#metabrainz

makes sense, i'll add to the comment on that proposal.

we only have 2 proposals for LB for the time being so that's it on this topic.

thanks for jumping into this -- I hope to be out of accounting jail tomorrow.

half day should do it.

alastairp: you may also want to look at the proposal above. particularly the time zone storage stuff, i had assumed PG had a built in type for it but apparently there isn't one.

ah cool. nice

<akshaaatt> "Why would you need to share..." <- Actually half of them were just MBID examples embedded with the links haha.

I removed them, so it should be fine now ig.

lucifer: yeah, I have it open now. that's a good question

the answers i saw on SO, suggested text column plus a constraint check.

Great Prathamesh! All the best :)

<alastairp> "PrathameshG: you can make a..." <- Yes, I am doing both.

Including a full link for the google doc, as well as a not-as-good-looking proposal written in Markdown on discourse.

lucifer: where text is the name of an official tz database name?

yes.

right. I wonder how often new ones get added/removed there? :)

akshaaatt: Thanks akshaaatt :))

oh well there was a great controversy about that recently.

yes, I remember

lucifer: oh, urhg. you know what I realised

mostly about pre 1970 dates though so not concerning us much. iirc PG chose to track the fork or revert changes from offical repos

for when DST changes, we should probably store an offset with the listen, indicating what the offset was at the time that it was submitted

there was that interesting post on HN a few weeks back about how to store time in calendars for future events

do we need an offset for that? i was assuming knowing the timezone would be enough and a datetime lib could figure out the offset.

nope

oh!

because for example, CEST happens from the last sunday of march to the last sunday of october

but what if next year they decide to start it on the first sunday of april

alastairp: datetime lib should be able to handle that.

Using the tzdata.

hmm, yes, if it takes into account historical data then it should

yeah i was assuming tzdata tracks the offset for each zone over period

It does.

Freso: I have some obligations at :30 after the hour. I'd love to go sooner than later, plz.

lucifer: in any case, some careful things to think about with this proposal so that we don't screw it up again like we did with the listened at timestamp

I'll take a look at it

It’s International Monday for Mine Awareness and Assistance in Mine Action!

One review received, so… atj, go!

atj says…

"""

I was occupied with work and some IRL matters last week so was unfortunately unable to dedicate much time to MB.

However I did resolve a minor issue with shorewall due to a reference to an invalid log file and hopefully fixed an recurrent issue with apt tasks failing when bootstrapping servers. I also reviewed a PR created by zas that added a role to manage timezone and NTP configuration.

The playbooks are now mature enough that zas was able to fully deploy a server from scratch in 30 minutes today, which is a great achievement.

"""

Others up: alastairp, mayhem, akshaaatt, zas, yvanzo, lucifer, bitmap, monkey, reosarevok, CatQuest, Freso – anyone else who wish to give review, let me know ASAP!

mayhem: Go!

thx!

Np :)

great to see the playbooks going well!

alastairp: yes makese sense 👍

!m atj and zas

You're doing good work, atj and zas!

last week was all about the metadata view, or more specifically the MB datadata cache coming together.

massive SQL queries ftw.

lucifer and I also cleaned up some mapping stuff, fixing some things that were not mapped and also renaming things to be more consistent.

we now also have a python class for making it super simple to run an SQL query to create a new table and then to rotate that table into production inside a transaction.

those bulk tables can be chained as well, so more than one table can be swapped in in the same transaction.

this is the meat and potatoes of the mapping work, so this will make me more efficient in the future.

then I paid people and did quarterly accounting today.

fin, Freso, go!

🙋

It's up!

mayhem alastairp akshaaatt : Please check it out if you find some time :))

Feel free to criticize it wherever possible. I'd be happy to even rewrite it if required!

Sure PrathameshG[m]! Will do so after a while :)

will do tomorrow afternoon

Mostly the usual, but I also poked at an issue with logging in to the forums from the iOS… and got that solved. So anyone using the Discourse iOS app should now be able to use it with our instance, too. :)

fin.

alastairp: Go!

hello

Thanks a lot akshaaatt mayhem :)

last week I finished work on the metadata viewer backend to update out of date rows, and planned with mayhem how to actually find out of date rows. I also started work on data dumps for the canonical tables

I finally got the BU upgrade in CB finished and released and some people tested that. It's looking good so we'll merge and roll that out. with lucifer's help I also merged and released some really old CB PRs (from before when lucifer was a contractor!)

