When `xpack.ml.enabled` is set to false, entering Kibana's machine learning page has a error in the log

[linghengqian]

• I am doing some tests in docker-compose.yml, I noticed that when I set the Elasticsearch environment variable to contain XPACK_ML_ENABLED: false, the machine learning page of the Kibana instance will jump to some urls like crazy, and the Kibana page also has the corresponding log to Show errors.

[2022-03-31T12:47:10.150+08:00][INFO ][plugins.security.routes] Logging in with provider "basic" (basic)
[2022-03-31T12:47:10.946+08:00][INFO ][status] Kibana is now available (was degraded)
[2022-03-31T12:55:26.827+08:00][INFO ][plugins.fleet] Beginning fleet setup
[2022-03-31T12:55:26.889+08:00][INFO ][plugins.fleet] Fleet setup completed
[Error [UnsubscriptionError]: 1 errors occurred during unsubscription:
1) ReferenceError: Cannot access 'sub' before initialization] {
    ReferenceError: Cannot access 'sub' before initialization
  errors: [
        at Subscription.<anonymous> (/usr/share/kibana/src/plugins/bfetch/server/streaming/create_compressed_stream.js:52:5)
        at Subscription.unsubscribe (/usr/share/kibana/node_modules/rxjs/internal/Subscription.js:40:30)
        at FinallySubscriber.Subscription.unsubscribe (/usr/share/kibana/node_modules/rxjs/internal/Subscription.js:53:29)
        at Subscriber.unsubscribe (/usr/share/kibana/node_modules/rxjs/internal/Subscriber.js:86:38)
        at FinallySubscriber.Subscriber.unsubscribe (/usr/share/kibana/node_modules/rxjs/internal/Subscriber.js:86:38)
        at Subscriber.Subscription.unsubscribe (/usr/share/kibana/node_modules/rxjs/internal/Subscription.js:53:29)
        at SafeSubscriber._unsubscribe (/usr/share/kibana/node_modules/rxjs/internal/Subscriber.js:241:27)
        at SafeSubscriber.Subscriber.unsubscribe (/usr/share/kibana/node_modules/rxjs/internal/Subscriber.js:86:38)
        at SafeSubscriber.Subscription.unsubscribe (/usr/share/kibana/node_modules/rxjs/internal/Subscription.js:40:30)
[2022-03-31T12:47:07.897+08:00][INFO ][plugins.screenshotting.chromium] Browser executable: /usr/share/kibana/x-pack/plugins/screenshotting/chromium/headless_shell-linux_x64/headless_shell
        at Subscriber.complete (/usr/share/kibana/node_modules/rxjs/internal/Subscriber.js:78:18)
        at SafeSubscriber.complete (/usr/share/kibana/node_modules/rxjs/internal/Subscriber.js:191:26)
        at Subscriber._complete (/usr/share/kibana/node_modules/rxjs/internal/Subscriber.js:96:26)
        at FinallySubscriber.Subscriber._complete (/usr/share/kibana/node_modules/rxjs/internal/Subscriber.js:96:26)
        at FinallySubscriber.Subscriber.complete (/usr/share/kibana/node_modules/rxjs/internal/Subscriber.js:78:18)
        at CatchSubscriber.Subscriber.complete (/usr/share/kibana/node_modules/rxjs/internal/Subscriber.js:78:18)
        at CatchSubscriber.Subscriber._complete (/usr/share/kibana/node_modules/rxjs/internal/Subscriber.js:96:26)
        at MergeMapSubscriber._complete (/usr/share/kibana/node_modules/rxjs/internal/operators/mergeMap.js:88:30)
        at processTicksAndRejections (node:internal/process/task_queues:96:5)
        at Observable.subscribe (/usr/share/kibana/node_modules/rxjs/internal/Observable.js:30:22)
        at Observable._subscribe (/usr/share/kibana/node_modules/rxjs/internal/util/subscribeToArray.js:7:16)
        at MergeMapSubscriber.Subscriber.complete (/usr/share/kibana/node_modules/rxjs/internal/Subscriber.js:78:18)
        at Observable._trySubscribe (/usr/share/kibana/node_modules/rxjs/internal/Observable.js:44:25)
        at Observable.subscribe (/usr/share/kibana/node_modules/rxjs/internal/Observable.js:25:31)
        at MergeMapOperator.call (/usr/share/kibana/node_modules/rxjs/internal/operators/mergeMap.js:37:23)
        at CatchOperator.call (/usr/share/kibana/node_modules/rxjs/internal/operators/catchError.js:30:23)
        at FinallyOperator.call (/usr/share/kibana/node_modules/rxjs/internal/operators/finalize.js:27:23)
        at createStream (/usr/share/kibana/src/plugins/bfetch/server/streaming/create_stream.js:20:74)
        at createCompressedStream (/usr/share/kibana/src/plugins/bfetch/server/streaming/create_compressed_stream.js:53:7)
        at /usr/share/kibana/src/plugins/bfetch/server/plugin.js:58:45
        at /usr/share/kibana/src/core/server/context/container/context.js:62:16
        at runMicrotasks (<anonymous>)
        at Router.handle (/usr/share/kibana/src/core/server/http/router/router.js:163:30)
        at exports.Manager.execute (/usr/share/kibana/node_modules/@hapi/hapi/lib/toolkit.js:60:28)
        at exports.execute (/usr/share/kibana/node_modules/@hapi/hapi/lib/handler.js:31:20)
  ]
}
        at Request._execute (/usr/share/kibana/node_modules/@hapi/hapi/lib/request.js:281:9)
        at Request._lifecycle (/usr/share/kibana/node_modules/@hapi/hapi/lib/request.js:371:32)
        at Object.internals.handler (/usr/share/kibana/node_modules/@hapi/hapi/lib/handler.js:46:20)
        at handler (/usr/share/kibana/src/core/server/http/router/router.js:124:50)
[2022-03-31T12:56:57.264+08:00][INFO ][plugins.fleet] Fleet setup completed
[2022-03-31T12:56:56.810+08:00][INFO ][plugins.fleet] Beginning fleet setup

• Also the Elasticsearch instance has a warn, but I'm not sure if it's relevant.

