#metabrainz

/

      • alastairp
        šŸ‘
      • 2022-05-03 12312, 2022

      • yvanzo
        reosarevok: This is the same issue as sir vTBD we discussed earlier :)
      • 2022-05-03 12319, 2022

      • reosarevok
        Yeah, thought so
      • 2022-05-03 12344, 2022

      • reosarevok
        So, it seems we need to figure out how to set stuff to R-O (if we want less downtime)
      • 2022-05-03 12305, 2022

      • reosarevok
        And to make sure we have a proper sir version
      • 2022-05-03 12308, 2022

      • reosarevok
        And that's about it
      • 2022-05-03 12309, 2022

      • hackerman4 joined the channel
      • 2022-05-03 12335, 2022

      • hackerman has quit
      • 2022-05-03 12335, 2022

      • hackerman4 is now known as hackerman
      • 2022-05-03 12325, 2022

      • zas
        yvanzo: it causes openresty to return 503s for all upstreams configured via consul -> https://github.com/metabrainz/openresty-gateways/…
      • 2022-05-03 12348, 2022

      • zas
        that's not a perfect solution, because it concerns only stuff going through gateways
      • 2022-05-03 12314, 2022

      • zas
        but basically it prevents openresty to pass requests to backends
      • 2022-05-03 12303, 2022

      • yvanzo
        Thank you!
      • 2022-05-03 12337, 2022

      • reosarevok
        yvanzo: should I merge all the open sql PRs into a local branch and try to at the very least upgrade.sh the sample db?
      • 2022-05-03 12300, 2022

      • reosarevok
        I guess we should eventually run the upgrade on a full DB too, but maybe starting with a sample is good to see if we find any quick and obvious issues
      • 2022-05-03 12344, 2022

      • reosarevok
        And once we properly merge all, then we run a full proper upgrade test or two
      • 2022-05-03 12334, 2022

      • yvanzo
        something like schema-change-2022-alpha.0 so that it will never be mistakenly used as a final thing by anyone?
      • 2022-05-03 12326, 2022

      • reosarevok
        I was thinking literally local, as in, in my laptop only :D
      • 2022-05-03 12331, 2022

      • yvanzo
        ah ok :)
      • 2022-05-03 12356, 2022

      • hackerman5 joined the channel
      • 2022-05-03 12341, 2022

      • zas
        yvanzo: I need your help on static files sync for rudi (sshd-nginx-static-resources container), it seems the server IP needs to be add to https://github.com/metabrainz/docker-server-confi… , so that's not very dynamic (I guess it implies a restart of all musicbrainz-website-* containers
      • 2022-05-03 12351, 2022

      • hackerman has quit
      • 2022-05-03 12351, 2022

      • hackerman5 is now known as hackerman
      • 2022-05-03 12349, 2022

      • zas
        do you think we can improve that somehow?
      • 2022-05-03 12311, 2022

      • yvanzo
        zas: I would first have to understand how it works for MB website :/ (MB sitemaps is easier to get.) I guess it could be set through DBDefs instead? Can make a ticket not to forget about it?
      • 2022-05-03 12307, 2022

      • yvanzo
        reosarevok: Are you more familiar with how MB websites updates staticbrainz? (I don’t find any call to rsync-staticbrainz-files but in rsync-sitemaps.)
      • 2022-05-03 12333, 2022

      • reosarevok
        I am not, not directly, but IIRC this was discussed at some point, let me see
      • 2022-05-03 12326, 2022

      • reosarevok
        "in MB we have a STATIC_RESOURCES_LOCATION config option that points to staticbrainz in prod" is that relevant?
      • 2022-05-03 12355, 2022

      • yvanzo
        Not directly
      • 2022-05-03 12309, 2022

      • reosarevok
        There's also StaticBrainz.md in syswiki but I guess you've seen that?
      • 2022-05-03 12327, 2022

      • yvanzo
        STATIC_RESOURCES_LOCATION is for getting resources from staticbrainz, not the other way around.
      • 2022-05-03 12310, 2022

      • reosarevok
        Oh, sure
      • 2022-05-03 12349, 2022

      • reosarevok
        Well, StaticBrainz.md does mention a bit of the other way around
      • 2022-05-03 12310, 2022

      • yvanzo
        Yes but it is not complete. It might be in a private repo
      • 2022-05-03 12356, 2022

      • BrainzGit
        [listenbrainz-server] 14alastair opened pull request #1978 (03master…metadata-cache-check-exists): Check if the canonical_release_redirect table exists before running https://github.com/metabrainz/listenbrainz-server…
      • 2022-05-03 12359, 2022

      • yvanzo
        zas, reosarevok: found it, server IP is set in musicbrainz-staticbrainz project in Jenkins.
      • 2022-05-03 12323, 2022

      • yvanzo
        It currently has herb & kiki only.
      • 2022-05-03 12349, 2022

      • zas
        can you add rudi?
      • 2022-05-03 12354, 2022

      • yvanzo
        ok
      • 2022-05-03 12337, 2022

      • zas
        I'm not sure the setup is correct on rudi's side, we'll find out soon
      • 2022-05-03 12329, 2022

      • yvanzo
        I updated the project configuration, now trying to synchronize test.mb.o resources
      • 2022-05-03 12352, 2022

      • reosarevok
        yvanzo: can you update the syswiki file?
      • 2022-05-03 12355, 2022

      • reosarevok
        Merged all except for https://github.com/metabrainz/musicbrainz-server/… locally, now checking
      • 2022-05-03 12309, 2022

