I installed Drill(1.16 version) on 3 nodes and use distributed mode. Becuz when i implemented query(join multiple table), I want to distribute the memory usage.
But When I enter Drill Web UI, Only a drillbit that I access gives information about usage. other drillbit says Not Available
.
I wonder if distributed mode is working well or not.
Just for test, When I implement a query which joins 3 tables on different node(PostgreSQL), only single major fragment is created and performed.(that means single drillbit is working and does not distribute the memory usage)
This is Web UI. you can see the Not available
sign.
This is conf/drill-override.conf
drill.exec: {
cluster-id: "drillbits1",
zk.connect: "<ip1>:2181,<ip2>:2181,<ip3>:2181"
}
If I misunderstand the Drill structure, Please explain me. Thanks.
This issue was fixed in the scope of https://issues.apache.org/jira/browse/DRILL-7582 and https://issues.apache.org/jira/browse/DRILL-7619, so on the current master version, these metrics should be displayed.
As a workaround for older versions, you may update your /etc/hosts
file to map nodes IPs with their hostnames.