www-infrastructure-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jonathan Hsieh (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (INFRA-3764) Migrate JIRA project from https://issues.cloudera.org/browse/FLUME to https://issues.apache.org/jira/browse/FLUME
Date Wed, 20 Jul 2011 01:16:57 GMT

    [ https://issues.apache.org/jira/browse/INFRA-3764?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13068115#comment-13068115
] 

Jonathan Hsieh commented on INFRA-3764:
---------------------------------------

A group of us (medthomas, gmcdonald, _hen, jmhsieh, abayer, esammer) had a chat this afternoon
in #asfinfra about options for porting jira information from cloudera.org to apache.org. 


There seem to be a few main concerns:
* Preserving the name and description of particular issues.
* Preserving attribution and comment history of particular issues.
* Ambiguous issue names (FLUME-xxx on one jira vs the other) that will cause confusion in
user communities and public histories. 
* Restrictions from Cloudera's side about privacy and personally identifying information.

* ASF policy about link direction, jira history, attribution, and "issue wholeness".

We enumerated 4 options:

Option 1: Direct jira import from cloudera.org to apache.org with anonymized emails / passwords.
 
* Cloudera exports jira db
* Cloudera anonymizes emails and passwords (and publically exposed user names?)
* asfinfra imports using JIRA import/export mechanism.
* cloudera.org jira goes read-only afterwards, with open issues being closed with link pointing
to Apache JIRA
+  Both JIRA's would be called FLUME/SQOOP, all history and comments preserved.

Option 2: New Apache jira with new name to avoid name clash.
* Create new jira called FLUME2/SQOOP2 or something like that to avoid naming clash/change
pain.  
* Export summary and description of existing open issues, and get them imported as new issues
in the FLUME2/SQOOP2 jira.
* cloudera.org jira goes read-only afterwards, with open issues being closed with link pointing
to Apache JIRA

Option 3: Bump apache jira number up artificially to avoid name clash.
* Check to see if ASF ok with artificially inflating jira number
* If ok, add remove issues until we get to a number like 1000.
* Export summary and description of existing open issues, and get them imported as new issues
in the apache.org FLUME/SQOOP jira.
* cloudera.org jira goes read-only afterwards, with open issues being closed with link pointing
to Apache JIRA

Option 4: "manually" port description to be consistent with cloudera.org jira.
* Manually copy summary and description from cloudera.org jira issues to corresponding apache.org
jira.  This is done so that the old jira issues numbers correspond to the numbers in both
jiras.
* Backlink apache.org jira to point to cloudera.org jira. 
* cloudera.org jira goes read-only afterwards, with open issues being closed with link pointing
to Apache JIRA.  All new comments and progress would happen on apache.org jira.
---- 

Option 1 seems to be the technically preferred option.  It requires the least work by all
parties but also faces the most difficult privacy/legal concerns from the Cloudera side.

Option 2 and Option 3 seem acceptable,  IMO, Option 3 is preferred from podling-side but is
not clear acceptable from the ASF policy side.

Option 4 would be sufficient and more ideal than option 2 or 3 but would also likely require
the most effort on the cloudera side.  It is also unacceptable by ASF from a policy stance.

* Whole issues should live in apache, so "backpointer" from apache.org jira to cloduera.org
jira not ok without the summary, description, comments, and attribution.  Note however, forward
pointer from cloudera.org to apache.org is acceptable.
* This constraint essentially forces us back to Option 1.

Speaking for the folks in the Podling, we'd be willing to do "busy work" necessary port issues
manually or inflate jira numbers.

> Migrate JIRA project from https://issues.cloudera.org/browse/FLUME to https://issues.apache.org/jira/browse/FLUME
> -----------------------------------------------------------------------------------------------------------------
>
>                 Key: INFRA-3764
>                 URL: https://issues.apache.org/jira/browse/INFRA-3764
>             Project: Infrastructure
>          Issue Type: Wish
>      Security Level: public(Regular issues) 
>          Components: JIRA
>            Reporter: Arvind Prabhakar
>
> The incubating Flume project was previously hosted by Cloudera and the issue tracking
system was JIRA located at https://issues.cloudera.org/browse/FLUME. We would like to get
the information in Cloudera hosted JIRA to be migrated to the ASF JIRA. 

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message