Learn how the recent adjustments in Impala have impacted its behavior in this version of
Cloudera Runtime compared to earlier versions.
- Summary:
- Buffering Differences in Impala-Shell with
Python 3 in DataHub
- Previous behavior:
- The Impala-shell relied on Python 2
installed on CentOS 7. Python 2 performed more extensive
buffering of output, which typically had no noticeable
impact. However, in scenarios where Impala-shell was run
without consuming stdout or stderr, the buffering allowed
the command to complete successfully as long as the output
fit within the buffer.
- New behavior:
- Starting with CDP DataHub 7.2.17, Red Hat
8 with Python 3 is now the default. Python 3 buffers less
input than Python 2, which can lead to issues when running
impala-shell without consuming stdout or stderr. If the
buffer fills and is not consumed, the command may hang,
preventing it from completing.
