Prasanth Jayachandran created HIVE-24866:
--------------------------------------------
Summary: FileNotFoundException during alter table concat
Key: HIVE-24866
URL: https://issues.apache.org/jira/browse/HIVE-24866
Project: Hive
Issue Type: Bug
Affects Versions: 2.4.0, 3.2.0, 4.0.0
Reporter: Prasanth Jayachandran
Assignee: Prasanth Jayachandran
Because of the way combinefile IF groups files based on node and rack locality,
there are cases where single big orc file gets spread across 2 or more combine
hive split. When first task completes, as part of jobCloseOp the source orc
file of concatenation is moved/renamed which can lead to FileNotFoundException
in subsequent mappers that has partial split of that file.
A simple fix would be for the mapper with start of the split to own the entire
orc file for concatenation. If a mapper gets partial split which is not the
start then it can skip the entire file.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
