#metabrainz

ruaok, bitmap: How do we update (old) search server/indexer FTP downloads?

we copy the jar/war files out to where the FTP dirs live.

I forget which machine those live on. zas?

(this happens on hetzner and they get synced from there)

ruaok: on williams

are we having a meeting today

yep

ok

ruaok: /home/ftp/ftp/root

yvanzo: ^^

thanks zas

iliekcomputers: we have 2 functions in artist.py fetch_unclustered_gids_for_artist_mbids_using_recording_artist_join and fetch_unclustered_gids_for_artist_credit_mbids. So, using_x is kind of required. similar for other places. So, removing using_x doesnot seem like an easy task 😂. I renamed get_artist_gids_from_recording_using_mbids_and_recording_artist_join

yvanzo: I'm about the leave the office -- got something for me to run?

ruaok: yes, go for it

ruaok: use the master branch of musicbrainz-vm

I did not update search server and indexer yet.

so this is just another test then?

yes

Use sample-data if you prefer a short test

kartikeyaSh: yikes

😂

bitmap, samj1912: Why is mmd-schema/production is one year old? Is mmd-schema/master still compatible with old search server?

the production branch should be what the search server uses, master is bleeding edge and may contain schema changes (but I wouldn't expect the old search to be broken with it)

if the new search isn't using the prod branch that should be fixed, but samj1912 would know better

It uses master

But I don't think we have any breaking changes

I don't see a prod branch?

Oh, nvm

It wasn't there on the overview

The only difference in the commit that solr uses is updated plugins

We don't use the genre updates

This is the commit it uses https://github.com/metabrainz/mmd-schema/commit...

<BANG>

s/Friday/Monday/

People up: Freso, reosarevok, bitmap, zas, ruaok, yvanzo, samj1912, iliekcomputers, rsh7, bukwurm, Leo_Verto, chhavi_, kartikeyaSh – anyone else, please let me know ASAP.

Leo_Verto: go!

Leo_Verto says:

"""

as mentioned in my GSoC proposal, I won't be around for the meeting this and next week as I'm volunteering at a sort of summer camp but here's my review for last week:

This week I set up Celery as a task queue for SpamBrainz (thanks for the suggestion, samj1912!) as well as an API token system and then started writing tests for both.

After moving things around and renaming them way too much I'm finally happy with this setup and should be able to pretty much plug the machine learning and report writing code into it.

I've also spent some time working with docker and feel like I've got a much better grasp on it which helped quite a bit with making the config system ready for containers.

"""

chhavi_: Go!

Or… samj1912, go!

Hi

Or… iliekcomputers go!

Hi again!

:p

:)

Hey

This week the biggest (or most expensive) bug I have written was uncovered.

😅

I guess ruaok will have more to say on that.

o/

Anyways, I've been dealing with the aftermath

Whoops.

Fixing logging and reporting more errors to sentry to make sure stuff like this doesn't happen.

Other than that, the usual PR reviews for kartikeyaSh and rsh7[m]

GSoC second evals start this week and rsh7[m] has a good PR up that im gonna get to soon.

Fin.

o/

chhavi_: go

Hey

Helo :) Managed to be productive over the weekend make some components. Pinged people for reviews on mockups. Fin!

samj1912 go

Hey, mostly spent the week relaxing, reading up on the new tasks and background solr management

Mostly enjoying life in Bangalore :p

I'll be tackling PG load balancing next

Solr has had a few improvement requests over this last week. Am going through them and fixing the ones that are high priority.

Apart from that typical week

bitmap: go

hi

I finished some prs for converting the event & series search results to react

fixed an i18n bug related to the search form options being mistranslated

worked on writing tests for the new i18n string parsing/interpolation code

fixed a couple strings (instrument/link type descriptions) in the DB that had broken html and couldn't be parsed

ah, and created a gdoc to compare some pg load balancing options

fin...go Freso!

Last week, I spent a lot of time going through reported (spam) editors and compiling them into a neat list for yvanzo's (and possibly Leo_Verto's) consumption.

A few more to go to clear my inbox of those, but far most of them are there.

Other than that, mostly being around and about and handling reported users and content.

fin!

rrrrrrrreosarevok, go!

Orr…

Hi!

Aha.

This week I did support and focused on looking into https://community.metabrainz.org/t/community-cl... :)

Luckily, it seems a bit more active than last week, and with loujine's new help (at the end of that thread) it should get a bit better still :)

Plan for next week is mostly more of the same

I guess CatQuest is on holiday, so I need someone new to give it to! :D

zas: go?

Hey

(People still up: ruaok, yvanzo, rsh7, bukwurm, kartikeyaSh – anyone else, please let me know ASAP.)

I did a lot of cleanup in SOLR/Zookeeper installation docs, and splitted scripts

Also a lot of work on matching graphs, we have a nice dashboard now, with almost all needed infos

I did a bunch of system upgrades too

A short point about how SOLR is performing atm

it works well, we don't have any error, but response times degrade quite quickly during peak hours, we are cpu-bound atm

samj1912 and i are investigating ways to improve the situation (without spawning more cpus)

the whole system is very stable til now, and handle single node downtime well. The main issue is that a SOLR node down during peak hours will cause some slowdowns

i'll experiment moving zookeeper on non-SOLR nodes this week

but overall, we can say "it works!" ;)

🙌

phew

FYI, i'll be travelling on friday (to Malta)

!m zas and samj1912

You're doing good work, zas and samj1912!

fin. ruaok ?

k

last week was paying people, accounting (mostly transferwise accounting setup and writing data import scripts)

I helped build VMs for yvanzo and worked on the conflict resolution protocol.

more PRs on LB and MSB, on which I am behind yet again.

then on friday I ran off to Nowhere for a short 36 hour stint to make sure my world isn't too calm and nice.

re PRs they were over the weekend so you can't be blamed :)

saturday evening got a wee bit strange, chaotic, sad, frustrating and just WTF.

never in my years or going to burning events has it gotten that far chaotic.

it was great. :)

oh yeah, and I had to do some cleanup on some monster BQ bill. lol.

yvanzo: go!

hi!

I worked on updating MusicBrainz VM mainly, opened a Vagrant Cloud account for further releases.

Also reviewed some PRs (thanks legoktm), and worked on usual React/Spam stuff.

FYI, I won’t be available on Friday either.

Finito, go rsh7?

hey!

(Only bukwurm / bukwurm_ and kartikeyaSh left on my list. Last chance for anyone else to let me know they want to give review too!)

I spent the starting days of the week working on iliekcomputers' reviews on #280 and #281, data insertion in musicbrainz schema was taking a lot of time initially which I have fixed now and it's pretty fast now.

Then, I worked on keeping musicbrainz schema updated with the actual MB db and used replication packets for that and I have opened a PR for the same (AB-350).

AB-350: Update musicbrainz schema in AB whenever there is any update in actual MusicBrainz database https://tickets.metabrainz.org/browse/AB-350

!m rsh7

You're doing good work, rsh7!

This all was the major part of the project. We would next work on testing both the methods and then to decide which would be better for implementing features in AB.

thanks iliekcomputers :D