#metabrainz

/

      • Etua joined the channel
      • Etua has quit
      • jesus2099 joined the channel
      • jesus2099 has quit
      • BrainzGit
        [listenbrainz-server] 14dependabot[bot] opened pull request #2011 (03master…dependabot/pip/relations/ujson-5.2.0): Bump ujson from 2.0.3 to 5.2.0 in /relations https://github.com/metabrainz/listenbrainz-serv...
      • [listenbrainz-server] 14dependabot[bot] opened pull request #2012 (03master…dependabot/pip/ujson-5.2.0): Bump ujson from 1.35 to 5.2.0 https://github.com/metabrainz/listenbrainz-serv...
      • [sir] 14dependabot[bot] opened pull request #133 (03master…dependabot/pip/ujson-5.2.0): Bump ujson from 1.35 to 5.2.0 https://github.com/metabrainz/sir/pull/133
      • [listenbrainz-server] 14dependabot[bot] opened pull request #2013 (03master…dependabot/pip/listenbrainz/mbid_mapping/ujson-5.2.0): Bump ujson from 2.0.3 to 5.2.0 in /listenbrainz/mbid_mapping https://github.com/metabrainz/listenbrainz-serv...
      • [troi-recommendation-playground] 14dependabot[bot] opened pull request #55 (03main…dependabot/pip/ujson-5.2.0): Bump ujson from 1.35 to 5.2.0 https://github.com/metabrainz/troi-recommendati...
      • [musicbrainz-server] 14mwiencek opened pull request #2544 (03master…react-18): Update react, react-dom to v18 https://github.com/metabrainz/musicbrainz-serve...
      • yvanzo
        O’Moin
      • bitmap, reosarevok: Thanks for having hotfixed MBS in production. I updated MBS releases in Tickets and added the postponed release.
      • s1b1 has quit
      • BrainzGit
        [sir] 14yvanzo closed pull request #133 (03master…dependabot/pip/ujson-5.2.0): Bump ujson from 1.35 to 5.2.0 https://github.com/metabrainz/sir/pull/133
      • yvanzo
        lucifer: I disabled Dependabot security updates because of Python 2 ^ but kept security alerts on just in case.
      • lucifer
        makes sense
      • yvanzo
        By the way is there anything I can do about the tests PR?
      • lucifer
        there is a weird issue. tests pass locally and on wolf but fail in jenkins.
      • i have been unable to debug it yet and hence haven't added the remaining tests yet.
      • if you could help debug it, that would be great
      • yvanzo
        Will do
      • lucifer
        (due to current debugging attempts, the commits and changes are a bit haphzard. once issue is resolved will clean up)
      • thanks!
      • yvanzo
        bitmap, reosarevok: Is the prod hotfix supposed to resolve this issue as from next Saturday? https://blog.metabrainz.org/2022/05/16/musicbra...
      • reosarevok
        yvanzo: yes, that's the same person who wrote to support later, we've already answered :) (but we can always answer there too just so others who see it know, I guess?)
      • yvanzo
        Yes, it is probably even worth a ticket for the next release if you already investigated this issue.
      • reosarevok
      • BrainzBot
        MBS-12400: cover_art_archive dump uses event_art_archive.art_type
      • reosarevok
        (although the issue ended up being wider than that, as per the PR)
      • yvanzo
        reosarevok: Thanks, updated the ticket and replied on the blog too.
      • reosarevok
        Thanks!
      • MRiddickW joined the channel
      • yvanzo
        mayhem: which document?
      • lucifer
        >[off] https://docs.google.com/document/d/1Doi5s99Pjry... -- please see section Use Cases.
      • this one probably
      • mayhem
        moin!
      • indeed, that one. thanks!
      • lucifer
        morning
      • mayhem: lmk when you have some time, want to discuss an issue with listens in LB dev.
      • mayhem
        now is fine.
      • lucifer
        cool, so as you know we use listen_user_metadata table to store timestamps and fetch listens in production. those timestamps are updated by a cron job but that job doesn't run in development so timestamps are never updated in dev. they are always set to NULL in development.
      • these are coalesced to 0 when fetched so the check to load listens in dev, ends up executing this if: https://github.com/metabrainz/listenbrainz-serv...
      • mayhem
        less than ideal.
      • if wonder if we can do some on-demand calculating this info...
      • atj
        lucifer: is port 80 only required to be open on the spark leader?
      • lucifer
        atj: yes, it currently running a nginx server
      • mayhem
        atj: kinda yes, kinda no.
      • we need it at times for debugging purposes.
      • atj
        OK, so open on all nodes then
      • mayhem
        no, we only need it on michael.
      • atj
        oh, I see
      • mayhem
        lucifer: there is a manage.py script we can run manually to calculate this data, yes?
      • what if we ran that at the beginning on develop.sh? or command of develop.sh?
      • lucifer
        mayhem: yes there is a manage.py script to update timestamps. that is the issue i had in mind, do we want contributors to have to run this command
      • mayhem
        I fear that automating it would get in the way.
      • alastairp
        if it was at develop.sh, the'd have to run this every time they added somethign to the db? or only once, and it'll do the slow method afterwards?
      • could the timescale writer do it in dev mode?
      • mayhem
        what if we added a note tot he UI in the most prominent place that if DEBUG and no listens, show a "hey you need to run this script" message
      • ?
      • lucifer
        we have a slow method to take over later. but this command needs to be run once for that to kick in.
      • mayhem
        oh, TS writer is a pretty good idea!
      • once the TS writer goes idle and DEBUG=True, run the script.
