#metabrainz

/

      • Mr_Monkey
        `docker volume ls`
      • 2019-02-01 03256, 2019

      • alastairp
        ruaok: iliekcomputers: Got an exception on acousticbrainz regarding redis, it looks like it might not be configured to allow as much memory usage? Is this something that you can look at?
      • 2019-02-01 03207, 2019

      • bukwurm
        Yeah, I a couple of them
      • 2019-02-01 03222, 2019

      • bukwurm
        *see a
      • 2019-02-01 03238, 2019

      • ruaok
        do you know how much you need?
      • 2019-02-01 03239, 2019

      • Mr_Monkey
        If you want a fresh start, find the ones associated with the BB containers, and delete them.
      • 2019-02-01 03213, 2019

      • bukwurm
        I am running compose up, if everything goes well - I think I'll continue
      • 2019-02-01 03217, 2019

      • alastairp
        no idea sorry. All I can say is "it worked fine when we were using LB's!"
      • 2019-02-01 03223, 2019

      • ruaok
        start_redis 'acousticbrainz-redis' '--maxmemory 8MB --maxmemory-policy allkeys-lru --save ""'
      • 2019-02-01 03236, 2019

      • ruaok
        previously it was unlimited, I think.
      • 2019-02-01 03238, 2019

      • alastairp
        the only thing we store in the cache at the moment is the recent list of recordings on the homepage, and some data about stats
      • 2019-02-01 03241, 2019

      • alastairp
        I don't know how much that is
      • 2019-02-01 03259, 2019

      • ruaok
        admittedly 8MB is rather paltry.
      • 2019-02-01 03221, 2019

      • alastairp
        e.g. if we could upgrade it to ~500 MB I think that would be fine, but that means we should fix all our cache code to consider the possibility of the cache being full
      • 2019-02-01 03226, 2019

      • ruaok
        boingo:~->free -h
      • 2019-02-01 03226, 2019

      • ruaok
        total used free shared buff/cache available
      • 2019-02-01 03226, 2019

      • ruaok
        Mem: 62G 11G 14G 672M 36G 49G
      • 2019-02-01 03236, 2019

      • ruaok
        I think we can do that. lol.
      • 2019-02-01 03247, 2019

      • alastairp
        OK, let's put it at 2GB or something, then?
      • 2019-02-01 03256, 2019

      • ruaok
        ok.
      • 2019-02-01 03210, 2019

      • ruaok
        can I get a "awww yisss" on this please?
      • 2019-02-01 03211, 2019

      • ruaok
      • 2019-02-01 03224, 2019

      • ruaok
        Mr_Monkey, zas, alastairp, bitmap, iliekcomputers: ^^
      • 2019-02-01 03225, 2019

      • alastairp
        not from me you can't!
      • 2019-02-01 03254, 2019

      • ruaok
        for once. 😆
      • 2019-02-01 03209, 2019

      • Mr_Monkey
        That was quick.
      • 2019-02-01 03226, 2019

      • ruaok
        I added 3 characters. hardly work.
      • 2019-02-01 03243, 2019

      • Mr_Monkey
        But 3 important characters, mind you.
      • 2019-02-01 03251, 2019

      • culinko has quit
      • 2019-02-01 03252, 2019

      • ruaok
        yvanzo: ^^ can you have a look so I can deploy this fix?
      • 2019-02-01 03248, 2019

      • alastairp
        mmm, got the python3.6 base image down from 800MB to 400MB!
      • 2019-02-01 03258, 2019

      • alastairp
        PR pending
      • 2019-02-01 03210, 2019

      • yvanzo
        ruaok: increasing memory x256? :)
      • 2019-02-01 03222, 2019

      • alastairp
        yvanzo: don't blame him, it was me who asked for it :)
      • 2019-02-01 03253, 2019

      • yvanzo
        as long as it is a power of 2
      • 2019-02-01 03223, 2019

      • Mr_Monkey
        alastairp: 50% gain? How?
      • 2019-02-01 03233, 2019

      • Mr_Monkey
      • 2019-02-01 03233, 2019

      • alastairp
        Mr_Monkey: it's all about layers
      • 2019-02-01 03235, 2019

      • alastairp
        like an onion
      • 2019-02-01 03252, 2019

      • yvanzo
        There are 10GB free on boingo atm, and we will retire MB search server soon.
      • 2019-02-01 03225, 2019

