Notes from #MetaBrainz Meeting 2024-11-18

[aerozol]

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

Agenda

• Reviews
• proxmox

Last meeting: 2024-11-11
Next meeting: 2024-11-25

@reosarevok: Welcome to another MetaBrainz Monday Meeting!

Reviews

@derat

[mailed in] I updated the yambs web interface to support seeding edits for all of a release’s recordings [forum link]. Nothing else to report, I don’t think.2.

@akshaaatt

[mailed in] Last week has been quite a lot for me. Life getting hold of me with my sister’s engagement coming up and work in general. I made some progress to the apps, but not much really! Looking forward to pacing things this week!

@Ashutosh_Aswal

[yellowhatpro] I couldn’t get much time last week. Got free this weekend, and worked on the JsonValue to structs refactor for extracting urls from edit types.
Came up with a draft PR (working on tests currently)
also checking to add more docs to the project, I feel contributing section can be better, so looking forward to work on that

@yvanzo

Last week, I investigated 3 Weblate issues, 2 of which are due to MB Server, and I reported the 3rd one upstream.
Otherwise kept working on Solr 9 data dumps and upgrades.
Looking forward to make it happen this week with atj, bitmap, reosarevok and zas

@julian45

Last week I learned a bunch from zas about the internals of MeB infra and discussed a few things we might try in regards to server hosting and security (e.g., bastion server)
One such topic was Proxmox, which is later in the agenda for today’s meeting

@rob

[mayhem] last week I doubled down (with monkey) on making LB much more stable.
I started picking off bugs and adding service level monitoring for LB.
opened a bunch of PRs and with zas’s help we’ll create a new Telegram group for LB services.
that way, when some service (dumps, stats, incoming listens or recs) is late, we know about it before our users report it to us. was way overdue.
we still have some more open data bugs, but in general LB is running 100% today.
I also helped welcome Julian to the team and did the usual background MeB tasks, including prepping for a board meeting which will happen right after this meeting.

@aerozol

I feel like I keep not getting ahead of things, and stuff keeps coming up… so once again, no major design work done
As always, staying on top of socials, tickets, and so on
Adding a few events to the statistics timeline
And the MB team pulls me in regularly to do reviews of PR’s with UI elements, which is always nice!
And caught up on adding the backlog of cover and event art, in the evenings…

@anshgoyal31

[ansh] Last Week, I worked with monkey on getting the donation flairs (almost) ready.
I worked on CritiqueBrainz to complete one of my PR CB#401 to add endpoints to fetch reviews which was required by MusicBrainz.
[CB-427: Support entity endpoint in the API including average rating]
Fixed some issues I encountered while navigating through LB, and worked on some tickets.

@mr_monkey

[monkey] As mentioned, last week I worked with ansh on integrating the new css flairs for donors. It’s looking pretty cool!
Also worked with mayhem on LB infrastructure stability, which included bugging zas for help in debugging a consul-template/uwsgi issue
And I started reviewing kellnerd’s remaining PRs for his gsoc project
And more PR review

@bitmap

last week I restored image uploads on the beta server via reosarevok’s PR, after regenerating all passwords/API keys and re-queueing previous failed artwork-indexer events
we saw some reports about errors with some images, but they seem to be loading now
after the meeting I plan to do the production release which will re-enable uploads on the main site
I also worked with Jade to test and deploy MBS-13815, so messages to editors now use the new mb-mail-service on beta
[MBS-13815: Allow sending contact emails through new mail service]
other than that I mostly did code review and debugged some rabbitmq isues in the github actions branch

@zas

Last week, I spent some time with Julian, for an initial tour of our infrastructure
I also had a chat with Monkey, we debugged and fixed an issue with LB containers related to process handling
I also reviewed a bunch of Picard / Picard Plugins PRs
And I did the usual maintenance, upgrades, minor fixes, general supervision, etc
Also, worked with mayhem on LB status
And finally I uploaded ~160 cover art images on beta for releases I added recently…

@outsidecontext

not so much again. Did code review, handled new tickets and some forum support.

@lucifer

Last week I attended ISMIR in San Francisco and presented ListenBrainz at the conference. Met lots of new and interesting people. Also a couple of new projects with similar goals to LB.

