#metabrainz

You're doing good work, MetaBrainz!

thanks

last week I created more ansible roles to add functionality requested by zas and also include some improvements from me (like adding SMART monitoring for drives)

zas and I re-installed floyd on Ubuntu 20.04, and the process went quite smoothly overall, which was a good improvement

I think we're getting more confident that this is the right direction, and will help to reduce the maintenance overhead going forward

there's still plenty of work to do, but it's positive so far

👏

that's about it, I just wanted to give digress from the usual flow (you can skip this if you want Freso) and give a few shoutouts

note that if I don't mention you, please don't take offence, it's purely some things I've noticed whilst being more involved over the recent few weeks

firstly to reosarevok for being so damn productive, I honestly don't understand how you manage it

secondly to akshaaatt for persevering with learning git and battling with GitHub (I know how frustrating it can be)

woah. you just called a spaniard productive? thats a first.

finally to bitmap for being such a wizard and just doing the failover from floyd without a sweat, then saying "I wasn't sure that would work" afterwards!

zas: you're up...

Hey

!m atj

You're doing good work, atj!

Apart usual supervision, upgrades, user support, mostly worked on floyd upgrade with atj

I agree with giving kudos to reo and bitmap especily. (but yea everyone is awesome)

!m zas too

You're doing good work, zas too!

I confirm what atj said, the move to Ansible looks good, we still have a lot of to do, but the gain is obvious

\o/

I ordered a new server to replace burnside (it should be ready this week, we got no news from hetzner yet)

once it will be deployed, we'll move yehudi containers to it, and retire yehudi

it will free a spot in our rack, that'll ease future moves

have to bail early sorry everyone, have an appointment. enjoy your morn/eve's!

I'll order a new server to replace herb very soon, we'll then start to work on new gateways

\o aerozol

plus usual Picard PRs, and edits, fin. CatQuest ?

Meow! `😺`

Have been updating recorders, pictures and adding aliases

(People still up: yvanzo, reosarevok, akshaaatt, alastairp, Freso – anyone else who want to give review, let me know ASAP!)

also working on testing bookbrainz-site and adding several tickets/sorting tickets

♒

fin, go reosarevok !

Hi!

I worked a lot on documenting and improving tests

And I helped akshaaatt with his PRs

Also worked quite a bit on changing slave to mirror in MBS

I'm sure there's more, but that's the main bits :)

yvanzo: you?

hi!

I reviewed a few MBS PRs and worked on the drafted schema change.

Updated and successfully tested a patch to address the recent failures of SIR.

Also helped with maintenance, fin.

Go akshaaatt!

Last week we made a lot of progress on the MB revamp stuff, thanks to reosarevok and bitmap!

(Only alastairp and myself (Freso) left for reviews. Last call for anyone else who may want to give review!)

The Design system integrated Figma to now allow us to have the designs(drag and drop designs from Figma) and code.

monkey and I also had some good discussions regarding the BrainzPlayer redesign and I also made some progress fixing the bootstrap updates in LB.

I also got to review some amazing work from yellowhatpro, who is working on cool fixes and additions for us on the MB app.

That's about it for me! fin :)

Go alastairp !

hi there

last week I continued working on some CB tasks, upgrading BU and making changes related to how we retrieve MB entities in it,

I reviewed some of Ansh's new PRs, and talked with monkey about CB/BU integration

I had a few planning discussions with monkey and mayhem and lucifer

CB/BB ?

yes, that one

that's it, Freso ?

🙋

Spent time on reports etc. I especially sunk a lot of time into unravelling one sockpuppet(?) web. Not so much a sockpuppet as someone who seems to use throwaway e-mail addresses to sign up and then makes a new account after a few days. :\

fin.

And that wraps up the reviews. Thank you all for yours!

And no more items on the agenda, so this wraps up today’s meeting too.

Thank you all for your time!

</BANG>

!m Freso

You're doing good work, Freso!

thanks everyone

thanks Freso

lucifer: around?

yup

I didn't realise earlier that pg_upgrade requires binaries for both versions. that's interesting

right so we could install on bono or build a docker image?

*install on gaga

bitmap: did you run https://github.com/metabrainz/docker-postgres-c... inside docker, or not?

I'm just looking, it seems that the timescale docker image is based on alpine: https://github.com/timescale/timescaledb-docker...

also we probably can't use `--link` option, the directory structure of the volume is different.

