IMPALA-8341 introduces a data cache for remote reads. In particular, it caches data for non-local reads (e.g. S3, ABFS, ADLS). The data cache can be enabled setting the startup flag --data_cache=<dir1>,<dir2>,...,<dirN>:<quota> in which <dir1>,...,<dirN> are directories on local filesystem and quota is the storage consumption quota for each directory. Note that multiple Impala daemons running on the same host must not share cache directories.