#metabrainz

(People still up: atj, skelly37, yuzie – anyone else who wish to give review, let me know ASAP!)

Nearly completed diagrams, continued writing docs, looked into MBS dependencies.

Also reviewed a few PRs, and made a bit of maintenance of Jira/Tickets.

Fin. Go atj?

Or skelly37?

Or yuzie?

hello

atj, skelly37, yuzie: Any of you around? :D

Ayy.

Go skelly! :)

Last week I've been improving the single instance mode for Picard. Very annoying bug got removed, URLs support is almost ready. Passing commands to the existing instance was discussed

quite a productive week

Also I was trying to move imp -> importlib but got stuck for now

When it comes to personal life: well,finally got my first job (with some help of the gsoc experience, thanks y'all!) :D

fin, go atj

\o/

or yuzie

Hi!

Last week I add backend API for resetting user timezone and also modified the corresponding frontend UI a bit.

But I’m stuck with problem related to CORS policy.

That’s for me.

I guess atj is not around, so I’ll wrap up after yuzie. If atj or anyone else are around and want to give review, let me know now!

Well, let’s wrap up then.

Thank you all for your reviews, and thank you all all for your time!

Stay hydrated, masked, and take care of yourselves!

</BANG>

yuzie: the endpoint seems to work fine for me. https://usercontent.irccloud-cdn.com/file/60C74...

can you share what you see in network and console logs?

Thanks

Sure. Please wait a minute.

thanks everyone

Just a reminder for all the GSoC folks that midterm evaluation should be open in 15 minutes :)

Pratha-Fish: hi, do you want to get LB running?

alastairp: Sure, if you don't mind

I hope it isn't too late for you tho

nah, it's fine

Alright, then let's get on to it :)

OK, so the idea is to clone the LB repo, configure it so that you don't run over anyone else's version on the same server, and then connect it to the main MB mirror database so that you can run the necessary apps

lucifer: still around? got a moment

yup

lucifer: https://sentry.metabrainz.org/organizations/met...

I looked into this, it was the same thing we ended up with a few weeks back - the service was trying to restart but failed to bind() to the port, causing an error and a restart

there are a _lot_ of error messages for it. I stopped and restarted the container

I can't remember what container we had this issue on last time. but it seems like our restarter script isn't sleeping between executes

Pratha-Fish: ok, go ahead and clone LB into wolf (I'm following along to make sure we make the correct changes)

i think it was the labs api container last time as wel

cloning

lucifer: Please see https://drive.google.com/file/d/1E9ojyAVTuKb-HC...

alastairp: Are there any changes required in CB#441?

CB-437: Add entity metadata to review get endpoints: https://github.com/metabrainz/critiquebrainz/pu...

yuzie: you need to API_URL in your config.py.

*need to set API_URL to localhost:8100 in your config.py

alastairp: done cloning

OH, I see. I set it to API_URL = 'https://api.listenbrainz.org'

ansh: I don't think so. I just wanted to take another look at the database query to understand what information the endpoint is returning

Pratha-Fish: great. first change - edit devlop.sh, line 13 in function invoke_docker_compose, change the argument after -p to `listenbrainz_prathfish` or something similar

doesn't matter what it is, just something different to the normal. your login name, or your name, or something

yuzie: setting it to api.listenbrainz.org will send the api request to LB production servers where this endpoint doesn't exist thus giving you a 404 error

when you change it to localhost:8100, it will go to your local container.

alastairp: done

bitmap, yvanzo: not sure there's enough ready to merge for a beta yet?

alastairp: Okay :) If there are any changes, can I make them tomorrow?

Probably not

ansh: yes, of course. sorry for the delay

Thanks :D

Pratha-Fish: cool, now you can run `./devlop.sh build` to build your listenbrainz containers

Pratha-Fish: these instructions are at https://listenbrainz.readthedocs.io/en/latest/d... btw

oopsie

hmm

did we ever set you up with docker access? perhaps not, as you were just using jupyter

one sec

yep

are you doing this through a terminal in vscode, or separate ssh?

vscode terminal

ok, we need to convince vscode to restart itself and log in again

alrighty

is there a command (control-shift-p) called "Remote-SSH: kill VS Code Server on Host" ?

alastairp: yes, there is :D

ok, let's try that

Too late, VScode's restarting :D

the problem is that restarting vscode locally won't fix it

because when it reconnects, it'll connect to an existing long-running process on the server

we need to shut down and restart those ones

lucifer: hi, it doesn't look like we have a local dev setup for labs api, right?

alastairp: yup don't have it.

ah, but there _is_ a `docker/docker-compose-labs-api.yml`

did I make that?

yes, looks like I did. it's a local only file :)

ah no, i see it here as well

huh

its docker/docker-compose.labs.api.yml here

. vs -

yep

good, lots of fixes for Pratha-Fish to put in his PR ;)

however, its local as in you need a mb db and typesense server also available.

yeah, but the typesense server on wolf is connected to the musicbrainz network

so that should be perfect

yup makes sense

Pratha-Fish: any luck? I see a few new vs code services started (though I also see some old ones running)

I've restarted it for the second time, apparently it's still not working

The kill SSH command didnt do much either TBH

docker commands still fail?

yeah. I remember looking into this when ansh tried it. it seems like there's no great solution

Pratha-Fish: anyway, perhaps you could just ssh in with a separate terminal?

sure

I'll kill all of your servers on wolf meanwhile

alastairp: should I port forward while connecting? (just in case)

they should all be shut down and restarted, maybe your terminal there works too now

not yet

we need to decide what to do about ports

hm

alastairp: Alright, I reconnected on vscode, and ran the command there. It worked :D

great. has it finished?

yep

great

so actually, this part of LB is still a bit developer-unfriendly

do you know about docker and docker-compose?

Not really

ok

we have this helper script, develop.sh. in the background it uses `docker-compose`

if you look at the file `docker/docker-compose.yml`, you'll see that it defines a bunch of services

these are all of the programs that LB needs to run. the flask webserver, a database server, redis, some other stuff

I see it in the docker folder

and when you run `./develop.sh up` it starts all of these services and makes them available on a port

however, the mapper, which we're looking at is separate from this infrastructure at the moment

take a look, we have a config file `docker/docker-compose.labs.api.yml`

I see it

this says "run a service called labs_api", the `build` section tells us how to set up the code, the `command` tells us what to run to start it, and `ports` tells us which ports it should be available on in the server

the `networks` section tells us that we want to be able to connect to the musicbrainz database which is running on wolf

reosarevok: not atm but I'll try to review stuff tonight so we have more to look at tomorrow

I see

Okay, sounds good

so, we can do 2 things: either run all of LB, or just run the mapper

as we just need to fix the mapper, let's do that instead

we have another little helper script which allows us to run the mapper, it's `./docker/start-labs-api.sh`

which works like the develop.sh script

so let's edit this file, and again change the argument to `-p` to the same that you put in develop.sh

Done

(this is called the docker-compose "project", it's a prefix that docker-compose uses to keep track of its processes. normally it's OK to leave it on "listenbrainz" but because we're on a shared server, it'll cause conflicts if multiple people run it at the same time)

ok, let's try and run it

What does the -p represent here though? I am guessing "profile"

./docker/start-labs-api.sh build

^ project :)

ah

Looks like the ```./docker/start-labs-api.sh build``` command is throwing an error