[IMPALA-5916] Add truncating files to test_scanners_fuzz.py

• Type: Improvement
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: Impala 2.11.0
• Component/s: Infrastructure
• Epic Color: ghx-label-9
• Assignee: Unassigned
• Reporter: Lars Volker
• Created: 09/Sep/17 17:25
• Updated: 09/Sep/17 17:29

Description

We should add tests for truncated files to test_scanners_fuzz.py. Impala can see truncated files when users overwrite tables (e.g. using Hive) while Impala is scanning them.

The fuzz test already truncates files already, however we need to add support for running a query on an untruncated table first, so that the metadata and the file diverge. See IMPALA-5890 for details.

Issue Links

• relates to: IMPALA-5890 Segmentation fault in ScannerContext::Stream::GetBytesInternal(long, unsigned char**, bool, long*)