Notes from #MetaBrainz Meeting 2023-06-12

[aerozol]

MetaBrainz Meeting 2023-06-05

Meeting start: IRC Logs for #metabrainz | MetaBrainz Chatlogs
Meeting wiki page: MetaBrainz Meeting - MusicBrainz Wiki

Agenda

• Reviews
• static.mb/LE
• Google passkeys

Last meeting: 2023-06-05
Next meeting: 2023-06-19

@reosarevok: It is Monday. That means you get to put up with me again

Reviews

@yvanzo

[mailed in] Last week I helped with zas-driven maintenance operations around Consul & RabbitMQ.
Reviewed some MBS PRs and worked on docker server configs.
Progressed in Weblate with outsidecontext, rdswift and reosarevok but ended up investigating CAA issues.
I won’t have anything to report on the 19th.

@aerozol

[mailed in] Winter’s setting in and I’ve slowed to a crawl around our poorly insulated house. Drinking lots of hot chocolates to try prepare for hibernation.
I did some comms stuff, tackled a bunch of smaller tickets, none of which are exciting enough to link to imo.

@rob

last week I reviewed some contracts and did the usual background emails.
and we have a new unicorn supporter, but not all details are ironed out yet, I’ll announce the company tomorrow.
I also worked with our accountant go once again conclude that quickbooks sucks.
and her fix made me shudder. if we engineered like that, the internet wouldn’t work. at all.
srsly.
I also worked on troi to generated recs for all users only to realize that it must be done in spark.
so lucifer replace the periodic-jams patch with 12 lines of SQL.
humbled!
over the weekend I worked on the lb-radio (was artist-radio) which was a fun change, but things are so busy right now, that I can only work on it in my spare time.
there were lots more things last week, but Imma shut up now.

@mr_monkey

Last week I continued working on the new LB recommendations page
Also pushed some PRs along an did some ticket triage
On the BookBrainz side I’ve been working on some old PRs that I really want to see merged
So lots of reviewing and working on finishing them up
And a little bit of GSoC discussions.
[added later] I will be mostly off this week and part of next week, and won’t be at the meeting next monday. It’s my wedding celebration on Saturday !

@akshaaatt

I reviewed lots of PRs, completed the realtime listening now on the android app and other misc stuff
We should be doing a final beta soon for next prod release
That’s pretty much it for now. Great stuff upcoming!

@lucifer

last week, i worked on adding the bulk-generation of playlists on spark, some more work on meb-oauth and some additions to the android app (login workaround and websockets for now playing).

@atj

last week I worked with zas on finishing up the consul migration
there were some issues due to unexpected behaviour of services with a single registered client
this meant a downtime-less migration wasn’t possible but we managed to get it done reasonably quickly and promptly
after that I spent time doing some ansible role maintenance and discussing the next project with zas, which will be migrating openresty from Docker to native packages on the gateway hosts
deployed and managed using ansible of course

@zas

last week was mainly about finishing the move to new gateways, implying the deployment of the new consul cluster, which required a small downtime, and thanks to the whole team and atj it was done is quite short time
then we discovered a misbehavior on rudi (one part of the new gateways), so I asked Hetzner support to have a look, and they replace RAM
I also ordered one machine for the future replacement of pink & floyd, this one should arrive soon (there was a delay of few days), and replace now unused herb
I also fixed a bunch of graphs, alerts, and bits related to the move to new gateways
yesterday we had a weird issue with keydb (sort of redis) causing problems with websites using Letsencrypt certs. I fixed it by clearing existing files on rudi, then resyncing from rex. But I upgraded keydb today since it was likely related to a bug in the app.
apart that, I worked on some big cleanup patches for upcoming Picard 2.9. We should release beta2 this week.
Also spend some time discussing options related to openresty/ansible with atj.
Plus usual stuff, supervision, MB edits.

