[
https://issues.apache.org/jira/browse/TAJO-1321?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14300884#comment-14300884
]
ASF GitHub Bot commented on TAJO-1321:
--------------------------------------
Github user jinossy commented on a diff in the pull request:
https://github.com/apache/tajo/pull/369#discussion_r23907071
--- Diff:
tajo-core/src/main/java/org/apache/tajo/master/TajoMasterClientService.java ---
@@ -428,9 +426,7 @@ public GetQueryListResponse
getFinishedQueryList(RpcController controller, GetQu
infoBuilder.setState(queryInfo.getQueryState());
infoBuilder.setQuery(queryInfo.getSql());
infoBuilder.setStartTime(queryInfo.getStartTime());
- long endTime = (queryInfo.getFinishTime() == 0) ?
- System.currentTimeMillis() : queryInfo.getFinishTime();
- infoBuilder.setFinishTime(endTime);
+ infoBuilder.setFinishTime(System.currentTimeMillis());
--- End diff --
It seems like last request time..
I suggest that the {{isFinishState}} branch in
{{QueryInProgress.hearbeart}} sets the finish time
> Cli prints wrong response time
> ------------------------------
>
> Key: TAJO-1321
> URL: https://issues.apache.org/jira/browse/TAJO-1321
> Project: Tajo
> Issue Type: Bug
> Components: cli
> Reporter: Jihoon Son
> Assignee: Jihoon Son
> Priority: Critical
> Fix For: 0.10
>
>
> When I run Tajo on two different clusters, the cli prints wrong response time.
> Here is examples.
> - Cluster1: the response time decreases when the query is finished.
> {noformat}
> tpch100> select count(*) from customer;
> Progress: 30%, response time: 1.834 sec
> Progress: 30%, response time: 1.834 sec
> Progress: 30%, response time: 2.235 sec
> Progress: 30%, response time: 3.036 sec
> Progress: 43%, response time: 4.037 sec
> Progress: 43%, response time: 5.038 sec
> Progress: 100%, response time: 4.878 sec
> ?count
> -------------------------------
> 15000000
> (1 rows, 4.878 sec, 9 B selected)
> {noformat}
> - Cluster2: the response time jumps when the query is finished.
> {noformat}
> tpch_hdfs> select count(*) from customer;
> Progress: 0%, response time: 0.908 sec
> Progress: 0%, response time: 0.909 sec
> Progress: 0%, response time: 1.31 sec
> Progress: 0%, response time: 2.111 sec
> Progress: 48%, response time: 3.111 sec
> Progress: 100%, response time: 24.104 sec
> ?count
> -------------------------------
> 15000000
> (1 rows, 24.104 sec, 9 B selected)
> {noformat}
> I suspect that the wrong information is stored when the query history is
> written.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
