[DRILL-3829] Metadata Caching : Drill should ignore a corrupted cache file

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.12.0
• Component/s: Metadata
• Assignee: Vitalii Diravka
• Reporter: Rahul Kumar Challapalli
• Created: 23/Sep/15 18:15
• Updated: 16/Aug/17 11:06
• Resolved: 16/Aug/17 11:06

Description

git.commit.id.abbrev=3c89b30

Drill should validate the cache file structure and ignore it if it detects any corruption to its contents.

I placed an empty cache file in the directory and executed a count query on top of the directory. Below is what I got

select count(*) from dfs.`/drill/testdata/metadata_caching/lineitem`;
Error: SYSTEM ERROR: JsonMappingException: No content to map due to end-of-input
 at [Source: com.mapr.fs.MapRFsDataInputStream@293240cd; line: 1, column: 1]


[Error Id: 88f77d37-aff3-4adc-bb0e-6c13b49e7776 on qa-node190.qa.lab:31010] (state=,code=0)

At the very least we should inform that the cache file has been corrupted.

Issue Links

• duplicates: DRILL-3827 Empty metadata file causes queries on the table to fail
• Is contained by: DRILL-5660 Drill 1.10 queries fail due to Parquet Metadata "corruption" from DRILL-3867