[ 
https://issues.apache.org/jira/browse/IMPALA-8059?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16738501#comment-16738501
 ] 

Sahil Takiar commented on IMPALA-8059:
--------------------------------------

Thanks for showing me that doc [~tarmstrong] helps a lot. Agree, would be great 
to link it in the wiki as well.

I think the main thing that is confusing me is the {{sleep}} function in 
Impala. I would assume that while {{sleep}} is running, the query would remain 
in RUNNING state and only transition to FINISHED state once the sleep has 
completed. The query {{select sleep(60000)}} doesn't produce any rows for at 
least 60 seconds (the {{sleep}} function returns {{true}}) so it should stay 
RUNNING for 60 seconds. However, testing on a mini-cluster, this doesn't seem 
to be the case. The query almost immediately transitions to the FINISHED state, 
and then waits for 60 seconds before returning any rows.

Perhaps thats by design just for the {{sleep}} function? I would assume other 
Impala functions would not behave that way - e.g. {{select 
some_long_running_function}} would be in the RUNNING state until the function 
produces at least one row, right?

> TestWebPage::test_backend_states is flaky
> -----------------------------------------
>
>                 Key: IMPALA-8059
>                 URL: https://issues.apache.org/jira/browse/IMPALA-8059
>             Project: IMPALA
>          Issue Type: Bug
>            Reporter: Pooja Nilangekar
>            Assignee: Sahil Takiar
>            Priority: Blocker
>              Labels: broken-build, flaky-test
>
> test_backend_states is flaky. The query reaches the _"FINISHED"_ state before 
> it's state is verified by the python test. Here are the relevant log: 
> {code:java}
> 07:33:45 ----------------------------- Captured stderr call 
> -----------------------------
> 07:33:45 -- executing async: localhost:21000
> 07:33:45 select sleep(10000) from functional.alltypes limit 1;
> 07:33:45 
> 07:33:45 -- 2019-01-08 07:31:57,952 INFO     MainThread: Started query 
> 7f46f15ed4d6d0f6:4d58cdbc00000000
> 07:33:45 -- getting state for operation: 
> <tests.common.impala_connection.OperationHandle object at 0x5b3db90>
> {code}
> This bug was introduced by IMPALA-7625.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-all-unsubscr...@impala.apache.org
For additional commands, e-mail: issues-all-h...@impala.apache.org

Reply via email to