State of Things

Gary Kramlich grim at
Thu Dec 1 06:33:46 EST 2022

TL;DR Trying to define a statement of work and figure out a rough timeline.

On Tue, Nov 29, 2022 at 9:23 PM Gary Kramlich <grim at> wrote:
> On Tue, Nov 29, 2022 at 8:25 PM Richard Laager <rlaager at> wrote:
> >
> > [Off-list because I’m replying from my phone which doesn’t have the appropriate address setup. It’s fine to quote this publicly.]
> >
> > as the service name seems fine.
> >
> > If it were me… I think it’d be worth some small amount of time (like 2 or 4 hours) to try to integrate the auth with hub. If it’s trivial, why not. If you investigate and you’re halfway there, then continue a little longer. If you investigate and it’s going to be a mess, then bail and use built in auth. But ultimately it is up to you, as you are the one doing it.
> That was kind of my thought, the questionable part is importing the
> user list from mailman then means we're bringing people into more..
> but I dunno, the more i think about it, the more i'm thinking the
> username import might not be worth it.

I've got discord running on, the node's
hostname is carrier and can be reached via The root
account has it has ssh public keys for both Richard and I.  Please do
not do anything with this yet. I started looking at the SSO and it
requires a little bit of work and some more reading on my part. I also
messed up the email settings initially so I will most likely just nuke
the server and start over with the right settings this time.

That said, how much more should I be doing before we agree to a
statement of work? Last time we did something like this, I ended up
doing all the work without a statement of work and everyone was
disappointed and we're very quickly falling down that path again. I'm
around 4 hours in and we have a rough statement of work, but we
haven't yet defined a rate or min/max expenditure so I'm hesitant to
continue moving until these are defined.

Some additional information that may help with this. My normal hourly
rate for contracting is 100 USD per hour. I'm currently 4 hours in,
not counting time spent on these emails, that's all research and doing
stuff. As I mentioned previously I will most likely wipe the machine
and start over to fix my previous mistakes as we're using a 1-click
setup on DigitalOcean that makes that trivial. I still need to do some
additional getting up to speed on Discourse but that's going to be
ongoing regardless so I don't want to block too much on that either.
But that leaves the following list of items.

Not sure where these two should go, so I'm just listing them first but
I mention later where they're required.
* Richard sets up email accounts at wiktel (as much as possible at this point)
* Add wiktel to the spf record and any other email based dns records
that we can.

* Decrease ttl on and so I
can recycle the vps and get a new one without my mistakes.
* Recycle the vps to fix my email mistake that ended up in the "admin"
account being used and forcing the shared administrator account to
have a username of "administrator". (The idea here is that the admin
account is accessible by any officer or anyone that has access to the
1password imfreedom vault can get into it if necessary, but each
officer's account will also be an admin.)
* Get
up and running for SSO with
* In Discourse, create categories support, discussion, infrastructure,
and maybe plugins?. Add tags for all of the projects (pidgin, finch,
libpurple, libgnt, talkatu, gplugin, traversity, plugin pack, any
others like adium?). Not sure if we want a development category? I'd
really rather keep that on youtrack for pidgin, but maybe there's a
compelling argument otherwise?
* Add some sticky threads in each category mentioning their use and so on.
* Create and dns
records that point to the kubernetes cluster.
* Create tar.zstd archives of the raw mailing lists on rock and upload
them to
* Create containers for and using the previously created tar.zstd
archive and get them deployed to the kubernetes cluster.
* Put mail man into read only mode (we should pick a date for this and
announce it).
* update tar.zstd archives for both sets of mailing lists and rebuild
the containers
* decrease ttl for and

* wiktel email needs to be set up for everything after this.

* Point hub, youtrack, teamcity, reviewboard, and discourse at the
wiktel email server.
* update dns records to make wiktel the primary mx and to remove from the list.
* update dns record for to point to the kubernetes
cluster and put the ttl back to normal.
* update the ingress for to instead use (we'll leave listarchives but have it redirect to the
original domain)
* update dns record for to point to the kubernetes
cluster and put the ttl back to normal.
* update the ingress for to instead use (we'll leave listarchives but have it redirect to
the original domain)
* shutdown Back up its disk image to a safe place,
preferably the pidgin/imfreedom vault that it looks like i haven't
published yet (i can discuss this project in more detail later,
basically it's a ark where I keep a copy in a fire safe and send
another copy to John).
* destroy the vps that's running at wiktel.

I'll be codifying all of this as an epic in the infrastructure project
in youtrack once it's agreed upon, but I can do that early if we

Also I'm sure I missed something here, because as you can tell it's a
lot of work, which also means I have no idea how long it's going to
take me to actually complete all of this as there's plenty of room for
error and things to be forgotten.

> > --
> > Richard


Gary Kramlich <grim at>

More information about the Board mailing list