On 16/02/2015 13:46, Stephen Mallette wrote:
> Hi Mark, 
> 
> Thanks for explaining the process.  There are about 500 issues to import:
> 
> https://github.com/tinkerpop/tinkerpop3/issues

I just did a test import to see how long it would take (not very). Take
a look at:
https://issues.apache.org/jira-test/browse/TINKERPOP3/?selectedTab=com.atlassian.jira.jira-projects-plugin:summary-panel

Make a list of what doesn't look right and let me know. I'll see if any
of it can be fixed in the import process.

> I think it would be worth some of my time to double-check Jira/GitHub
> accounts for those who have been most active in the issue list, so as
> not to create duplicate accounts by accident.  I think it would also be
> worthwhile starting a separate thread in the dev and user mailing lists
> to let folks know that we will be migrating and what the process will be
> so that they can project what their account will look like in apache
> jira once everything is moved over - that would give folks a chance to
> create a ASF Jira account with the same email as github or whatever.
> 
> I think it will take some coordination to get everyone notified properly
> of what we're doing as i just described and to figure out an "off-time"
> for github issues so that the migration can occur.  What is your
> availability to do the actual migration?  Do you have any particular
> dates that are good or bad to pull the trigger?

The import is easy. It is the user merging that takes the time.

I'm an infra volunteer so I need to fit this around $dayjob and $family.
This looks to be a relatively simple import so - assuming it would take
a couple of hours - it could be done late one weekend evening (probably
Saturday evening UTC). Any weekend from March onwards looks like a
possibility at the moment.

