Hello,
I think this could be very helpful as some way of making a
comparison between the different options as I get a bit confused
by the technical descriptions and determining what options are
best and what are the basic necessities in terms of serviceability
and reliability.
One criteria that seems to stick out is finding a location that is
compatible with and sympathetic to what NETTIME is about.
best
a
On 2023. 06. 21. 16:04, John Preston
via Nettime-tmp wrote:
Our ticket at Riseup has been assigned but there is no
response yet. I will keep checking and let y'all know as soon as
there is any movement. 🙂
We seem to have a few options floating around now
but perhaps little shared clarity around what the different
factors are we want to manage, and how much we all care about
those. May I suggest someone set up a
Jamboard/Miro/Padlet/Trello/... so that we can start to collect
together the different
factors/dimensions/concerns/properties/benefits/risks/opportunities...
of potential solutions? That way it might be easier to come to a
collective decision. Happy to lead on this if people think it's
a good idea! 😊
Thanks,
John (they/them)
Hey Allan, others,
I've been pondering the trade off between
paying for a commercial host versus looking for non-paid
solutions. Initially i thought it "obvious" that we should
pick a service (whether turn-key or something offering full
control such as Panix indeed) and simply pay for it. I'd
still be more than happy to front up the first year's
hosting costs.
However, i reread with interest Ted, Felix and
David's state of nettime retrospective from 2015 [1], where
it's emphasised that money never changed hands. I think
that's laudable, and indeed, financial arrangements often
imply expectations and demands that could turn thing sour in
difficult moments. But on the other hand, "volunteer"
servers such as ljudmila.org are being paid for by someone,
so in that sense there would be no change.
To try and practically advance things without
getting too caught up in idealistic or theoretical
discussions, my draft proposal would be:
* Ensure wherever we host, we can make use of
our own domain name. Future hosting changes (which i expect
will be a fact of life) will at least be transparent to
subscribers,
* take Rich's suggestion of a trusted hosting
company where we're more likely to get a good-reputation
sending IP, like Panix,
* i'd be happy to pay the first yearly hosting
fee and do initial setup of the DNS and basic software to
run Mailman,
* of course, i'd ensure the other volunteer
janitors named by Christian get SSH access to the machine as
soon as possible so i do not become a SPOF.
* If additional moderators want to volunteer, i
welcome that, but i would suggest we at least introduce
ourselves during the next video sync, to build trust
(explicitly NOT an interview, we have no right to start
gatekeeping).
This would allow us to actually relieve Felix
as he has requested within a week or two, optimistically.
This is just my proposal - i can start on this
work this weekend, unless folks have concerns.
In the meantime, Jordan and John, have you
heard back from Riseup.net or Rhizome.org?
Thanks,
p.
Hello,
Obviously there are differences on how to solve the
technical issues just as there are questions regarding
the social issues; what would be helpful if those with
the expertise and experience dealing with the
technical questions proposed a method for sorting out
these questions; AND, secondly, what are the expenses
involved in these various solutions (hosting, etc.)
and how do these expenses get covered?
Also, I'm wondering - because the social and technical
questions are interrelated - how do the various
technical solutions impact how we are able to address
the social issues that have been mentioned? I think
this is why the hosting service really matters...
best
allan
On 2023. 06. 15. 0:42, Jaime Magiera via
Nettime-tmp wrote:
Yes, apparently it hit a nerve for you. The point was that a container would need to be hosted, one option would be on a Kubernetes cluster service. Otherwise, it’s just running on someone’s VM. There are many non-commercial options. Containerization and Kubernetes are my day job. Likewise, I managed mailman mailing lists for several decades. Portability is a key technical hurdle as the list hosting changes over time.
I don’t think anyone is prioritizing technical over social. The technological issue is clear and being discussed as a separate issue.
On Jun 14, 2023, at 6:06 PM, paul van der walt via Nettime-tmp <nettime-tmp@mail.ljudmila.org> wrote:
Hey Jaime, Christian,
On 2023-06-14 at 11:39 -04, quoth Jaime Magiera via Nettime-tmp <nettime-tmp@mail.ljudmila.org>:
I’ve been quietly following this discussion, but will pipe in on this aspect: Running from a
container would be a wise solution. The archives can be stored on a mount and backed up
elsewhere. I’m happy to provide my expertise in the area of containerization (and Kubernetes if so
desired) to help if this is the way folks decide to go.
I appreciate folks are just brainstorming, but i feel i should add my 2c too. It is my literal day job to support a fairly sizeable e-commerce website (millions to billions of SKUs, millions of requests per minute) with AWS infrastructure, and we use a lot of Kubernetes and Docker. In that context, the trade-offs make sense. But i guess my only plea would be, let's please not overcook and overcomplicate things from the get-go. Bringing Kubernetes into the discussion is almost the canonical example of over-complication for hosting a mailing list.
I think it's noble and understandable to want to do work up-front to make things infinitely lift-and-shiftable, but personally my philosophy is what is sometimes jokingly called "KISS - keep it simple, stupid". Concretely, that would mean i'd favour using (e.g.) plain-old Mailman from a package repository of Linux or indeed (Rich's suggestion) OpenBSD for stability and security.
If i'm to be involved in the technical side of things (and that's the main reason i volunteered for janitorial duties) i'd want to hold off on committing to any one particular hosting company / technological choice / etc. because, as others have pointed out, our main difficulties are social.
I hope my response is sufficiently measured, but the mention of Kubernetes hit a bit of a nerve for me :).
Cheers,
p.
# distributed via <nettime>: no commercial use without permission
# <nettime> is a moderated mailing list for net criticism,
# collaborative text filtering and cultural politics of the nets
# more info: https://mail.ljudmila.org/cgi-bin/mailman/listinfo/nettime-tmp
# archive: http://www.nettime.org contact: nettime@kein.org
# @nettime_bot tweets mail w/ sender unless #ANON is in Subject:
# distributed via <nettime>: no commercial use without permission
# <nettime> is a moderated mailing list for net criticism,
# collaborative text filtering and cultural politics of the nets
# more info: https://mail.ljudmila.org/cgi-bin/mailman/listinfo/nettime-tmp
# archive: http://www.nettime.org contact: nettime@kein.org
# @nettime_bot tweets mail w/ sender unless #ANON is in Subject:
# distributed via <nettime>: no
commercial use without permission
# <nettime> is a moderated mailing list
for net criticism,
# collaborative text filtering and cultural
politics of the nets
# @nettime_bot tweets mail w/ sender unless
#ANON is in Subject:
# distributed via <nettime>: no commercial use without permission
# <nettime> is a moderated mailing list for net criticism,
# collaborative text filtering and cultural politics of the nets
# more info: https://mail.ljudmila.org/cgi-bin/mailman/listinfo/nettime-tmp
# archive: http://www.nettime.org contact: nettime@kein.org
# @nettime_bot tweets mail w/ sender unless #ANON is in Subject:
|