[jira] [Created] (IMPALA-15559) Add support for running bin/single_node_perf_run.py against data in S3

["Joe McDonnell (Jira)" <ji...@apache.org>]

Joe McDonnell created IMPALA-15559:
--------------------------------------

             Summary: Add support for running bin/single_node_perf_run.py against data in S3
                 Key: IMPALA-15559
                 URL: https://issues.apache.org/jira/browse/IMPALA-15559
             Project: IMPALA
          Issue Type: Task
          Components: Infrastructure
    Affects Versions: Impala 5.0.0
            Reporter: Joe McDonnell


To work on larger scale datasets, we want to avoid dataload time. One way to do that is to have a dataset in S3 and run against it there. There are a series of things needed to make that work:
 * bin/single_node_perf_run.py needs to be able to run a custom dataload script (to create the tables to point to the S3 bucket)
 * We want to run with a data cache, so it would be useful for bin/single_node_perf_run.py to have a way to pass arguments directly to bin/start-impala-cluster.py
 * bin/start-impala-cluster.py should support scratch directories and multiple data cache directories (large cloud machines often have multiple SSD devices that can be mounted in different locations)
 * Fixes for running with TARGET_FILESYSTEM=s3



--
This message was sent by Atlassian Jira
(v8.20.10#820010)