Cloud native analytics grouping stops on a geo-redundant cluster

If live events are not being grouped by cloud native analytics policies, restart the gateway pod.

Problem

Live events are not being grouped by cloud native analytics policies if the connection for the cloud native analytics gateway is dropped. You can tell if the gateway is dropped if the row for the cloud native analytics gateway is missing in the catalog.connections table.

To check the catalog.connections table, run the following command:

oc rsh $(oc get pods|grep ncoprimary|head -1|awk '{print $1}') 
/opt/IBM/tivoli/netcool/omnibus/bin/nco_sql -user root -password $OMNIBUS_ROOT_PWD -server AGG_P
select * from catalog.connections where AppDescription  like 'CEA'
go

If the cloud native analytics gateway is connected, the command returns an entry similar to the following example:

ConnectionID LogName HostName                       AppName AppDescription     IsRealTime ConnectTime
 ------------ ------- ------------------------------ ------- ------------------ ---------- -----------
           28 gateway evtmanager-ea-noi-layer-eanoig GATEWAY CEA Gateway Reader 1          1661753131

Resolution

If the gateway is dropped, restart the gateway pod with the following command:

oc delete pod $(oc get pod|grep eanoigateway|awk '{print $1}' )

If the issue remains after you run the command, see Cloud native analytics data flow stops.