AXA Log Analytics 17.3.0 installation on a 5 node design, with 3 ElasticSearch nodes and a single ElasticSearch client.
There is an issue in the jarvis-es-utils-ForceMerge.log. In the log we get the following message:
This is all the job seems to report so it appears that it's not completing as expected, Checking the _cluster/state command and found that the node mentioned in the log above is for the ElasticSearch Client:
Where does the configuration come from to see this and what can we do about it ?
Release : 17.3
Component : AXA LOG ANALYTICS
ElasticSearch data nodes will have a box_type value of hot or warm, this can be noted with this query :
http://<AXA_HOST>:9200/_cluster/state?pretty
"cluster_name" : "CA_AOPlatform-ProdCluster", "version" : 35493, "state_uuid" : "lFkorqWRTaulnhQXA30x5A", "master_node" : "jacSrq7wTL2LFeX8BOmi9g", "blocks" : { }, "nodes" : { "jacSrq7wTL2LFeX8BOmi9g" : { "name" : "jacSrq7", "ephemeral_id" : "qudYEzTzTcihQAmRLQHHqQ", "transport_address" : "10.80.72.149:9300", "attributes" : { "box_type" : "hot" }
The ElasticSearch client node does not have this box_type value specified as it does not store data, hence the error.
The list of ElasticSearch hosts is configured in <AXA_HOME>jarvis/config/utils.properties but it is considered best to leave this as configured.
Using the log configuration in utils.log4j.properties
For the original messages in ForceMerge, it's possible to change the log level from INFO to ERROR in this config section, that would stop the reporting of any WARN errors and reduce impact of unnecessary errors.
logger.forceMergeLogger2.name=com.ca.jarvis.utils.services.EsClusterService
logger.forceMergeLogger2.level=INFO
logger.forceMergeLogger2.additivity=false
logger.forceMergeLogger2.appenderRefs=forceMergeAppender
logger.forceMergeLogger2.appenderRef.utilsAppender.ref=ForceMergeFile
logger.forceMergeLogger2.appenderRefs=console
logger.forceMergeLogger2.appenderRef.console.ref=STDOUT