Hi everyone,
I’m trying to connect KNIME Analytics Platform to Hive via JDBC using Docker and wanted to ask how others in the community are doing this successfully.
So far I’ve noticed:
-
Some Apache Hive Docker images start but don’t automatically expose HiveServer2 (port 10000).
-
Images like
bde2020/hiveseem to depend on HDFS, so HiveServer2 never starts without a Namenode. -
On Apple Silicon (ARM Macs), amd64 Hive images sometimes run but TCP port 10000 never becomes reachable, even when HiveServer2 appears to start.
From the KNIME side, the setup is straightforward (DB Connector + Hive JDBC driver), but the challenge seems to be finding a reliable Hive Docker image/setup.
Observed Issues
-
Apache Hive images (
apache/hive:3.x / 4.x)-
Containers often exit immediately or require manual HiveServer2 startup
-
HiveServer2 starts, but port 10000 is not reachable externally
-
-
bde2020 Hive images
-
Hive startup fails when HDFS is not running
-
Logs show connection failures to
fs.defaultFS(port 8020) -
HiveServer2 never becomes available
-
-
Docker checks
-
nc -vz localhost 10000consistently shows TCP connection refused -
Confirms HiveServer2 is not listening despite containers being up
-
-
Apple Silicon / ARM Macs
-
amd64 Hive images run under emulation
-
Hive processes start, but JDBC TCP ports do not bind reliably
-
Key Questions
-
Which Docker image/setup are you using for Hive with KNIME?
-
Are you running a full Hadoop stack, or a standalone HiveServer2?
-
Has anyone done this reliably on Apple Silicon?
Would love to hear what’s worked for others — or if there’s a recommended approach.
Thanks!