webMethods API Gateway Q&A - kibana -illegal_argument_exception] mapper [version] cannot be changed from type [long] to [integer]

Product/components used and version/fix level:

Version: 10.3 and above
Operating System Family: PC
Operating System: Windows Server 2016

Detailed explanation of the problem:

Unable to start Kibana after switching elasticsearch from the local installation to remotes installations.

There are two API Gateway servers clustered using TSA, three Internal Data Store servers clustered on remote separate hosts.

The local instance of Internal Data Store (elasticsearche) on the API Gateway servers is stopped with auto start set to false.

Error messages / full error message screenshot / log file:

The following Kibana startup error is seen:
"illegal_argument_exception] mapper [version] cannot be changed from type [long] to [integer]"

This suggests that the Kibana index in the external elasticsearch is not compatible.

kibana.yml updated -

The Elasticsearch instance to use for all your queries.

elasticsearch.url: “http://ESB10ESINUAT02:9240

How can the index be rebuilt?

Should the gateway_default_dashboard* indexes be deleted?

Example -
curl -X DELETE http://localhost:9240/gateway_default_dashboard*

The indexes can be rebuilt by deleting the gateway_default_dashboard* indexes.

Listing the elasticsearch indexes-
http://<ES-HOSTNAME>:9240/_cat/indices

The Kibana being used is the one bundled along with API Gateway and not an external one.

Assuming this is the case, please follow the below steps:

  1. Stop both the API Gateway instances
  2. Clean the dashboard indices from the external Elasticsearch cluster.

Example using the command line utility curl -
curl -X DELETE http://<ES-HOSTNAME>:9240/gateway_default_dashboard*

Postman could also be used.

  1. Start the API Gateway instances.

This will resolve the issue.

Alternate:

The indexes can be reconstructed by removing the gateway_default_dashboard* indexes.

Displaying the Elasticsearch indexes: http://<ES-HOSTNAME>:9240/_cat/indices

The Kibana being utilized is the one bundled together with API Gateway, rather than an external one.

Assuming this scenario, kindly follow the below steps:

Cease both the API Gateway instances.

Erase the dashboard indices from the external Elasticsearch cluster.

An example using the command line utility curl:

curl -X DELETE http://<ES-HOSTNAME>:9240/gateway_default_dashboard*

Alternatively, Postman can also be utilized.

Start the API Gateway instances.

This action will resolve the issue.

This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.