Let's investigate what the Apache Beam and Apache Superset teams have done and what we can borrow from them. Losing (a lot of) our JIRA would be a shame if there are tools and experiences that would allow us to migrate with relative ease.
Regards, Bart On Wed, Oct 26, 2022 at 6:45 PM Sergio Ramazzina <[email protected]> wrote: > Ciao Hans, > > unfortunately I don't have any prior experience to share on this topic. > However, looking at the options you depicted, I personally prefer the idea > of starting from a clean state and then manually re-create the tickets we > are really interested to work on discarding the rest as garbage. > > It could take longer and requires a bit of effort but I think it could > give better results. > > Cheers > > Sergio > > On 2022/10/26 07:42:19 Hans Van Akelyen wrote: > > Hi All, > > > > As discussed in our previous thread we would all like to move to GitHub > > Issues. > > Now we will have to come up with a plan on how to make this happen… And > it > > will require a bit of work. > > > > Unfortunately there is nog magic button that migrates all our issues from > > Jira to GH. We will have to create a cut-off point and put our current > Jira > > project in read-only mode for historical reference. > > The good news however is that we are not the first project doing this > > migration. If needed and based on our decisions we can ask around with > > other projects on how they handled the work. > > > > I think the 2 major components we have to align on is the WHEN and the > WHAT. > > > > About the when, we *could* migrate at any time depending on what we > decide > > for the what. Migration during a running release means we will have to > > create release notes based on 2 sources, or copy over the already done > > tickets to GH to generate a complete release note report. Switching to GH > > Issues directly after a release IMHO seems like the best approach. For > the > > 2.2 release I think our target should be end of November, no one wants to > > make a release during the end of year and it would go by unnoticed by the > > masses. > > > > Now the harder part, the what. We can put our Jira in read-only and > migrate > > nothing! We would start with a clean slate and would have to re-create > the > > tickets we are working on manually. This means we will not be taking over > > our current backlog and all tickets would have to be reported again. The > > advantage is we don’t have to do much work. > > The second option is using some tools/scripts (will have to ask around > with > > other projects) to migrate all open issues to GH Issues. I would still > > suggest we first do a cleanup of the backlog (we currently have 760 > tickets > > in backlog). Even after cleaning up we will still have plenty to do… > > The final option, we split the work validate each ticket and manually > > migrate/update and add more information. > > > > For the move to GH Issues, you can also take a look at how Apache Beam > and > > Apache Airflow are handling things, they have templates and bots to help > > with initial triage/labeling of the issues. > > > > As always, feel free to share ideas/experience/comments! > > > > Cheers, > > Hans > > >
