#metabrainz

[musicbrainz-server] 14mwiencek merged pull request #2937 (03schema-change-2023-q2…mbs-12107): MBS-12107: Enable replication with dbmirror2 by default https://github.com/metabrainz/musicbrainz-serve...

mayhem: If we’re not updating reports then shall we remove this link from the menu? It makes MetaBrainz looks like it’s not being maintained if it happens to be one of the first pages you visit: https://metabrainz.org/reports/2017

aerozol: that apiseeds lyrics question is odd. They say they registered an API key on apiseeds.com, but this has not been available since 2 years, the Picard plugin also got removed 2 years ago. 🤔

outsidecontext: Yeah I clicked to that after looking at the Google translate results a bit closer. Maybe someone wants to tell them they’re out of luck in French :D

There was happi.dev as the successor, and even that shut down earlier this year

Yeah, lyrics sites aren’t worth putting heaps of effort into supporting imo, unless they’re some ma$$ive outfit like Genius

aerozol: indeed. there have been several lyrics plugins come and go. everything with an open API will have a short time span :)

I reactivated my old reddit account and answered in English

<3

Hey reosarevok would the codebase for the bot be on MetaBrainz's GitHub profile, or my own GitHub profile? Just confirming

Probably MeB, I can open a repo at some point

Then you can work on your fork and send PRs against the MeB one as needed

reosarevok: Do we have a media at metabrainz email? Just making a generic comms/marketing form to fill out, and need a default email. I know you check the community manager email but not sure if that’s quite right

mayhem: hi! I had filled the Soundcloud form to change redirect uri about 8-9 days ago but haven't received any response. should we reach out to the person we had a meet with the other day? also, can we request him to make one more application for us that we can deploy on test.lb for testing.

About one hour of downtime is expected starting from 5pm UTC to upgrade to the MusicBrainz database to its latest schema.

See announced changes: https://blog.metabrainz.org/2023/03/15/schema-c...

And countdown to downtime: https://www.timeanddate.com/countdown/retiremen...

mayhem: busy week last week, so put my foot on the gas today re. making noise

Could you case your eyes over the following docs for me, and I’ll review tomorrow, before putting it all into actions with the datasets news

Marketing/comms form, hopefully easy for anybody to fill out (will also do a fillable PDF version): https://docs.google.com/document/d/1Jj40bPIpy9r...

Press Release template: https://docs.google.com/document/d/1bBjhfwtI31f...

aerozol: can you put your foot on the gas and then circle back to confirm we have synergy?

mayhem: the previous two + a media contact/email list: https://docs.google.com/document/d/1ZvgcROSXTFu... (viewable by MeB emails only)

atj: Is synergy a joke to you???

i prefer to pivot and then touch base to make sure we're in lockstep

everyone: I’m putting together a list of news sites to send news to - where do you get your tech/open source/music information from, that might suit us? Especially smaller nerdy outfits. New-fangled podcasts and youtube channels, etc, qualify just fine.

mayhem: maybe those docs are a good start re. stuff for Rassami to look at and feed back on or add to

okay, bed time for me. Hope the schema change goes well all!

Good night, same hope!

[metabrainz.org] 14mayhem opened pull request #412 (03master…remove-annual-reports-link): Remove the reports link https://github.com/metabrainz/metabrainz.org/pu...

aerozol: done ^^ (removing the reports link)

lucifer: go ahead and contact jordan directly, no problem.

aerozol: will have a look at the docs today.

[musicbrainz-server] 14yvanzo opened pull request #2939 (03master…mbs-13083-fix-comment): MBS-13083: Fix code comment https://github.com/metabrainz/musicbrainz-serve...

reosarevok: yvanzo: hey, I finished running upgrade.sh against a clone of the production DB (musicbrainz_schema_change_2023_q2 on aretha) last night, and didn't have any issues

👍

<3

also, I was checking the consistency of the first-release-date data

I set up a mirror using Docker Compose to be able to test upgrade after main server has been upgraded in proudciton.

