Can't set xpack.ml.enabled=false when running Hardened Elasticsearch image

[greenwoodma]

I'm trying to start up dhi.io/elasticsearch:9 (i.e. a hardened image for elasticsearch) and while it initially starts to bring up the instance it eventually crashes with the error:

Failure running machine learning native code. This could be due to running on an unsupported OS or distribution, missing OS libraries, or a problem with the temp directory. To bypass this problem by running Elasticsearch without machine learning functionality set [xpack.ml.enabled: false].

Now I've had that before with the normal docker images available direct from elastic, and in fact have already disabled the feature in my compose file in the environment section:

    environment:
      - xpack.ml.enabled=false

That works fine with the official images but seems to have no effect on the hardened images (I've tried it on a number of 9.x versions to see if it was just the latest that was a problem).

I don't know if it's related but in the logs there are a number of messages such as

INFO org.elasticsearch.common.util.FeatureFlag - The current build is a snapshot, feature flag [new_ml_memory_computation] is enabled

and I'm wondering if the snapshot building is causing the environment settings to be ignored. If so then how would we go about configuring the image properly?

[pxxx6094-del]

#> I'm trying to start up dhi.io/elasticsearch:9 (i.e. a hardened image for elasticsearch) and while it initially starts to bring up the instance it eventually crashes with the error:

Failure running machine learning native code. This could be due to running on an unsupported OS or distribution, missing OS libraries, or a problem with the temp directory. To bypass this problem by running Elasticsearch without machine learning functionality set [xpack.ml.enabled: false].

Now I've had that before with the normal docker images available direct from elastic, and in fact have already disabled the feature in my compose file in the environment section:

    environment:
      - xpack.ml.enabled=false

That works fine with the official images but seems to have no effect on the hardened images (I've tried it on a number of 9.x versions to see if it was just the latest that was a problem).

I don't know if it's related but in the logs there are a number of messages such as

INFO org.elasticsearch.common.util.FeatureFlag - The current build is a snapshot, feature flag [new_ml_memory_computation] is enabled

and I'm wondering if the snapshot building is causing the environment settings to be ignored. If so then how would we go about configuring the image properly?

[pxxx6094-del]

• [ ]

[pxxx6094-del]

https://github.com/orgs/docker-hardened-images/discussions/218#discussion-9533165

[pxxx6094-del]