@reosarevok

I spent most of my week working towards expanding derat 's code to show primary aliases for artists so that it works for all entities and in all pages
Turns out there’s a ton of places where changes need to be made and a lot to test, and I’m sure when we think it’s ready and it goes into beta someone will find five places I missed
But it’s progressing and I’m quite happy about it, maybe we’ll get less people adding duplicates and generally being confused
I also spent quite a bit of time working on the revived code for a better review UX, and even annoyed ansh into doing some CB work related to it
That’ll take a bit longer too, but progressed quite a bit there as well

@kellnerd

Last week I did a bit of maintenance for some of my smaller projects and got sucked deeper into MB editing after a long pause.
I am back to university (for research, no more lectures), so I will have less time available from now on.
Over the weekend I wanted to look into the OpenLibrary importer for BB, but got distracted (will get back to that tomorrow).
Instead I finally made the region input in Harmony persistent and made the region and the preferred providers available to the server as cookies.
This should be really useful for external tools which are linking to Harmony, they no longer have to hard-code these options into the URL. (I will post details in the forums later.)

@ApeKattQuest_MonkeyPython

[ApeKattQuest] Slowly working on the ektars!
Other things have taken my time this week, but RL hinders are now coming to a close!
I’ve also started translating again, because we figured out it wasa bug in the beta-ticket!
that prevented the login on weblate. very happy to be able to login again!

@jasjeet

[Jasje] Slowly working on the ektars!
Other things have taken my time this week, but RL hinders are now coming to a close!
I’ve also started translating again, because we figured out it wasa bug in the beta-ticket!
that prevented the login on weblate. very happy to be able to login again!

proxmox

@Julian45

As zas and I were discussing MeB infra, I asked where there might be any interest in testing Proxmox VE
[website], which is essentially a quite powerful FOSS version of [insert fancy hypervisor here] (VMware, Citrix, etc.)
Essentially, this would give MeB a chance to use its existing physical hardware as a cluster for hosting virtual machines, which would allow for a lot more manageability, safety (e.g., backups), and reliability (e.g., high availability in case one node in a cluster goes down)

@zas

(basically it could help with few historical SPoFs)

@Julian45

We would like to stand up a Proxmox testing environment on which to put non-critical workloads and evaluate Proxmox’s claims of stability and HA
[others ask what PoFs in particular/for an example use case]
For example, say we run a number of MeB services as VMs on a cluster of a few physical nodes. One day, a physical node craps itself out. Proxmox would allow workloads hosted on the failing node to automatically be migrated to the remaining nodes

@zas

Almost anything we run atm could run on a VM on proxmox, but perhaps high performance DB servers. Since we’ll use our own hardware, if a single node goes down, it would have no impact on the services (ie. redis)

@rob

[mayhem] Almost anything we run atm could run on a VM on proxmox, but perhaps high performance DB servers. Since we’ll use our own hardware, if a single node goes down, it would have no impact on the services (ie. redis)

@Julian45

Possibly. It can run Docker containers itself (I tend not to do this much in other situations due to lack of use case); that being said, given a way to expose those volumes from the existing hosts, I don’t see why the contents couldn’t be migrated into the appropriate place (e.g., within a new VM’s filesystem)

@rob

[mayhem] ok. that is one of those uses cases I would like to see.
moving some services is a real pain right now because of this.

@zas

The idea of the experiment is to evaluate it for our use. We think about a minimal cluster of 3 machines (we have 2 about to be retired we could use)
If we get a green light for this, we can organize a dedicated meeting about proxmox for people interested

rob/mayhem has to go, so the green light will be given another day.

absent: atj (1)

[rhetticent]

@mr_monkey

[monkey] I feel like I keep not getting ahead of things, and stuff keeps coming up… so once again, no major design work done
As always, staying on top of socials, tickets, and so on
Adding a few events to the statistics timeline
And the MB team pulls me in regularly to do reviews of PR’s with UI elements, which is always nice!
And caught up on adding the backlog of cover and event art, in the evenings…

Looks like @mr_monkey’s update this week may have gotten mangled. For anyone that keeps track of meetings via Discourse, here’s how I (think ) it’s supposed to read:

[monkey] As mentioned, last week I worked with ansh on integrating the new css flairs for donors. It’s looking pretty cool!
Also worked with mayhem on LB infrastructure stability, which included bugging zas for help in debugging a consul-template/uwsgi issue
And I started reviewing kellnerd’s remaining PRs for his gsoc project
And more PR review

[aerozol]

Thanks!! I’ve also fixed the OP