[listenbrainz-server] 14MonkeyDo merged pull request #2485 (03master…LB-1274): LB-1274: Small usability improvements of the MBID mapping modal https://github.com/metabrainz/listenbrainz-serv...

release_first_release_date is identical to a query that computes it from scratch, which is good

(very proudly)

I did notice that some first-release-date data in release_group_meta and recording_first_release_date is wrong/inconsistent, but then checked the production DB, and those problems already exist there. so it appears to be unrelated to withdrawn releases or the upgrade script

and I think I know why the recording ones are wrong, but it'll have to be a future schema change

Oh, it's that bad? :/

well, I'm pretty sure we're missing a trigger for medium updates - when releases are merged using the append strategy, some medium.release columns are updated, which affects all of those mediums' recordings' dates

I see

Fortunately this can be changed in production ahead of the next schema change release since it doesn't require any change on mirror side.

bitmap: Why is it recommended to restart Redis for the schema change (last year at least)?

I guess it was related to some change but couldn't find it. Not sure it is still needed this year.

I'm guessing just in case there are any code changes that affect cached entities (like changing the Moose entity models)

right, it may not be needed this year

Do mirrors get the date info from us or do they calculate it with their own triggers? I forget

Good question, not sure.

release_group_meta is replicated, but the other two materialized tables (release_first_release_date, recording_first_release_date) are not and are updated by triggers on mirrors.

we could rebuild recording_first_release_date as part of our upgrade steps and recommend mirrors do too. though it will lengthen upgrade process by a little bit and won't fix the underlying issue, just existing data.

Yeah, that's what I was thinking

How hard is that fix?

(as in, is there any chance we can offer a way to fix it for schema change users if they want to fix it as part of this SC even though it was not announced, or is it hard and we should wait with it anyway?)

well, I don't think writing the trigger will be hard, but would rather wait to test it properly. we could always mention the upgrade script in future MBS/docker release notes

I would just wait with it, we can probably come with a workaround to fix existing first release dates before the next schema change.

I'm sure we can, it's just they'd need to run it repeatedly :)

But I guess there's no reason we can't say that we'll publish an updated trigger later

And they can install it outside the schema change themselves

Now properly behind my screen too

yvanzo: https://github.com/metabrainz/docker-server-con... why should JSON dumps not be updated?

they use a separate mirror DB which has to be upgraded separately

Oh, ok

I forgot that part, sorry

(test.mb too, though that's a standalone DB)

bitmap: did you stop json-dump container already?

yup, just crossed that one

I also started the pg_dump of musicbrainz_db on aretha just now.

Ok

do we plan to merge https://github.com/metabrainz/musicbrainz-serve... today?

reosarevok ^

rn we're at the point where we should be merging the schema change branch into master soon

I'd like to because I feel any issues are minor enough to be resolved later, but I'm checking one last bit yvanzo brought up

(the "Cancel" button also fails to submit if a required field is empty)

reosarevok: Are you testing a fix or are you looking for a fix?

has_field 'cancel' => ( type => 'Submit' ); is the same as in many other forms, but I guess this is the first one with an actual required field, heh

Looking for, but this is a very minor issue too - so I think it'd be fine to amend it on next release tbh

But as you two feel :)

lucifer: hi, are there any LB cron jobs using the MB database? they can be stopped whenever you're able

I'm also ok with putting the SQL out today to prod, and the implementation later today in beta

Or tomorrow

That might even be better since I bet users will find some more small issues :)

yvanzo: does that sound like a good idea to you?

reosarevok: I agree, I’ve edited the PR description to update the condition #4 and to add the condition #5.

BTW, does monkey have to stop something for BB?

We can even add more small patches to beta tomorrow, and have a release next week.

lucifer has

Sounds good to me. bitmap?

I think BB should be fine, apart from logging in

the beta release sounds like a good idea to me

But I will check once you start the procedure and take it down if anything doesn't work

Ok, let's do that then :)

Then we should be good to merge the schema change branch into master (since you said you cherry-picked the SQL into it)

I'll merge schema-change-2023-q2 into master then

Thank you both

bitmap: You cherry-picked it *after* I changed it to change_editor as you asked?

yeah, it's all good

Ok :)

Yay

double-checked