      • lucifer
        i see, makes sense
      • we can't check idle easily in ts writer because its always listening to rabbitmq but can run the command after every insert.
      • alastairp
        in DEBUG=True I don't see a big problem with running it after every insert
      • mayhem
        we could change the reading method to non-blocking, right?
      • lucifer
        that becomes a problem if there it too much data but i don't think most local dbs will have those.
      • mayhem
        but that is a more drastic change. perhaps we should run it after every insert for now and see if that causes problems.
      • lucifer
        i had tried that when ts writer had issues 2 days ago, it doesn't work well.
      • mayhem
        ok, then lets KISS
      • lucifer
        (non-blocking thing)
      • 👍
      • skelly37 joined the channel
      • atj
        lucifer: FYI, michael is running "python3 -m http.server --bind 0.0.0.0 80" not nginx
      • lucifer
        ah i see.
      • atj
        also, there is a uhttpd container running and listening on port 49153
      • any idea what that is for?
      • hmm, looks like it was meant to serve the recommendations files too, but the python webserver is being used instead
      • mayhem
        hmmm.
      • I stopped the uhttpd container since that was not setup as expected.
      • atj
        thanks
      • mayhem
        ok, it is cleaned up now and the uttp contianer
      • is running on port 80. which is now not reachable.
      • mayhem will continue
      • atj: any idea why that is suddenly no longer reachable? it is listening on port 80, and the container has `-p "80:80"`
      • daschuer joined the channel
      • daschuer has quit
      • daschuer joined the channel
      • daschuer has quit
      • yvanzo
        To project maintainers, more reviews are needed at https://github.com/metabrainz/guidelines/pull/17
      • It is a first step before opening tickets for all projects using RabbitMQ and then doing the same for all other services (Redis, Postgres…).
      • alastairp
        thanks for the reminder yvanzo
      • atj
        mayhem: hmmm, just checking
      • reosarevok
        yvanzo: do I remember correctly that you wanted to try and update the DB diagrams after the schema change? :)
      • yvanzo
        reosarevok: Yes, won’t have time this week, but will open a PR next week. basically it is an adaptation of admin/GenerateSQLScripts.pl to generate diagrams, which is a good fit for parsing MBDB schema code.
      • reosarevok
        I don't think there's a huge rush anyway, so that sounds good :)
      • Thanks!
      • skelly37 has quit
      • atj
        mayhem: it's working now but I didn't do anything :) firewall rules looked correct
      • mayhem
        it times out for me: michael.metabrainz.org/
      • atj
      • mayhem
        yep
      • lucifer
        doesn't load for me eihter
      • atj
        ah, IPv6 works but not IPv4
      • alastairp
        some people have ipv6? :(
      • atj
        yes, it's great!
      • lucifer
        it fails for me on ipv6 as well
      • BrainzGit
        [design-system] 14akshaaatt opened pull request #77 (03master…add-prettier): Add prettier and Setup Lint GitHub workflow https://github.com/metabrainz/design-system/pul...
      • lucifer
        alastairp, yvanzo: hi! do you know to mock.patch all objects in a dict or all subclasses of a particular class?
      • alastairp
        magicmock comes to mind
      • lucifer
        oh patch.dict
      • alastairp
        or that :)
      • BrainzGit
        [guidelines] 14yvanzo opened pull request #20 (03master…docker-optional-ports): Avoid using EXPOSE directive for optional ports https://github.com/metabrainz/guidelines/pull/20
      • atj
        mayhem, lucifer: should be fixed now
      • do you want the container to autostart?
      • mayhem
        not for me.
      • I think starting it manually is ok for us for now.
      • atj
        ok
      • mayhem
        does the page load for you now lucifer ?
      • still times out for me.
      • alastairp
        works for me
      • lucifer
        yes loads for me now but without css
      • mayhem
        ah, FF insisting on https. Yes, works now.
      • I dont think there is any css, lucifer
      • lucifer
        oh nvm that's the index page
      • mayhem
        atj: what needed changing?
      • lucifer
        thanks atj!
      • mayhem
        yes, indeed -- thanks!
      • lucifer
        it appears docker was missing a firewall rule
      • mayhem
        ok, not our problem then.
      • atj
        mayhem: it was actually the docker configuration, it needed changing to disable iptables support
      • MRiddickW has quit
      • mayhem
        got it, thanks.
      • lucifer
        alastairp: oh well, actually need to mock 1 attribute of all items inside a dict here :(. i wondering if it'll be easier to just store an original copy and fix it manually later.
      • mayhem
        lucifer: monkey alastairp can I deploy some test containers to test.mb to debug an issue?
      • monkey
        If you mean test.lb, then OK for me
      • mayhem
        yeah, that one.
      • lucifer
        test.lb? :) yes should be fine, i deployed it last yesterday.
      • mayhem
        k, thx
      • leonardo
        i thiunk i accidentally listenbrainz again https://imgur.com/ViuRQQz.png
      • mayhem wants to know why the release_group tags in the DB are not appearing in the api
      • mayhem
        leonardo: did you just do an import?
      • lucifer
        leonardo: whats your username on listenbrainz?
      • leonardo
        lucifer: leofiore
      • mayhem: i tried a partial import this saturday and didn't succeded well, so i decided to delete my listens and reimport it again
      • lucifer
        leonardo: oh then, its a cached value. the count will reset to 0 in a few minutes.
      • leonardo
        i see, but the listen counts was wrong too
      • i had ~80k listens, and it shows ~55k
      • looks a bit weird
      • lucifer
        80k on last.fm?