      • D4RK-PH0ENiX has quit
      • 2019-02-01 03226, 2019

      • ruaok
        we can start with 1GB. that is probably sufficient. but we have loads of ram in general.
      • 2019-02-01 03207, 2019

      • D4RK-PH0ENiX joined the channel
      • 2019-02-01 03208, 2019

      • bukwurm
      • 2019-02-01 03212, 2019

      • bukwurm
        Mr_Monkey: /\
      • 2019-02-01 03242, 2019

      • ruaok
        alastairp: eeb26b090937 redis:3.2.4-alpine "docker-entrypoint..." 9 seconds ago Up 8 seconds acousticbrainz-redis
      • 2019-02-01 03248, 2019

      • Mr_Monkey
        bukwurm: That's odd. Maybe try running ./develop.sh again?
      • 2019-02-01 03215, 2019

      • Mr_Monkey
        It should wait for ES to be available before starting the web container, so that's strange
      • 2019-02-01 03235, 2019

      • bukwurm
        Ok.
      • 2019-02-01 03238, 2019

      • Mr_Monkey
        Is your dockr-compose any different than what's currently on master?
      • 2019-02-01 03250, 2019

      • Mr_Monkey
        .yml*
      • 2019-02-01 03209, 2019

      • D4RK-PH0ENiX has quit
      • 2019-02-01 03216, 2019

      • bukwurm
        That was probably because I was only running `docker-compose up --build bookbrainz-site`
      • 2019-02-01 03220, 2019

      • bukwurm
        It's connecting now
      • 2019-02-01 03258, 2019

      • Mr_Monkey
        Ahaa, yes. We've moved to using ./develop.sh
      • 2019-02-01 03207, 2019

      • dolina has quit
      • 2019-02-01 03255, 2019

      • D4RK-PH0ENiX joined the channel
      • 2019-02-01 03239, 2019

      • dolina joined the channel
      • 2019-02-01 03228, 2019

      • bukwurm
        Mr_Monkey: After multiple attempts (and getting the below), it opens up but still no search results
      • 2019-02-01 03231, 2019

      • bukwurm
      • 2019-02-01 03241, 2019

      • Mr_Monkey
        bukwurm: Hmm, never seen that. It looks like it's related to docker-compose. Maybe restart Docker on your machine, try again?
      • 2019-02-01 03229, 2019

      • Mr_Monkey
        Not sure what's happening there.
      • 2019-02-01 03236, 2019

      • dolina has quit
      • 2019-02-01 03227, 2019

      • bukwurm
        Mr_Monkey: You use a Mac right?
      • 2019-02-01 03234, 2019

      • bukwurm
        Are you on Mojave?
      • 2019-02-01 03247, 2019

      • Mr_Monkey
        On High Sierra
      • 2019-02-01 03251, 2019

      • bukwurm
        Ok
      • 2019-02-01 03220, 2019

      • bukwurm
        The build process was painfully slow and docker was getting unresponsive quite too often
      • 2019-02-01 03249, 2019

      • bukwurm
        I checked memory usage which was around 3-4 G, but CPU usage was 200%+
      • 2019-02-01 03252, 2019

      • bukwurm
        Is this normal?
      • 2019-02-01 03216, 2019

      • bukwurm
        I looked up and found quite recent discussion here https://github.com/docker/for-mac/issues/1759
      • 2019-02-01 03235, 2019

      • Mr_Monkey
        It's behaving on my machine, but I know some users have had issues with ElasticSearch. This might be a similar issue, with ES being to greedy
      • 2019-02-01 03236, 2019

      • Mr_Monkey
        bukwurm: If you want, you can try updating the ES image version in docker-compose, and see if that makes a difference for resources. I've seen better performance from more recent versions.
      • 2019-02-01 03251, 2019

      • pristine_ joined the channel
      • 2019-02-01 03252, 2019

      • dolina joined the channel
      • 2019-02-01 03232, 2019

      • Mr_Monkey
        It's bound to break some code, but just to test resource-wise it might give you a better idea if that's the issue
      • 2019-02-01 03230, 2019

      • pristine_
        ruaok , iliekcomputers , Mr_Monkey , alastairp : how about making the last-fm imports dynamic? Once they give permission, should we import listens as and when they come without the user having to do it explicitly.
      • 2019-02-01 03231, 2019

