[Alioth-staff-replacement] [dom at earth.li: Taking over lists.alioth.debian.org]

[Dominic Hargreaves]

Apologies, it was pointed out that I did not include the existing DSA
team (and alioth admins? I don't know what the overlap is) in this email.

It was also pointed out that this mail looked like I hadn't consulted
sufficiently, but consultation was precisely the aim of the message
below so please take it as such. I should have been clearer in describing
that I was expecting my assumptions to be tested.

Thanks,
Dominic.

----- Forwarded message from Dominic Hargreaves <dom at earth.li> -----

Date: Sun, 1 Oct 2017 21:58:43 +0100
From: Dominic Hargreaves <dom at earth.li>
To: alioth-staff-replacement at lists.alioth.debian.org
Subject: [Alioth-staff-replacement] Taking over lists.alioth.debian.org

Hi all,

As per my emails on d-d[1] I have a proposal for continuing to support
lists.alioth.debian.org for a while longer than the current team are
planning.

It sees that there are plenty of reasons to want to do this, including
retaining mailable maintainer fields (1/3 of the archive currently
uses alioth addresses); list archives (including the same URL structure
as currently) and (possibly) filling a niche for small mailing lists
that tracker / lists.debian.org / gitlab doesn't meet. The response to
my suggestion has been fairly positive with a number of individuals
offering to help out either in response to me, or in a more general way
on the thread.

My offer is as follow:

* to provide replacement infrastructure for a lists.alioth.debian.org
  mailman 2 installation, taking the form of a VM hosted on a system in
  London with sufficient capacity for the service, together with a
  pre-existing mail hub setup with exim+spamassassin+clamav[2] and backup,
  monitoring and configuration management capabilities
* to assemble an administrative team composed of DDs (thanks for those
  who have already offered; please let me know again if you are on board
  with this proposal, noting any particular area you'd like to work on/
  have expertise with). Note that the non-Debian specific parts of the
  infrastructure are managed by a collection of individuals including
  myself who are Bcced to this email (collectively spiny.org.uk)
* To plan and lead the migration of the service from existing infrastructure
  in line a schedule agreed with the others involved in the migration of
  alioth services

Assumptions
-----------

* that there is general consensus that taking over the service in this
  way is appropriate
* that DSA will be willing to change DNS entries relating to
  lists.alioth.debian.org to support the migration, and otherwise
  support the service (eg by providing SSL certificates and similar
  administrative/integration tasks)
* that the existing alioth admins can assist by providing access to the
  existing installation for migration
* that mailman 2 will continue to be supported in Debian for the
  foreseeable future

Timescales
----------

As I mentioned I don't have time to work on this properly until mid-November,
and I note that the current migration schedule[3] suggests mailman be
disabled on 1st February. This leaves, realistically, 2-3 months to
execute the migration, which should be plenty.

* November 2017: assemble team; plan installation and migration in
  more details; define administrative policies (eg new list creation,
  AUPs, dealing with abuse...)
* December 2017: set up infrastructure and start test migrations
* January 2018: switch over to new infrastructure

The details of the migration are tbd but I have in mind migrating
all lists to start with, and post-migration starting a health-check,
eg to weed out lists that are unneeded and can be either disabled
or deleted. Of course I would be interested to hear how things are
currently working in this regard. Also, I expect that in time it will
become clearer what valid use cases for lists will be, and then
we will define an appropriate timescale for winding down list types
which don't fit that criteria (which I expect to be at least one
stable release cycle)

Next steps
----------

Once my assumptions are validated, and if there are no other objections
or counter-proposals, I would like to email the project with a summary
of the plan (particularly since it represents a change compared to the
last official word on this in [3]).

Then I suggest we schedule an IRC meeting of interested parties in
early-mid November.

I look forward to hearing people's thoughts on this.

Cheers,
Dominic.

[1] <https://lists.debian.org/debian-devel/2017/09/msg00270.html>
[2] based on information from IRC, 285GiB of disk space and approx 1GiB/day
    of email traffic
[3] <https://lists.debian.org/debian-devel-announce/2017/09/msg00004.html>

_______________________________________________
Alioth-staff-replacement mailing list
Alioth-staff-replacement at lists.alioth.debian.org
https://lists.alioth.debian.org/mailman/listinfo/alioth-staff-replacement


----- End forwarded message -----