#metabrainz

akshaaatt[m]: looks awesome!! awesome work :D i'll try to review it later today.

lucifer I found AcousticBrainz Android Client and made a pull request on it 😃

akshaaatt[m]: oh nice. i intended to add that functionality to the app last year but there are some issues with the ffmpeg dependency so not done that yet.

zas: getting this error when trying to open stats.meb https://usercontent.irccloud-cdn.com/file/U6Xa4...

lucifer: try to force a full reload

(the site works for me atm)

ah that worked, thanks!!

alastairp: thoughts on adding a manage.py commnad to read an sql script to initialize the db and execute it with psycopg2.

i am working on simplifying meb.org dev setup and currently it uses subprocess to execute the scripts through psql

*to clarify, only db and extensions initialisation and dropping is using psql)

lucifer: the init_db methods in LB do that

oh nice! it makes sense to do that in other projects as well methinks.

plus, we could add a command that accepts a file name to execute for apply db migrations in prod, that way we don't need to include psql client in the images.

lucifer I had a chat with timusus, the guy with the KTagLib, regarding overwriting/saving artwork to the files and apparently we'll have to add that feature on our own since he's not planning on working on it as of now. On Musicbrainz tagger, I've added cover art supports from our APIs but yeah we lag the part about saving them locally.

Lack*

Btw a French guy reached out to me regarding our app and he's really happy with it. He was suggesting some features like default select the directory on tagger. After that I went through playstore for any other apps which provide such functionality and found that our app is not being indexed on playstore. We need to better write the description, add some relevant terms in the title like "MusicBrainz - Tag, Listen and Search" something

like that and even add a precise description.

Maybe the other team members can help us prepare really good content to go with the playstore which is our policy friendly and everything.

akshaaatt[m]: indeed, the app description hs much scope of improvements. regarding app indexing, its a conscious decision to not do that. the app is not on feature parity with the website and probably never will be because edits are only possible through the api so we probably don't want to redirect users using the website to the app.

*edit are _not_ possible through the api

Right

Do you have any recommended tutorial for me where I could learn about JNI and how these C++ libraries work? Because I am proficient with C++ but I still am not sure how these libraries are communicating with the Android internal hardware.

If I get the hang of it I could start adding more features to our app

the ndk tutorials on developer.android are nice and a good starting point. eg: https://developer.android.com/training/articles...

lucifer: yeah, doing migrations through python was an idea that I had, this is definitely a good feature

sometimes we do still want to get a psql shell somewhere, so we should have a way of doing that

right, i was thinking how to keep that.

perhaps we could do it through the server container instead of the python app container

for now I don't think it's a problem that we keep psql in the containers

i am looking to merge Dockerfiles for meb.org, i'll keep psql for now then

do you know if its expected that autocommit mode will not work if we try to execute multiple lines at once? maybe that's why we read a line at once.

I can't remember why, sorry

the docs do not mention anything of this sort but it looks like it still executes inside a transaction if i execute the entire file. but if i copy paste the function i linked above, it works fine.

i'll try to debug this later and see if i can figure what the issue is.

do you want to discuss what i should work on next in AB?

sure, let's do it

the continuation of #292 would be good

taking look at it

note that it's a PR into https://github.com/metabrainz/acousticbrainz-se...

i think we have discussed this ticket before. let me see if i can find that

I'd be tempted to just merge this PR as-is, then get the integration branch up to date with current master

i think we should that. we can continue doing the fixups in that branch then.

*should do

then try and make sure that we understand the workflow of how this branch works

if it's more or less good, then we could probably merge/release as-is and continue work as necessary. I recall that this PR just sets the groundwork anyway, we'll need additional features to actually make use of it

makes sense

i'll merge this PR then and start looking into the overall changes in the branch.

!m outsidecontext

You're doing good work, outsidecontext!

let's hope we can now focus on 2.7, this maintenance release wasn't really planned ;)

yep. but as I see it currently 2.7 is also still a bit out, and likely we will have a longer beta period this time. the changes are rather extensive

yup, imho we should even release a bunch of alpha versions on the path to first beta, I expect even more massive changes (for the good)

alastairp: just found my exact issue, https://github.com/psycopg/psycopg2/commit/7479...

ah, great. yeah, we have to be careful about parsing these lines then

added to docs only 3 weeks ago :)

I'll add a comment to the function in our codebase as well.

now that i think these functions could probably go into BU.

moooin!

zas: ping

the AX51-NVMe servers are ready at Hetzner -- we can't act on the williams replacement yet, but we can act on the prince replacement.

we should schedule the move for prince for.. tomorrow. you ok with that?

Yup, ok for me

what time should we say>?

and can you just confirm that we are replacing prince tomorrow?

Let me check, lunch atm

ok

prince is running bookbrainz, rabbitmq, redis

(and elasticsearch)

yes

we need to move all these to other servers, I guess it is feasible, but not sure how much it impacts services

elasticsearch is used only by bookbrainz?

Yes

We can do without test.bookbrainz for a bit if it helps (also means only one instance of elasticsearch instead of 2)

the main issue will be rabbitmq (it is used by lb, caa, sir)

how easy is it to move this to another server? can it be done without downtime (of related services)?

sir downtime is not a problem

caa not sure (bitmap?)

and lb is your stuff, so I guess you can tell

i think it should be easier now, lb was preped last time rabbitmq was moved.

there's also a redis instance, not sure what is using it (mb?)

ruaok: let's wait for bitmap to be around, and we decide appointment time depending on this, I guess somewhere end of afternoon tomorrow (17?)

the setup of the new machine should be fast (as most scripts should now work without much hassle), if everything goes smooth, we can expect the new server to be ready under one hour in any case

but in case of hardware issue with new server ofc

ah redis is bookbrainz-redis

so I guess only used by bb

what do we need bitmap for in this case? (curious, I'm fine waiting)

yvanzo: can you check how hard it is to move rabbitmq for sir?

ruaok: I'd prefer to have bitmap for caa/rabbitmq migration

got it.

we moved all the stuff to boingo?

s/moved/move/

load is low on both boingo & prince, so I see no problem with that

monkey: can you start to move bb to boingo today?

Yes

ruaok: any update from Youtube side regarding Ouath?

Nothing yet. I wonder if the rerequest went through ok.

👍

ruaok: is the mail container useful during meb.org local development?

Naw, it's used to send mails about customers signing up etc. Not needed, really.

👍, removing from dev setup.

alastairp: do we need to drop tables here since we already dropped the schema? https://github.com/metabrainz/listenbrainz-serv...

nvm, i see the dropped tables are not in the dropped schemas.

brainzgit seems down

zas: hey, we should be able to migrate caa queues without downtime. (basically, spin up a new CAA-indexer pointing to the new queue, point pg_amqp to the new queue, then bring down the old CAA-indexer once the old queue is empty.) I can work on that now

maybe something similar is possible for sir, not sure about lb

it'd be a bit easier for all projects if we had a multi-node cluster, but

bitmap, zas: I checked how to migrate rabbitmq from prince to boingo.

bitmap: To soften the migration, I’m using shovels like last time.

won't that duplicate messages?

no

I see, it consumes them first

ok, I'll work on something else then :)

bitmap: just wanted to have you around in case caa get down :)

https://github.com/metabrainz/acousticbrainz-se... +5,098 −21 😨

oh yeah, I was worried about that :)

good work in fixing conflicts