You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@trafodion.apache.org by "Alice Chen (JIRA)" <ji...@apache.org> on 2015/07/22 20:15:27 UTC
[jira] [Created] (TRAFODION-277) LP Bug: 1320387 - Update
statistics spawned a large number of tdm_arkcmps and never finished
Alice Chen created TRAFODION-277:
------------------------------------
Summary: LP Bug: 1320387 - Update statistics spawned a large number of tdm_arkcmps and never finished
Key: TRAFODION-277
URL: https://issues.apache.org/jira/browse/TRAFODION-277
Project: Apache Trafodion
Issue Type: Bug
Components: sql-cmp
Reporter: Weishiun Tsai
Assignee: Qifan Chen
Priority: Blocker
The following 3 statements hanged at the update statistics statement. The table only has 3 rows. After about 30 minutes, the update statistics statement spawned about 99 tdm_arkcmp processes on centos-mapr1, a 6-node cluster. The statement never finished before it was killed.
This was seen on the datalake v40535 build.
Here are the statements to reproduce this problem.
create table t (a int, b int, c int);
insert into t values (1,1,1),(2,2,2),(3,3,3);
update statistics for table t on a generate 100 intervals;
Here are the execution of the statements and snapshots of the tdm_arkcmp processes at each point. There were 2 tdm_arkcmp processes before update statistics statement was issued. 30 minutes after the statement was issued, the total number was 99 before the statement was killed.
[trafodion@centos-mapr1 scripts]$ sqps | grep tdm_arkcmp | wc -l
0
>>create table t (a int, b int, c int);
[trafodion@centos-mapr1 scripts]$ sqps | grep tdm_arkcmp | wc -l
1
>>insert into t values (1,1,1),(2,2,2),(3,3,3);
[trafodion@centos-mapr1 scripts]$ sqps | grep tdm_arkcmp | wc -l
2
>>update statistics for table t on a generate 100 intervals;
[trafodion@centos-mapr1 scripts]$ sqps | grep tdm_arkcmp | wc -l
99
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)