You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lens.apache.org by "Yash Sharma (JIRA)" <ji...@apache.org> on 2015/06/08 15:31:03 UTC

[jira] [Comment Edited] (LENS-539) Fix issues reported in 2.1 release

    [ https://issues.apache.org/jira/browse/LENS-539?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14577183#comment-14577183 ] 

Yash Sharma edited comment on LENS-539 at 6/8/15 1:30 PM:
----------------------------------------------------------

I would like to pick this up if no one is working on it already.

*Identified issues:*

1. Header in data files :
Data files that do not involve any creativity can be skipped for headers. Which looks like our case.

2. Justin's error for missing Database:
This error originates out of Hive and there is little we can do. The identified testcase is:
org.apache.lens.server.stats.TestStatisticsLogPartitionHandler.configureHiveTables(TestStatisticsLogPartitionHandler.java:128)
org.apache.lens.server.stats.TestStatisticsLogPartitionHandler.testPartitionHandler(TestStatisticsLogPartitionHandler.java:63)

We use hive.dropDatabase(LensConfConstants.DEFAULT_STATISTICS_DATABASE, true, true); 
which has ignoreUnknownDatabases flag as true but the stack trace still would still be visible on console. We should have a new JIRA to figure out similar testcases which have error stacks on console.

3. Enhance README :
This is a minor point and I can add that JAVA_HOME is expected. Let me know any other info thats on top of your mind.

4. MD5 content not formatted:
gpg --print-md5 works fine for me on Ubuntu. We can try something that would be consistent across environments-
generate with: gpg --print-md MD5 ${dist_name}.tar.gz  | tr -d '\n' > ${dist_name}.tar.gz.md5
verify with: gpg --print-md MD5 ${dist_name}.tar.gz | tr -d '\n' | diff - ${dist_name}.tar.gz.md5

5. Signed email has "(Works for Hadoop and ecosystem)”
Have less idea on this. Experienced contributors please share your thoughts.

6. pom.xml listing Devs:
I have less say on this and existing contributors must decide on this.
I can suggest to have a dedicated webpage for Lens Team on the official page. It would be easies to customize and more exposed to outer world. We can have details like- contributor name, Status (Committer/PMC/PPMC etc), And link to Github profile. Thoughts ?


was (Author: yash360@gmail.com):
I would like to pick this up if no one is working on it already.

*Identified issues:*

_1. Header in data files : _
Data files that do not involve any creativity can be skipped for headers. Which looks like our case.

_2. Justin's error for missing Database: _
This error originates out of Hive and there is little we can do. The identified testcase is:
org.apache.lens.server.stats.TestStatisticsLogPartitionHandler.configureHiveTables(TestStatisticsLogPartitionHandler.java:128)
org.apache.lens.server.stats.TestStatisticsLogPartitionHandler.testPartitionHandler(TestStatisticsLogPartitionHandler.java:63)

We use hive.dropDatabase(LensConfConstants.DEFAULT_STATISTICS_DATABASE, true, true); 
which has ignoreUnknownDatabases flag as true but the stack trace still would still be visible on console. We should have a new JIRA to figure out similar testcases which have error stacks on console.

_3. Enhance README :_
This is a minor point and I can add that JAVA_HOME is expected. Let me know any other info thats on top of your mind.

_4. MD5 content not formatted:_
gpg --print-md5 works fine for me on Ubuntu. We can try something that would be consistent across environments-
generate with: gpg --print-md MD5 ${dist_name}.tar.gz  | tr -d '\n' > ${dist_name}.tar.gz.md5
verify with: gpg --print-md MD5 ${dist_name}.tar.gz | tr -d '\n' | diff - ${dist_name}.tar.gz.md5

_5. Signed email has "(Works for Hadoop and ecosystem)”_
Have less idea on this. Experienced contributors please share your thoughts.

_6. pom.xml listing Devs:_
I have less say on this and existing contributors must decide on this.
I can suggest to have a dedicated webpage for Lens Team on the official page. It would be easies to customize and more exposed to outer world. We can have details like- contributor name, Status (Committer/PMC/PPMC etc), And link to Github profile. Thoughts ?

> Fix issues reported in 2.1 release
> ----------------------------------
>
>                 Key: LENS-539
>                 URL: https://issues.apache.org/jira/browse/LENS-539
>             Project: Apache Lens
>          Issue Type: Improvement
>          Components: build
>    Affects Versions: 2.1
>            Reporter: Amareshwari Sriramadasu
>            Priority: Critical
>             Fix For: 2.2
>
>
> Here are the mailing threads for more details :
> http://mail-archives.apache.org/mod_mbox/incubator-general/201505.mbox/%3C0E519FCC-EF9D-4548-9A7F-6AE6CE9EEE5F%40me.com%3E
> http://mail-archives.apache.org/mod_mbox/incubator-lens-dev/201505.mbox/%3CCABJEuZc1Vgi0Zs-6hfy7hMTGutd6FXDf7W5qJfPi_tUPy8NWww%40mail.gmail.com%3E
> and
> http://mail-archives.apache.org/mod_mbox/incubator-lens-dev/201505.mbox/%3CBD17B5FA-7653-4B9B-BA49-612A3391E5BB%40me.com%3E



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