nickSoul created HIVE-19950:
-------------------------------
Summary: Hive ACID NOT LOCK LockComponent Correctly
Key: HIVE-19950
URL: https://issues.apache.org/jira/browse/HIVE-19950
Project: Hive
Issue Type: Bug
Affects Versions: 2.3.2
Reporter: nickSoul
Hi,
When using Streaming Mutation recently, I found LockComponents were not locked
correctly by current transaction. I wonder if i use in a wrong way or
misunderstand about ACID in hive. Below is my test case:
# Begin the transaction with a transactionId 126, and the transaction locks a
table. Then hangs the transaction. The lock information were correctly restored
in mariaDB
{code:java}
MariaDB [hive]> select
HL_LOCK_EXT_ID,HL_LOCK_INT_ID,HL_TXNID,HL_DB,HL_TABLE,HL_PARTITION,HL_LOCK_STATE,HL_LOCK_TYPE,HL_ACQUIRED_AT,HL_BLOCKEDBY_EXT_ID,HL_BLOCKEDBY_INT_ID
from HIVE_LOCKS;
+----------------+----------------+----------+-----------+-----------+----
| HL_LOCK_EXT_ID | HL_LOCK_INT_ID | HL_TXNID | HL_DB | HL_TABLE | HL_PARTITION
| HL_LOCK_STATE | HL_LOCK_TYPE | HL_ACQUIRED_AT | HL_BLOCKEDBY_EXT_ID |
HL_BLOCKEDBY_INT_ID |
+----------------+----------------+----------+-----------+-----------+----
| 384 | 1 | 126 | test_acid | acid_test | NULL | a | w | 1529512857000 | NULL |
NULL |
+----------------+----------------+----------+-----------+-----------+----{code}
# Begin the other transaction with a transactionId 127 before previous
transaction 126 finished. Transaction 127 tries to lock the same table too, but
failed at first attempt.
{code:java}
//Log Information{code}
#
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
