Import is done. However, the issue keys do not match because JIRA uses a different strategy for re-creating the Subtask issues (it defers them until the end of the import). For example, in the following issue the subtasks have much higher keys, you can see: https://issues.apache.org/jira/browse/HOP-1482 Frankly, this is all I can do for now.

I have confirmed all existing 626 issues have been imported successfully, including the Resolution status.

-Max

On 04.10.20 13:35, Maximilian Michels wrote:
Hey Hans,

Great!

I've managed to add "Resolution" as label, so I can filter them afterwards and set the correct "Resolution" status. I also noticed JIRA doesn't update existing issues based on the issue ID. So I have to delete the already imported ones - we are anyways already off with the issue keys, so it doesn't matter.

Re-running the import now. Have to split the CSV into three packages, each 250 issues (limitation of the bulk import).

-Max

On 04.10.20 13:26, Hans Van Akelyen wrote:
Hi Max,

I disabled the ability to change anything in our jira when I did the
export. so we are in-sync now.
It seems changing the commit messages isn't that hard [1]. Will go ahead
and make a pipeline to find the HOP messages and increment the number by
one.

Then we only have to clean things back up a bit in the new Jira, and we are
ready to go!

Hans

[1]
https://docs.github.com/en/free-pro-team@latest/github/committing-changes-to-your-project/changing-a-commit-message


On Sun, Oct 4, 2020 at 1:02 PM Maximilian Michels <[email protected]> wrote:

(1) Turns out, the only way we can keep the existing issue keys, is to
re-create the Hop JIRA and then run the import. Otherwise all keys will
be off-by-one because HOP-1 has already been created. Even deleting the
issue won't solve that problem.

(2) Another possibility is to just run the import and adjust the Git
history to make up for the offset.

I've went ahead with option (2) because option (1) can still be applied
afterwards. The issues are marked with the "pre-apache" version. Let me
know if you want to update the import as it might have become stale in
the meantime (that's easily possible now because we can keep JIRA ids
which are unique identifiers; not possible for JIRA keys though).

The "Resolution" field couldn't be imported which is really a bummer.
I'll try to find a way to update the resolution status of the issues.

Current status: https://issues.apache.org/jira/projects/HOP/issues/

-Max

On 28.09.20 17:50, Maximilian Michels wrote:
  From what I just discovered, it is possible to keep the issue ids/keys
present in the CSV file:

We might have to move HOP-1 though if we want to avoid a conflict. I'll
try completing the migration tomorrow.

-Max

On 28.09.20 17:24, Hans Van Akelyen wrote:
Hi Max,

Unless someone else disagrees I think the mismatch between commits and
issue numbers pre-incubation is not a huge issue.
It's more important to have them correct for future development and
PR's.
To avoid confusion will just add a remark somewhere that pre HOP-xxx it
won't match.

The full issue export from our current jira project can be found here
<
https://drive.google.com/file/d/173CvE5N-zo0IXRFVXS7gxga8Jxe3Pi1V/view?usp=sharing

.

Thanks for the quick response.

Hans

On Mon, Sep 28, 2020 at 4:08 PM Maximilian Michels<[email protected]>
wrote:

Hi Hans,

I think it is possible to either keep the existing issues numbers or
start with an offset. There is an import feature [1], however it does
not seem to support keeping the numbering. So maybe we should import
the
existing issues first and then continue from there.

Perhaps you can prepare a CSV dump [2] and we can try? I'll also talk
with the INFRA team.

-Max

[1]


https://issues.apache.org/jira/secure/BulkCreateSetupPage!default.jspa?externalSystem=com.atlassian.jira.plugins.jira-importers-plugin:bulkCreateCsv
[2]


https://support.atlassian.com/jira-software-cloud/docs/create-issues-using-the-csv-importer/

On 28.09.20 15:56, Hans Van Akelyen wrote:
Small side question, when we move our code to the ASF github repo our
history is copied over?
We are currently already using the HOP- tag for issues, so I'm
guessing
jira will get a bit confusing until we pass the ticket numbers that
are
already included in our commits?
Or could we for example change the start numbering to 1000 so we are
not
in
conflict with our current tickets?

Cheers,
Hans

On Mon, Sep 28, 2020 at 3:44 PM Matt Casters <[email protected]
.invalid>
wrote:

I created HOP-1 : migration of the existing open JIRA cases so I
guess
it
works :-)
Thanks max!!

On Mon, Sep 28, 2020 at 2:38 PM Maximilian Michels<[email protected]>
wrote:
You can find the JIRA at
https://issues.apache.org/jira/projects/HOP/issues/

I might have to give you permissions to create issues, so far only
the
mentors should have full access.

-Max

--
Neo4j Chief Solutions Architect
*✉   *[email protected]
☎  +32486972937



Reply via email to