Calling options() on a client resets timeout related attributes

[redbaron4]

Elasticsearch version (8.1.1):

elasticsearch-py version (8.1.1):

Please make sure the major version matches the Elasticsearch server you are running.

Description of the problem including expected versus actual behavior:

When the options() method is called for an existing ElasticSearch client instance, the timeout related parameters are not preserved. These parameters are:

• self._request_timeout
• self._max_retries
• self._retry_on_timeout
• self._retry_on_status

This is problematic in the helpers module, particularly scan() because that method calls the options and resets any timeout related settings we may have made on the client. We can re-specify the request_timeout there but retries parameters cannot be replicated

Steps to reproduce:

>>> import elasticsearch as ES
>>> es=ES.Elasticsearch("http://esmasters:9200", sniff_on_connection_fail=True, sniff_on_start=True, min_delay_between_sniffing=600, request_timeout=600, sniff_timeout=300, max_retries=5, retry_on_timeout=True)
>>> es._retry_on_timeout
True
>>> es2=es.options()
>>> es2._retry_on_timeout
<DEFAULT>
>>> es._max_retries
5
>>> es2._max_retries
<DEFAULT>
>>> es2._retry_on_status
<DEFAULT>
>>> es._retry_on_status
<DEFAULT>
>>> es._request_timeout
600
>>> es2._request_timeout
<DEFAULT>

[sethmlarson]

Thanks for reporting this, this does seem like something worth fixing. Let me take a look.

[mwbenowitz]

+1 to this. It's called in the streaming_bulk method as well, and I believe it is not possible to re-specify any of the affected parameters at that point (most importantly request_timeout)

[meganvw]

Yes +1 it would be great to see this fixed - ran into this problem today

[itayB]

Assuming you guys used the retry_on_timeout in bulk operations, I wonder if you also faced this issue (value propagate but retries never happen)