@bitmap

last week I worked some more on testing the docker-postgres-cluster improvements to haproxy/pgbouncer containers
the goal is to have haproxy detect which node is the primary so we don’t have to swap consul service names during switchover, and make read-only queries easier to scale across all the nodes
would like to have that ready for the new DB server if possible
I also worked on simplifying how to update configuration inside these containers (since they don’t use consul-template)
other than that I was cleaning up some old databases on aretha/rakim and some unused files in docker-server-configs
submitted a PR for MBS-13120 [Users can probably sometimes modify edit notes older than a day] and did some code review
and restored/updated the test.mb database to recent dumps

@reosarevok

The main things from last week were a prod server release (which was kinda confusing at first because of some errors that turned out to be ssl issues from the old gateways and zas dealt with - thanks zas)
And more working on the JS tests
I also did a bunch of code review
And worked on adding what seemed like a small addition to the wikidata bot so it will set the “named as” qualifier in our links
The change wasn’t too large but it turned out the images weren’t building because the bot was based on a xenial ubuntu image, and postgres doesn’t even have a repo for that anymore
So I ended up updating that to a newer image with the help of atj
Hopefully that means it won’t break for a while longer

@CatQuest

Some more instrument images
With that I also found an instrument staticbrainz bug which reo/bitmap fixed
(also did aliases. it’sgoing slow now becasue it’s summer. more on that later)
Went to a Concert!!! it was [this musicbrainz event entity]
(badly added, but oh well)
summer here so I’m going to be faffing about on the beach and garden parties and such

@snaekboi

Last week I got started with the MusicBrainz-Bot
did a bunch of changes with respect to reo’s suggestions
also added some “docs” (if docstrings count as docs)
Now I am currently working on some new features like updating existing areas using the MusicBrainz-bot : D
If lord see fit the project shall be completed on time

@jasjeet

last week was slow but still productive
oh mai gawhh
did some work on search user
fixed some bugs
also did work on direct submission from Brainzplayer
its still in work
and i wish i were a bird

@kellnerd

Currently I’m very busy with a university project which should be finished by the end of the week
As you can imagine that left little time for GSoC, so I’ve only read docs and reviewed shivam_'s progresss but didn’t wrote any code myself…
Starting with next week I’ll have to catch up; good that I’ve included buffer weeks in my GSoC timeline.
Oh and I discussed my alternative multi-source release seeder implementation (aka a-tisket 2.0) with atj, so now you know the mysterious MB project which I had mentioned a few weeks ago.
It’s already useful but a few things are yet to be implemented before I’ll release it, but as I said there are other prios for the next weeks : )

@ShivamAwasthi

Last week I worked on creating the Admin Panel Search Page which will allow admins to search for users and then give them privileges
I am almost done with this bit of functionality, some code cleanup left before I create a chonky PR for feedback

@arsh331

Last week I worked on my interactive atist graph GSOC project and have a demo running for the same.
Most of my time was spent building and fixing the main network graph, it is in a lot better shape now
but still needs to be polished further to look pleasing.
Overall I am happy with my progress
and on track for my project

@vardansaini

I am working with Apple Music player
trying to integrate with brain player

static.mb/LE

@zas

We are using a LetsEncrypt certificate for this rather critical domain, staticbrainz.org
And our setup is rather fragile atm regarding this matter
So we have different options, one is of course to make things tougher so we don’t have any issue with LE certs, but that’s not that easy as it involves many moving pieces
So the simplest option is buying a certificate for this domain
the second it to move to a domain we already have a paid cert for
static.metabrainz.org was my example (I configured it and checked the cert we have work, it does)
the question arises because we avoided a lot of issues last years by using hard certificates on critical websites
and this one became critical now (as many websites depend on it)
buying a new cert implies more time to manage/deploy/renew
using existing one, has no extra cost
strenghening LE stuff will require time and will never be as resistant as an hardcoded cert (or it will require a lot of resources, we’ll see)

Too much chatter to effectively summarize, please read the logs if you want to read all the input.

No objections are raised to:

1. Move to static.mb.org now
2. Plan to make LE robust so we can use it for all certs.

zas notes that we’ll keep staticbrainz.org, but it will redirect to static.meb, so no links will break until LE stuff breaks. But things will have to be changed internally.

Google passkeys

@rob

Google has enabled passkeys for our google accounts.
which are like SSH keys, but for logins.
I haven’t been able to make it work, but if anyone does, please let me know.

absent: pixelpenguin (1), riksucks (1), jivte (3, removed from the regulars list)

[UltimateRiff]

congrats @mr_monkey!~ wishing the best to you and soon-to-be @mrs_monkey~

@kellnerd

a-tisket 2.0

ooh~ I like the sound of that~ I can’t wait!~

…but I guess I have to, because that’s how time works…

@arsh331

Last week I worked on my interactive atist graph GSOC project and have a demo running for the same.

oh, I hadn’t seen this project before… looks real exciting~ perhaps it could be added to the artist pages I’ve seen are in the works?