hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Junping Du (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-679) add an entry point that can start any Yarn service
Date Fri, 26 Aug 2016 16:50:21 GMT

    [ https://issues.apache.org/jira/browse/YARN-679?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15439375#comment-15439375
] 

Junping Du commented on YARN-679:
---------------------------------

Agree with [~templedf] that this is well written javadoc that can be good example for reference
in future. :)
Good to see we finally figure out Jenkins' problem last week. Sorry I haven't figure out too
much time this week to review it thoroughly. Just have some level comments here:
1. It looks like all changes are in hadoop-common project currently. If so, shall we turn
it into Hadoop JIRA instead of a YARN issue? Or we think these APIs only get used for YARN
application. If so, may be we should consider to refactor most code to under hadoop-yarn-project
(like: yarn-api)?

2. The lifecycle of yarn service looks good to me. However, I haven't see an detail example
to show case how an application can use these code to launch yarn service. Do we have a quick
dummy application here or we can plan to add one in future.


3. I saw we were wrapping a lot of error code here around all kinds of exceptions or HTTP
code.
{noformat}
+ * <pre>
+ *    0-10: general command issues
+ *   30-39: equivalent to the 3XX responses, where those responses are
+ *          considered errors by the application.
+ *   40-49: client-side/CLI/config problems
+ *   50-59: service-side problems.
+ *   60+  : application specific error codes
+ * </pre>
{noformat}
It sounds like 11-20 is vacant. Is that reserved for other purpose. Also, if error code is
used out (like service side problem is more than 10 kinds). How can we extend according to
current design?

I need more time for check more details of the patch which is very huge now. 

> add an entry point that can start any Yarn service
> --------------------------------------------------
>
>                 Key: YARN-679
>                 URL: https://issues.apache.org/jira/browse/YARN-679
>             Project: Hadoop YARN
>          Issue Type: New Feature
>          Components: api
>    Affects Versions: 2.8.0
>            Reporter: Steve Loughran
>            Assignee: Steve Loughran
>         Attachments: YARN-679-001.patch, YARN-679-002.patch, YARN-679-002.patch, YARN-679-003.patch,
YARN-679-004.patch, YARN-679-005.patch, YARN-679-006.patch, YARN-679-007.patch, YARN-679-008.patch,
YARN-679-009.patch, YARN-679-010.patch, YARN-679-011.patch, org.apache.hadoop.servic...mon
3.0.0-SNAPSHOT API).pdf
>
>          Time Spent: 72h
>  Remaining Estimate: 0h
>
> There's no need to write separate .main classes for every Yarn service, given that the
startup mechanism should be identical: create, init, start, wait for stopped -with an interrupt
handler to trigger a clean shutdown on a control-c interrupt.
> Provide one that takes any classname, and a list of config files/options



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org


Mime
View raw message