flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF GitHub Bot (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-8686) Improve basic embedded SQL client
Date Wed, 18 Apr 2018 16:50:00 GMT

    [ https://issues.apache.org/jira/browse/FLINK-8686?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16442835#comment-16442835
] 

ASF GitHub Bot commented on FLINK-8686:
---------------------------------------

Github user walterddr commented on the issue:

    https://github.com/apache/flink/pull/5867
  
    Hi @twalthr, having JVM heap size configurable is definitely a great benefit. Just to
clarify, this is only changing the Client JVM heap size, correct? I am assuming this is mainly
for manipulating large size data when retrieving results from the query. 
    
    I think adding it to `flink-conf.yaml` should be fine. There are already specific configurations
commented out for components like Kerberos, ZK, etc. which makes adding a specific conf to
sql-client less disturbing. We can always break it out to another file later when extra configuration
files are needed for sql-client in the future. 
    
    What do you guys think?


> Improve basic embedded SQL client 
> ----------------------------------
>
>                 Key: FLINK-8686
>                 URL: https://issues.apache.org/jira/browse/FLINK-8686
>             Project: Flink
>          Issue Type: Sub-task
>          Components: Table API &amp; SQL
>            Reporter: Timo Walther
>            Assignee: Timo Walther
>            Priority: Major
>             Fix For: 1.5.0
>
>
> This issue describes follow-up issues that should be fixes in order to make the SQL client
more stable:
>  - Add more tests for executor
>  - Configure JVM heap size
>  - Limit changelog and table buffers
>  - "The input is invalid please check it again." => add allowed range
>  - Load dependencies recursively
>  - Clean up results in result store
>  - Improve error message for unsupported batch queries
>  - Add more logging instead swallowing exceptions
>  - List properties in error message about missing TS factory sorted by name
>  - Add command to show loaded TS factories and their required propeties
>  - Add command to reload configuration from files (no need to restart client)
>  - Improve error message in case of invalid json-schema (right now: {{java.lang.IllegalArgumentException:
No type could be found in node:<root>}}
>  - Add switch to show full stacktraces of exceptions
>  - Give error message when setting unknown parameters {{result-mode=changelog}} does
not give an error but should be {{execution.result-mode=changelog}}



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

Mime
View raw message