Firstly I like and appreciate the role that bodhi has in our
infrastructure and this discussion is not about changing the current
workflow.
But the current state of the project deserves some visibility. There are
a few issues currently, just to list some that I have picked up:
- FTBFS state of `bodhi-client`
- Web UI freezing on big updates
- Interface/ability to CC the main maintainers when packit builds on rawhide
This not to complain about this state, as I am grateful as anybody here
for the work done in this area, and would gladly contribute as I can.
But as Mattia puts it [1] "IMO there would a lot of reasons to just
replace bodhi with a new tool with a target for more automation", see
the reference for yet another list of reasoning from the maintainer's
perspective.
Short term future
Right now bodhi still works fine, there is only one upcoming known
breakage due to the src.fp.o migration to forgejo, that is still
possible to address and is being worked on [2]. So for the time being
there should not be any disturbances expected and we can continue with
the usual workflow and the bugs we came to appreciate :). (Unless anyone
else knows of more breakages that will be coming up?)
Long term future
So what do we do for the long term? Mattia has a point about the need to
rewrite or at least refactor away from the unmaintained dependencies.
But life happens to us all and how many people do we have available to
work on that?
In tangentially relevant insight from Redhat, they are also trying to
find a solution for a system for more general aggregation of test
results, and they are open to expanding the scope to manage gating, so
it may not be too far off to consider how bodhi could fit into this. The
team being formed to work on this expressed interest in accommodating
the Fedora workflow needs, and the planning is still in early stages
that this can be adapted to best fit a bodhi replacement or individual
components.
There could well be other services that are already developed and
maintained that we could adapt to fit in or around bodhi. Do name those
out if anyone know or ping someone who might know. Yes, let's address
the konflux shaped elephant waiting for this email to be sent.
In the same spirit as the bugzilla replacement ticket [3], I think we
should start with gathering everyone's user stories on what exactly is
bodhi's role in the current workflow in the broader picture. What do you
see as the good features of the current workflow, and what major things
you find missing or would need attention? In looking or designing an
alternative, what are some key designs that should be considered (e.g.
should it be a standalone service, or part of a bigger service)? We also
need to gather the technical side how everything connects around bodhi.
First step
Before anything, where should the tracker for this thing be. As a releng
or fesco ticket? Hopefully not as an email/discussion thread?
[1]:
https://github.com/fedora-infra/bodhi/issues/6129#issuecomment-4801000412
[2]: https://github.com/fedora-infra/bodhi/issues/6146
[3]: https://forge.fedoraproject.org/fesco/tickets/issues/3664
--
_______________________________________________
devel mailing list -- [email protected]
To unsubscribe send an email to [email protected]
Fedora Code of Conduct:
https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives:
https://lists.fedoraproject.org/archives/list/[email protected]
Do not reply to spam, report it:
https://forge.fedoraproject.org/infra/tickets/issues/new
