, in init
    i.save(using=using)
  File "elasticsearch_dsl/index.py", line 260, in save
    self.put_settings(using=using, body=settings)
  File "elasticsearch_dsl/index.py", line 430, in put_settings
    return self._get_connection(using).indices.put_settings(index=self._name, **kwargs)
  File "elasticsearch/client/utils.py", line 76, in _wrapped
    return func(*args, params=params, **kwargs)
  File "elasticsearch/client/indices.py", line 547, in put_settings
    '_settings'), params=params, body=body)
  File "elasticsearch/transport.py", line 318, in perform_request
    status, headers_response, data = connection.perform_request(method, url, params, body, headers=headers, ignore=ignore, timeout=timeout)
  File "elasticsearch/connection/http_requests.py", line 90, in perform_request
    self._raise_error(response.status_code, raw_data)
  File "elasticsearch/connection/base.py", line 125, in _raise_error
    raise HTTP_EXCEPTIONS.get(status_code, TransportError)(status_code, error_message, additional_info)

And this is the actual reponse from elastic:

{
    "error": {
        "root_cause": [
            {
                "type": "remote_transport_exception",
                "reason": "[1-*****][x.x.x.x:9300][indices:admin/settings/update]"
            }
        ],
        "type": "illegal_argument_exception",
        "reason": "Can't update non dynamic settings [[index.number_of_shards]] for open indices [[data/***]]"
    },
    "status": 400
}

The problem seems to be that you’re trying to apply the settings more than once to settings that are not writable.

The problem is that in now i don’t know how to just update the mapping on that index.

[hoIIer]

@HonzaKral I just ran into this exact issue on latest when I mistakenly thought Post.init() was idempotent. Is this supposed to be fixed?

Is it safe to do this instead?

if not index := Index('posts').exists():
    Post.init()

I received error:

RequestError(400, 'illegal_argument_exception', "Can't update non dynamic settings [[index.number_of_shards]] for open indices [[posts-staging/sOmqF5m3RymtcVj2-2vflA]]")

elasticsearch==7.6.0
elasticsearch-dsl==7.1.0

[honzakral]

I provided a hotfix in a991fae, I will add some more tests around the functionality and, while this is not a guaranteed solution for all settings, for your use case this should now be fixed with the next release that should be coming soon.