On Tue, Jul 5, 2011 at 12:43 PM, Pedro F. Giffuni <giffu...@tutopia.com> wrote: > > --- On Tue, 7/5/11, Rob Weir <apa...@robweir.com> wrote: > ... > >> >> I'd like to explore the technical feasibility of migrating >> to JIRA. >> > > +1 > >> I see this page here describing the migration from JIRA's >> perspective: >> >> http://www.atlassian.com/software/jira/tour/bugzilla-importer.jsp >> >> This appears to require db admin credentials. >> >> Does anyone have experience doing this kind of migration? >> > > I am sure infra@ has the experience. >
OK. cc'ing infrastructure. > I can't find it in the archives but someone in the list > offered to make available the bugzilla archive. We should > probably just make an experimental conversion and keep it > if things go well. > How does that help? The JIRA import feature seems to operate on the live Bugzilla instance, or at least the live database. I can see that doing a migration over the internet would be slow and possible prone to time out errors. I suspect there are a variety of approaches possible: Would it be unnecessarily complex to do a Bugzilla dump, import to a temporary Apache Bugzilla instance, then import into to the real Apache OOo JIRA instance? Advantages are if we need to redo the JIRA conversion due to problems, we can iterate on that step much faster. It is also an opportunity to ensure we're starting (hopefully) from the latest stable Bugzilla rev, which may migrate easier. >> And from the project perspective, I assume we want to >> migrate all issues, old, new, open closed, everything? >> And we must preserve IDs. > > The priority is new open, but the most we can preserve > the better. I would expect all the code in there is > covered by the grant too, so we are probably getting a > lot of code that was initially rejected but is part of > some unnamed forks. > Good point on the patches. -Rob > Pedro. >