{"@timestamp":"2022-03-31T04:57:50.880Z", "log.level": "WARN",  "data_stream.dataset":"deprecation.elasticsearch","data_stream.namespace":"default","data_stream.type":"logs","elasticsearch.elastic_product_origin":"kibana","elasticsearch.event.category":"api","elasticsearch.http.request.x_opaque_id":"465c0681-cbf7-4ca6-9aab-0c334b9fab85;kibana::management:","event.code":"open_system_index_access","message":"this request accesses system indices: [.async-search, .security-7, .tasks, .transform-internal-007], but in a future major version, direct access to system indices will be prevented by default" , "ecs.version": "1.2.0","service.name":"ES_ECS","event.dataset":"deprecation.elasticsearch","process.thread.name":"elasticsearch[node01.elasticsearch.test.com][transport_worker][T#1]","log.logger":"org.elasticsearch.deprecation.cluster.metadata.IndexNameExpressionResolver","trace.id":"35a2fbeff0bf419a2debf9e82cd6aef7","elasticsearch.cluster.uuid":"-cezmxW7TMKuqZFy_YccFA","elasticsearch.node.id":"9traWmKDSZmC228Ks3JC8A","elasticsearch.node.name":"node01.elasticsearch.test.com","elasticsearch.cluster.name":"test-elasticsearch"}

• I think if Kibana detects that the setting XPACK_ML_ENABLED of the corresponding ElatsicSearch is false, would it be better not to display this machine learning entry directly? But I'm not sure how to hide it, and now I always get an error when entering this page. At the same time the page flickers.
• 
• Below is the docker-compose.yml I use for testing. After startup, I registered Kibana with ElasticSearch through the token generated by bin/elasticsearch-create-enrollment-token -s kibana.

version: "3.8"
services:
  elasticsearch-node01:
    image: elasticsearch:8.1.1
    hostname: node01.elasticsearch.test.com
    environment:
      TZ: Asia/Shanghai
      ES_SETTING_BOOTSTRAP_MEMORY__LOCK: true
      ES_SETTING_CLUSTER_NAME: "test-elasticsearch"
      ELASTIC_PASSWORD: "6123AF57-35CA-B2E0-BE8B-12345D5C123E"
      XPACK_ML_ENABLED: false
      ES_JAVA_OPTS: "-Xms4g -Xmx4g"
    volumes:
      - elasticsearch-data01:/usr/share/elasticsearch/data
    ports:
      - "9200:9200"
      - "9300:9300"
    ulimits:
      memlock:
        soft: -1
        hard: -1
  kibana-node01:
    image: kibana:8.1.1
    hostname: node01.kibana.test.com
    environment:
      TZ: Asia/Shanghai
      XPACK_SECURITY_ENCRYPTIONKEY: fc12345ea223eb1234dac7b5678ac62e
      XPACK_ENCRYPTEDSAVEDOBJECTS_ENCRYPTIONKEY: 1bb12c4e9ed1c36ae4a8ead1234567f1
      XPACK_REPORTING_ENCRYPTIONKEY: af5cbdd2e2f4da20a72f60f1e123450f
      XPACK_REPORTING_ROLES_ENABLED: false
      XPACK_REPORTING_KIBANASERVER_HOSTNAME: localhost
      SERVER_PUBLICBASEURL: http://localhost:5601
      I18N_LOCALE: zh-CN
    ports:
      - "5601:5601"
    volumes:
      - kibana-data01:/usr/share/kibana/data
volumes:
  elasticsearch-data01:
    driver: local
  kibana-data01:
    driver: local

[dolaru (David Olaru)]

Hi @linghengqian.

I think if Kibana detects that the setting XPACK_ML_ENABLED of the corresponding ElatsicSearch is false , would it be better not to display this machine learning entry directly?

You are correct. Kibana should detect that ML is disabled in Elasticsearch and not show the app in the sidebar.

After setting XPACK_ML_ENABLED: false for your Elasticsearch container, did you restart the Kibana container as well?

[linghengqian]

• After restarting Kibana, this situation no longer occurs. It seems that the wrong log is caused by the device on my side. But when I enter the machine learning page, I still see that it is switching urls like crazy, causing Chrome's address bar to flicker. Until after a while it goes to localhost:5601/app/ml/datavisualizer. (Yes, no error or warn log in Kibana )
• 
• 

[dolaru (David Olaru)]

I tried to reproduce the issue with with a docker-compose setup identical to the one mentioned in your first post, but I didn't notice any issues with the URL flickering or any redirect loops when navigating to the Machine Learning page.

Could you please update Chrome to the latest version, clear the cache and let me know if you're still seeing this issue?

Later edit: Is this the flickering you're talking about?

If yes, this is normal.

[linghengqian]

• I restarted a new compose instance locally and it didn't reproduce. When I restarted the old instance with docker compose down --volumes and docker compose up -d to restart, the problem disappeared, it seems to be because of the old instance, the /usr/share/kibana/data of kibana of the instance is damaged.
• thank you very much for your help~