Mark

> 
> Thanks,
> 
> Stephen
> 
> 
> On Sun, Feb 15, 2015 at 8:36 AM, Mark Thomas <[email protected]
> <mailto:[email protected]>> wrote:
> 
>     On 14/02/2015 19:02, James Carman wrote:
>     > I do not know exactly what the requirements are.  Perhaps we should
>     > formally engage the infra folks (copied) for more help.
> 
>     Installing the Github importer plug-in shouldn't be a problem although
>     (see below) that might not be necessary.
> 
>     Resolving the usernames is (relatively) simple. We create (or re-use an
>     existing test instance just pointing to a new database) a clean Jira
>     install. Then we import the github issues into it and then use some SQL
>     we already have to fix the user names. The mapping would be via e-mail
>     address and would look something like:
>     for each e-mail/username pair in the import
>       if the email is associated with a user in the ASF Jira instance
>         update the imported username to the ASF username
>       else if the username also exists in the ASF Jira instance
>         update the imported username to append -tinkerpop to the end
>       else
>         use the imported username/email as is
> 
>     If the same person has different email addresses associated with the ASF
>     Jira account and their github account they will end up with two accounts
>     in the ASF Jira instance (and infra is unlikely to want to spend the
>     time to merge them). The simplest solution is to make sure the same
>     e-mail is addressed for both.
> 
>     Once the data has been imported into the test instance and the usernames
>     amended, we can export from the test instance and import into the live
>     instance.
> 
>     How many issues are we talking about here (so I have an idea of how much
>     work this might be)?
> 
>     Mark
> 
> 
> 
>     >
>     > On Thu, Feb 12, 2015 at 8:46 PM, Stephen Mallette
>     <[email protected] <mailto:[email protected]>> wrote:
>     >> I think I understand what you mean by "resolve the usernames" but
>     could you
>     >> please elaborate?  What would we have to do exactly to allow
>     proper import
>     >> to jira?
>     >>
>     >> I would also like to know what the alternatives are to "resolve the
>     >> usernames" in case that ends up being overly burdensome?  Do we
>     re-tread
>     >> old ground about continuing with github issues as that would be
>     easy to
>     >> migrate to the mirror I think (I seem to remember opposition to
>     that but I
>     >> don't recall the exact nature of the arguments either way)? Can
>     anyone say
>     >> what other projects have done in this situation?
>     >>
>     >> If we could have some options so that we could we weigh the level
>     of effort
>     >> vs impact to all the historical information we have in those
>     github issues,
>     >> I think we could make a decision and get this major piece of the
>     >> infrastructure migration done.
>     >>
>     >> Thanks,
>     >>
>     >> Stephen
>     >> On Feb 12, 2015 8:11 PM, "James Carman"
>     <[email protected] <mailto:[email protected]>> wrote:
>     >>
>     >>> Well they gave us our option, which looks like setting up a
>     trial instance
>     >>> and importing it there.  The caveat is the usernames.
>     >>>
>     >>> On Thursday, February 12, 2015, Stephen Mallette
>     <[email protected] <mailto:[email protected]>>
>     >>> wrote:
>     >>>
>     >>>> Hi James, on the subject of JIRA, you'd last said that there
>     was some
>     >>>> discussion with the infrastructure folks with respect to how we
>     migrate
>     >>>> over.  I was wondering if there was any new updates on that?
>     >>>>
>     >>>> On Mon, Feb 9, 2015 at 8:40 PM, James Carman
>     <[email protected] <mailto:[email protected]>
>     >>>> <javascript:;>>
>     >>>> wrote:
>     >>>>
>     >>>>> I have been chatting with the infra folks about this.  The
>     migration
>     >>>> plugin
>     >>>>> might work but they are reluctant to add more stuff to our
>     JIRA at this
>     >>>>> time since JIRA is very slow already.  Another option is to
>     spin up a
>     >>>> trial
>     >>>>> JIRA instance and migrate into it.  Then migrate from trial
>     JIRA -> ASF
>     >>>>> JIRA.  The versions would have to match exactly and we would
>     need to
>     >>>>> resolve usernames first.
>     >>>>>
>     >>>>> On Monday, February 9, 2015, Stephen Mallette
>     <[email protected] <mailto:[email protected]>
>     >>>> <javascript:;>>
>     >>>>> wrote:
>     >>>>>
>     >>>>>> I'm not sure if we've had other discussion around this topic,
>     so i'll
>     >>>>> just
>     >>>>>> use this thread to continue on the topic of the issue
>     tracker.  At
>     >>> this
>     >>>>>> point, we have our mirror to github, pull request integration is
>     >>>> working
>     >>>>>> nicely and we have a separate thread going about getting Travis
>     >>>>> integration
>     >>>>>> running against the mirror - all excellent.
>     >>>>>>
>     >>>>>> I'm now turning my attention to migrating the issues in the old
>     >>>>> tinkerpop3
>     >>>>>> repository to a new home.  I assume that home is JIRA - two
>     >>> questions:
>     >>>>>>
>     >>>>>> 1. how do we get that setup? do i create a JIRA ticket in INFRA?
>     >>>>>> 2. is there a known, easy way to migrate all of our github
>     issues to
>     >>>>> JIRA?
>     >>>>>> i found this
>     >>>>>>
>     >>>>>>
>     >>>>>
>     >>>>
>     >>>
>     
> https://answers.atlassian.com/questions/133617/moving-to-jira-from-github-issues
>     >>>>>> but don't know if it's good or not.
>     >>>>>>
>     >>>>>> Thanks,
>     >>>>>>
>     >>>>>> Stephen
>     >>>>>>
>     >>>>>>
>     >>>>>>
>     >>>>>> On Fri, Feb 6, 2015 at 12:10 PM, Rob Vesse
>     <[email protected] <mailto:[email protected]>
>     >>>> <javascript:;>
>     >>>>>> <javascript:;>> wrote:
>     >>>>>>
>     >>>>>>> Marko
>     >>>>>>>
>     >>>>>>> Not an Infra person I'm afraid, just an interested (and very
>     new)
>     >>>> IPMC
>     >>>>>>> member trying to help answer your projects questions where I can
>     >>>>>>>
>     >>>>>>> Dave as VP Infra can better advise on what needs to be done
>     to get
>     >>> a
>     >>>>>> GitHub
>     >>>>>>> mirror, I honestly don't know if this is automatic or if you'll
>     >>> need
>     >>>> to
>     >>>>>>> file
>     >>>>>>> a JIRA
>     >>>>>>>
>     >>>>>>> Rob
>     >>>>>>>
>     >>>>>>> From:  Marko Rodriguez <[email protected]
>     <mailto:[email protected]> <javascript:;>
>     >>>> <javascript:;>>
>     >>>>>>> Date:  Friday, 6 February 2015 08:44
>     >>>>>>> To:  "[email protected]
>     <mailto:[email protected]> <javascript:;>
>     >>>> <javascript:;>"
>     >>>>>>> <[email protected]
>     <mailto:[email protected]> <javascript:;>
>     <javascript:;>>
>     >>>>>>> Cc:  David Nalley <[email protected] <mailto:[email protected]>
>     <javascript:;> <javascript:;>>,
>     >>>> Rob Vesse <
>     >>>>>> [email protected] <mailto:[email protected]>
>     <javascript:;> <javascript:;>>
>     >>>>>>> Subject:  Setting up GitHub Issues? --- or do we go JIRA?
>     >>>>>>>
>     >>>>>>>> Hi David/Rob,
>     >>>>>>>>
>     >>>>>>>> We have asked Rob to set up the Apache TinkerPop mirror on
>     >>> GitHub.
>     >>>>>>>>
>     >>>>>>>> https://github.com/apache/tinkerpop
>     >>>>>>>>
>     >>>>>>>> Stephen and I see that mirrors allow Pull Requests. Can we just
>     >>> use
>     >>>>> the
>     >>>>>>> issue
>     >>>>>>>> tracker on GitHub mirror as well OR do we have to go JIRA. Our
>     >>>>>>> preference, is
>     >>>>>>>> GitHub mirror issue tracker.
>     >>>>>>>>
>     >>>>>>>> Thank you very much,
>     >>>>>>>> Marko.
>     >>>>>>>>
>     >>>>>>>> http://markorodriguez.com
>     >>>>>>>>
>     >>>>>>>
>     >>>>>>>
>     >>>>>>>
>     >>>>>>
>     >>>>>
>     >>>>
>     >>>
> 
> 

Reply via email to