      • ruaok
        is that possible?
      • 2019-02-01 03255, 2019

      • iliekcomputers
        Could be done like we do with Spotify probably
      • 2019-02-01 03215, 2019

      • bukwurm
        Mr_Monkey: I increased memory from 2GB to 4GB and build happened in 2-4 seconds with already running containers
      • 2019-02-01 03229, 2019

      • ruaok
        oh yeah, duh. the api is open and we've handled all the bad parts about it.
      • 2019-02-01 03234, 2019

      • iliekcomputers
        Things to think about are rate limits and possible duplicates
      • 2019-02-01 03237, 2019

      • bukwurm
        But now docker is taking 6GB ram
      • 2019-02-01 03245, 2019

      • ruaok
        duplicates. that's the main problem.
      • 2019-02-01 03248, 2019

      • bukwurm
        which is fine I guess
      • 2019-02-01 03218, 2019

      • ruaok
        I think we may need better duplicate support and removal of duplicates before we add the feature.
      • 2019-02-01 03232, 2019

      • Mr_Monkey
        :/ hungry hungry BB.
      • 2019-02-01 03232, 2019

      • Mr_Monkey
        So it's all working now, bukwurm ? Would you mind documenting your findings and solution in the repo?
      • 2019-02-01 03201, 2019

      • bukwurm
        Mr_Monkey: It's pretty fast now, CPU usage is also down around 100% though RAM usage is high
      • 2019-02-01 03215, 2019

      • bukwurm
        But I am still not getting any search results :/
      • 2019-02-01 03259, 2019

      • Mr_Monkey
        Did you try slamming a heavy object on the computer?
      • 2019-02-01 03259, 2019

      • bukwurm
        The logs give 200 Ok though
      • 2019-02-01 03205, 2019

      • Mr_Monkey
        Hm.
      • 2019-02-01 03210, 2019

      • bukwurm
        Lol
      • 2019-02-01 03249, 2019

      • bukwurm
        I'll try checking if indexes have been created or not in my case
      • 2019-02-01 03207, 2019

      • Mr_Monkey
        Yep
      • 2019-02-01 03240, 2019

      • ruaok
        pristine_: iliekcomputers: this is what I worry about: https://blog.dbrgn.ch/2012/7/29/mail-loop-from-he…
      • 2019-02-01 03245, 2019

      • ruaok
        🤣
      • 2019-02-01 03248, 2019

      • Mr_Monkey
        I was going to say, try localhost:9099/search/reindex .
      • 2019-02-01 03248, 2019

      • Mr_Monkey
        You'll need to have a look at src/server/routes/search.js L129 before that
      • 2019-02-01 03256, 2019

      • Mr_Monkey
        bukwurm: ^
      • 2019-02-01 03245, 2019

      • pristine_
        ruaok: Lol.
      • 2019-02-01 03235, 2019

      • ruaok
        pristine_: I think focusing on apple music first is a good idea. we have *some* form on import for last.fm. let's do something we have no support for all...
      • 2019-02-01 03202, 2019

      • iliekcomputers
        >I managed to bring down 13 CIA offices, all FBI offices east of the Mississippi, and the entire Southeastern university Research Network.
      • 2019-02-01 03210, 2019

      • iliekcomputers
        Noiice
      • 2019-02-01 03208, 2019

      • dolina has quit
      • 2019-02-01 03218, 2019

      • Mr_Monkey
        bukwurm: I meant Line 133
      • 2019-02-01 03232, 2019

      • pristine_
        ruaok: yeah. Sounds good.
      • 2019-02-01 03236, 2019

      • bukwurm
        *132
      • 2019-02-01 03256, 2019

      • reosarevok
        iliekcomputers: see, you only cost money. Not so bad! :p
      • 2019-02-01 03210, 2019

      • dolina joined the channel
      • 2019-02-01 03236, 2019

      • pristine_
        ruaok: I did some testing last night. Tried consuming data from rabbitMQ on lb-server side and then pushed them to db.
      • 2019-02-01 03213, 2019

      • pristine_
        We need to create a producer at lb-playground
      • 2019-02-01 03243, 2019