I also read and reviewed some SoC proposals, and reviewed some PRs

finally, I worked on getting AB dumps working, so that we can distribute an archive of submissions when we shut the site down. 2 of the 3 dump types are working, so almost finished there

that's me, akshaaatt next?

Hi everyone!

Last week I worked massively on the MusicBrainz revamp work

You can experience it at https://test.musicbrainz.org, thanks to bitmap and reosarevok!

Also, I have been working at full pace on the android app side by side.

That's it for me! fin

Go bitmap!

hey

last week I did a fair amount of code review and did more schema change work

mostly related to https://github.com/metabrainz/musicbrainz-serve... but also writing scripts for the dbmirror2 changes and performing other tests

the font is stoo huge :O

Yay, less broken db mess

I also opened a PR to improve our CAA error logging and deployed that to hopefully help debug an issue Zas was seeing

earlier in the week I helped perform another primary database switchover from pink to floyd so that pink could be upgraded, then restored standby service on pink later that day. and I apologize for not better communicating which services would be affected by that, I've noted it for the future!

fin. yvanzo go!

Hi!

Last week I hotfixed bugs in the wrapper that runs MB background tasks.

(Still up: zas, lucifer, monkey, reosarevok, CatQuest – anyone else who wish to give review, let me know ASAP!)

Helped with pink update, and investigated some alerts (CAA, SIR…).

Updated the documentation PR about Docker volumes.

Resumed updating the PR about releasing MBS.

Also reviewed a few PRs and provided support.

Finito, go zas!

Hey

Mainly worked on moving servers to ansible, last week we did pink, patton and ludwig

It implied fixing/improving our playbooks, but today's ludwig was up and running within 30 minutes from scratch (including formatting/installing from rescue)

!m zas and atj

You're doing good work, zas and atj!

I also continued on new gateways, fixed few issues, and started to benchmark (using vegeta), there's some bottlenecks related to SSL I need to identify

basically I achieved 30k+ req/s on http on a single server, but this goes down to 2.5k in https with keepalive, and only 530 without ka

(that's per second)

plus usual supervision, upgrades, edits, PR reviews. fin, lucifer?

hi all!

(Only monkey, reosarevok, and CatQuest left.)

i reviewed some LB PRs last week and SoC proposals. also did some on work on metadata viewer backend and related electron app. also some dumps debugging. that's it for me.

reosarevok: next?

Hi!

I released the server, worked on a few small issues, and took a couple chill days to rest and bird and cycle

Also spent some time reviewing akshaaatt's code :)

:bird:

I didn't know that birds could cycle

:D

cyclebirds

It's in the Lion King, dude

eeehht

It's the cyyyyycle of biiiirss

*birds

shouldn't thst be birbs

Oh, yeah, of birbs

Anyway, this week plan is to hopefully start working on moods

If I'm in the mood, that is

And some more helping akshaaatt probably :)

We should work on merging some of the open schema change PRs too

But anyway, that's it for me. CatQuest ?

Meow! `😺`

last week was spent mostly nagging people to help write a stupid blogpost

then reosarevok basically rewrote the whole thing :​(

but it ended up a success so that was a good thing! :​D

this week I've nagged monkey and shub but they are doing really good work

bai!

so I'll be working on testing bookbrainz and seeing if I can be at all useful :​) ♓

fin, go monkey !

(also ps happy aprilfish day!)

Hello everyone !

Last week I worked on UI mockups for the LB metadata viewer, with a bit of help from aerozol (thanks!)

I tweaked BB's automated GitHub deployment configuration a bit, and deployed a new version in production.

We’ve already identified a couple of regressions (unrelated to the deployment config), so not the best deployment ever…we’re discussing better deployment cycle practices on the #bookbrainz channel.

I also reviewed PRs, mostly for BB, and spent some time helping and directing new contributors, participating in forum discussions and reviewing GSoC project proposals.