Daniel Becker has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/17026 )

Change subject: IMPALA-10640: Support reading Parquet Bloom filters - most 
common types
......................................................................


Patch Set 27:

The problem seems to be with this test: 
tests/failure/test_failpoints.py::TestFailpoints::test_failpoints, and this 
query in particular:

SET mt_dop=4;
SET batch_size=0;
SET num_nodes=0;
SET disable_codegen_rows_threshold=0;
SET disable_codegen=False;
SET abort_on_error=1;
SET exec_single_node_rows_threshold=0;
SET 
debug_action=0:PREPARE_SCANNER:MEM_LIMIT_EXCEEDED|COORD_BEFORE_EXEC_RPC:JITTER@100@0.3;
SELECT STRAIGHT_JOIN *
FROM functional_kudu.alltypes t1
JOIN /*+broadcast*/ functional_kudu.alltypesagg t2 ON t1.id = t2.id
WHERE t2.int_col < 1000;

It seems that when reading a Kudu table, the code for kudu::BlockBloomFilter 
may come from different places
- from the Kudu (binary?) dependency and
- be/src/kudu/util/block_bloom_filter.h

The crash happens because an assert in Kudu as 
kudu::BlockBloomFilter::hash_algorithm_ has an invalid value.

I think it is caused by binary incompatibility between our modified 
kudu::BlockBloomFilter and the original one, used in Kudu. I have added two 
fields just before kudu::BlockBloomFilter::hash_algorithm_ and probably that 
data is treated as if it were the hash_algorithm_ field by the original Kudu 
code. When I put the new fields to the end of the class, the crash doesn't 
happen.


--
To view, visit http://gerrit.cloudera.org:8080/17026
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I7119c7161fa3658e561fc1265430cb90079d8287
Gerrit-Change-Number: 17026
Gerrit-PatchSet: 27
Gerrit-Owner: Daniel Becker <daniel.bec...@cloudera.com>
Gerrit-Reviewer: Csaba Ringhofer <csringho...@cloudera.com>
Gerrit-Reviewer: Daniel Becker <daniel.bec...@cloudera.com>
Gerrit-Reviewer: Impala Public Jenkins <impala-public-jenk...@cloudera.com>
Gerrit-Reviewer: Tamas Mate <tm...@cloudera.com>
Gerrit-Reviewer: Zoltan Borok-Nagy <borokna...@cloudera.com>
Gerrit-Comment-Date: Wed, 21 Apr 2021 15:50:44 +0000
Gerrit-HasComments: No

Reply via email to