      • bukwurm
        Mr_Monkey: Reindexing works
      • 2019-02-01 03251, 2019

      • bukwurm
        \o/
      • 2019-02-01 03208, 2019

      • Mr_Monkey
        bukwurm: OK, good.
      • 2019-02-01 03216, 2019

      • Mr_Monkey
        Everything working?
      • 2019-02-01 03220, 2019

      • bukwurm
        Looks like it
      • 2019-02-01 03231, 2019

      • Mr_Monkey
        \o/
      • 2019-02-01 03206, 2019

      • bukwurm
        One more thing: we're not using nodemon in docker run?
      • 2019-02-01 03234, 2019

      • bukwurm
        Mr_Monkey: /\
      • 2019-02-01 03257, 2019

      • Mr_Monkey
        bukwurm: No. Suposedly stopping and rerunning ./develops.hs should be fast enough. But you can definitely change the run command in your docker-compose file if you prefer
      • 2019-02-01 03227, 2019

      • code_master5 joined the channel
      • 2019-02-01 03251, 2019

      • code_master5
        reosarevok: mind if I pm?
      • 2019-02-01 03202, 2019

      • bukwurm
        Mr_Monkey: Cool, and thank you!
      • 2019-02-01 03222, 2019

      • Mr_Monkey
        My pleasure, nice to see you around ! :)
      • 2019-02-01 03241, 2019

      • bukwurm
        > Would you mind documenting your findings and solution in the repo? 5:35 PM <bukwurm>
      • 2019-02-01 03212, 2019

      • bukwurm
        Should I add it in the README?
      • 2019-02-01 03212, 2019

      • bukwurm
        Yeah, I have been totally out of the loop for a couple of months now, hopefully I'll be able to contribute more :)
      • 2019-02-01 03211, 2019

      • alastairp
        pristine_: ruaok: remember the reason to do last.fm in the browser is so that it was the user who did the requests to l.fm, not us
      • 2019-02-01 03252, 2019

      • Freso
        amCap1712: Log out of forums and back, that should carry it over IIRC.
      • 2019-02-01 03234, 2019

      • ruaok
        alastairp: oh yes. Good point.
      • 2019-02-01 03209, 2019

      • ruaok
        pristine_: yeah, I need to start working on that very soon.
      • 2019-02-01 03219, 2019

      • reosarevok
        code_master5: go ahead I guess?
      • 2019-02-01 03235, 2019

      • Mr_Monkey
      • 2019-02-01 03240, 2019

      • Mr_Monkey
        In there please :)
      • 2019-02-01 03230, 2019

      • amCap1712
        Freso: it worked. thanks
      • 2019-02-01 03237, 2019

      • pristine_
        alastairp: oh yes:)
      • 2019-02-01 03213, 2019

      • amCap1712
        I had a question. Its been half a month that I am working on the app. Should I post an update in the GSoC proposal about what has been materialized in these days?
      • 2019-02-01 03204, 2019

      • pristine_
        ruaok: yeah. Would love to assist you in that 😆
      • 2019-02-01 03214, 2019

      • ruaok
        Well, if you want to start, we can make it your task and I'll mentor you...
      • 2019-02-01 03244, 2019

      • Mr_Monkey
        ruaok: I've made some good UI improvements on the listens page on branch spotify-player-styling. When you get a chance, can you try it out and give me your feedback?
      • 2019-02-01 03255, 2019

      • Mr_Monkey
        Same for iliekcomputers :)
      • 2019-02-01 03235, 2019

      • pristine_
        ruaok: sounds wow :)
      • 2019-02-01 03246, 2019

      • iliekcomputers
        ruaok: Note that we're talking about the stats movement from the cluster to hetzner and not listen streaming spark writer. :)
      • 2019-02-01 03217, 2019

      • iliekcomputers
        I think pristine_ has a consumer all ready for that.
      • 2019-02-01 03249, 2019

      • iliekcomputers
        I can take the spark writer stuff in the meanwhile if needed.
      • 2019-02-01 03216, 2019

      • pristine_
        ruoak: I made a PR yesterday. It calculates some stats for users. It is very basic as of now. I was thinking of using the same script to calculate stats of users such that we build a json doc for every user and then push it into the queue. That ways it would be easier to retrieve data from the other end and fill the db. What do you think?
      • 2019-02-01 03226, 2019

      • alastairp
        iliekcomputers: hi, I've finished all the docker python PRs, just got one small one to finish regarding the push.sh script
      • 2019-02-01 03213, 2019

      • alastairp
        I wanted to chat with zas and/or bitmap one final time about what we discussed yesterday regarding metabrainz-specific version tags in dockerhub (e.g. :3.6-2 or :3.6.8)