(line 38)

i see

alastairp: yeah inside docker

bitmap: ah, so you exec into a running 9.5 container then manually install v12?

right, I just installed both versions side by side inside the running container

lucifer: so maybe we could try and extend the timescale image with the new version of postgres?

although, I'd also start to consider doing this in multiple steps - can we upgrade timescale on the old posgres (still supported?)

and then in a second step, once we're on new timescale upgrade the pg version

yes that is the plan. do pg upgrade first then ts.

or that way around, sure

tell me the versions of each that we're upgrading?

pg11->pg14, ts 2.2.0->2.6.0

ts 2.4 is the last version supporting pg11. but it doesn't support pg14. 2.5.0 is the one that supports pg14.

so there's no timescale docker image for ts 2.2 + pg 14, or for ts 2.6 + pg 11

nope

do ts upgrades have to be done 1 at a time? looking at the dockerfile it seems like it copies in the "old version" shared libraries into the image

acc to my understanding, we can do it all in 1 go.

"This upgrades TimescaleDB to the latest installed version, even if you're several versions behind.", right

the pg11 - pg14 thing might be an issue. if we want to remain compatible at all points, we need to do 12 or 13 first.

ah, I was wrong - https://github.com/timescale/timescaledb-docker... I thought it only copied in the last ts version, but it copies in the last 5

so pg 11 -> ts 2.3 -> pg 13 -> ts 2.5 ?

and then we put off the pg 14 upgrade till next time?

wait 8 months and jump straight to 15 when it gets released

because pg11+ts2.3 is just a matter of starting a new image and then installing the updated extension, right?

or we can do pg 11 -> pg 13 -> ts 2.6. 1 step less

yes

yes we currently don't have any particular pg 14 feature requirement so even 13 should be fine.

oh right, 2.2 supports pg 13 too. I agree with you

although, one sec

OK, I just checked - the timescale/timescaledb:2.6.0-pg13 image does contain the 2.2.0 timescale .so

the alpine postgres images build postgres from source, that's annoying :(

we can write a small custom image, installing both versions?

I don't know enough about how pg works to know how much something like this might affect us: https://github.com/docker-library/postgres/issu...

we don't need to use the alpine version though?

it seems that timescale only comes in the alpine variant

but its a pg database so while upgrading pg we can skip using ts images, no?

so, maybe it'd work if we had a custom alpine image which had pg 11 and pg 13 installed (no timescale??!), do the upgrade, then start it again with 2.6.0-pg13 image?

yeah, we could also try and do the upgrade with a debian based image with both 11 and 13 installed

yeah i think debian based should be fine but we can confirm with the folks in ts slack if needed. also we'll test it on bono tomorrow

sure, if you could confirm with them then that'd be great

sure will do

I'm working tomorrow so can help with testing too

Updating test.LB

sounds good

interesting, there's no discussion on https://docs.timescale.com/timescaledb/latest/h... about if the versions of postgres that you're upgrading need the timescale .so files installed

lucifer: might be a good idea to ask them about that too

+1

[bookbrainz-site] 14dependabot[bot] opened pull request #794 (03master…dependabot/npm_and_yarn/compression-webpack-plugin-9.2.0): chore(deps-dev): bump compression-webpack-plugin from 8.0.1 to 9.2.0 https://github.com/metabrainz/bookbrainz-site/p...

[bookbrainz-site] 14dependabot[bot] opened pull request #795 (03master…dependabot/npm_and_yarn/babel/runtime-7.17.2): chore(deps): bump @babel/runtime from 7.16.3 to 7.17.2 https://github.com/metabrainz/bookbrainz-site/p...

[bookbrainz-site] 14dependabot[bot] opened pull request #796 (03master…dependabot/npm_and_yarn/babel/plugin-transform-runtime-7.17.0): chore(deps-dev): bump @babel/plugin-transform-runtime from 7.16.4 to 7.17.0 https://github.com/metabrainz/bookbrainz-site/p...

[bookbrainz-site] 14dependabot[bot] opened pull request #797 (03master…dependabot/npm_and_yarn/webpack-dev-middleware-5.3.1): chore(deps-dev): bump webpack-dev-middleware from 5.3.0 to 5.3.1 https://github.com/metabrainz/bookbrainz-site/p...

[bookbrainz-site] 14dependabot[bot] opened pull request #798 (03master…dependabot/npm_and_yarn/babel/cli-7.17.6): chore(deps-dev): bump @babel/cli from 7.16.0 to 7.17.6 https://github.com/metabrainz/bookbrainz-site/p...

lucifer: ok, thanks. let's talk tomorrow!

thanks you too!

interesting. it's kind of like our develop.sh

bitmap: do you know if collation parameters in postgres (e.g. LANG) result in data being written to disk differently? (I'm thinking things like the on-disk format of a index)

alastairp: Are there any changes required in the Remove Gravatar PR?

although I would expect that the collation data would be stored alongside it, so pg would know how to read the data in any case?

Ansh: I've just finished work for the day, but I'll look at these tomorrow. The Remove Gravatar one looks good

* Gravatar PR? https://github.com/metabrainz/critiquebrainz/pu...

Okay! Also does the /artist endpoint look good to you?