      • reosarevok
        (yvanzo: when you have some time please rebase that and fix conflicts :) )
      • 2022-05-03 12354, 2022

      • yvanzo
        zas, reosarevok: no error when rsync-ing MB staticbrainz files from Jenkins to rudi :) looking into prod and beta now
      • 2022-05-03 12318, 2022

      • zas
        yvanzo: perfect :)
      • 2022-05-03 12331, 2022

      • zas
        thank you for the help on this
      • 2022-05-03 12349, 2022

      • yvanzo
        zas: apparently it did rsync static files for all of prod/beta/test at once, further runs based on beta/prod images did not change anything to rudi.
      • 2022-05-03 12323, 2022

      • BrainzGit
        [listenbrainz-server] 14alastair opened pull request #1979 (03master…metadata-cache-docs): LB-1079: Fix mb_metadata_cache update methods to work with bulk table code https://github.com/metabrainz/listenbrainz-server…
      • 2022-05-03 12333, 2022

      • yvanzo
        zas: Should I also add MB sitemaps to rudi?
      • 2022-05-03 12349, 2022

      • uZer has quit
      • 2022-05-03 12338, 2022

      • zas
        I think so
      • 2022-05-03 12316, 2022

      • zas
        In fact, anything set up for kiki/herb should also be for rudi
      • 2022-05-03 12321, 2022

      • yvanzo
        zas: Is LGTM for the changes and/or the question about dropping set -x on https://github.com/metabrainz/docker-server-confi… ?
      • 2022-05-03 12350, 2022

      • BrainzGit
        [listenbrainz-server] 14alastair opened pull request #1980 (03master…consul-config-listenstore): Use consul-template if-statement to check if the listenstore is up https://github.com/metabrainz/listenbrainz-server…
      • 2022-05-03 12326, 2022

      • zas
        yvanzo: ah yes, sorry, I think you can drop set -x, it was there for debugging purposes
      • 2022-05-03 12338, 2022

      • zas
        so LGTM for both
      • 2022-05-03 12338, 2022

      • yvanzo
        Ok, thanks!
      • 2022-05-03 12324, 2022

      • yvanzo
        I updated MB sitemaps to use rudi too, I will run again at :30, will check it then.
      • 2022-05-03 12351, 2022

      • yvanzo
        "cron job will run"
      • 2022-05-03 12311, 2022

      • yvanzo
        reosarevok: opened a PR, please review if that it is understandable enough :)
      • 2022-05-03 12349, 2022

      • adhawkins
        I think my slave container setup has corrupted its database.
      • 2022-05-03 12321, 2022

      • adhawkins
        Is it likely to be rescuable, or shall I just download a database dump and start again?
      • 2022-05-03 12337, 2022

      • reosarevok
        yvanzo: checking, let me see
      • 2022-05-03 12316, 2022

      • yvanzo
        adhawkins: how corrupted?
      • 2022-05-03 12351, 2022

      • adhawkins
        yvanzo: My replication was throwing errors. However, I've just kicked it off again and it seems to be proceeding. Will monitor it.
      • 2022-05-03 12342, 2022

      • BrainzGit
        [listenbrainz-server] 14alastair merged pull request #1978 (03master…metadata-cache-check-exists): Check if the canonical_release_redirect table exists before running https://github.com/metabrainz/listenbrainz-server…
      • 2022-05-03 12342, 2022

      • BrainzGit
        [listenbrainz-server] 14alastair merged pull request #1979 (03master…metadata-cache-docs): LB-1079: Fix mb_metadata_cache update methods to work with bulk table code https://github.com/metabrainz/listenbrainz-server…
      • 2022-05-03 12324, 2022

      • odnes has quit
      • 2022-05-03 12359, 2022

      • yvanzo
        adhawkins: Can you share some errors through pastebin so I can check these?
      • 2022-05-03 12332, 2022

      • adhawkins
        If I see any more, I will yvanzo.
      • 2022-05-03 12343, 2022

      • adhawkins
        It was something along the lines of duplicate index entries I think.
      • 2022-05-03 12353, 2022

      • adhawkins
        Sorry, I appreciate that's not particularly helpful :)
      • 2022-05-03 12320, 2022

      • outsidecontext
        rdswift: Hi, I merged picard-docs master to release for new 2.8.0rc1, and pages build as far as I can see. But does not show up on the website. Any idea why? Usually that worked rather quickly, or is this just GH pages being slow with updating? gh-pages branch looks good to me
      • 2022-05-03 12301, 2022

      • reosarevok
        yvanzo: hmm. as of now, upgrade.sh fails with ERROR: schema "dbmirror2" does not exist
      • 2022-05-03 12318, 2022

      • outsidecontext
        zas: are you around later for publishing the website or shall we do it now already?
      • 2022-05-03 12333, 2022

      • zas
        I'll be around
      • 2022-05-03 12355, 2022

      • reosarevok
        So I guess the upgrade script needs to include CREATE SCHEMA dbmirror2;
      • 2022-05-03 12309, 2022

      • outsidecontext
        ok, I'll ping you. I'm a bit AFK for now
      • 2022-05-03 12331, 2022

      • rdswift
        outsidecontext, I have no idea. It looks like the build went okay. Let me check further.
      • 2022-05-03 12310, 2022

      • mayhem
        yvanzo: I think lucifer can help you with some of the things you might need help with for the schema change. can you please get lucifer up to speed on what you need help with?
      • 2022-05-03 12300, 2022

      • atj
        zas: any objection to adding "metabrainz.org" as a search param in the netplan DNS configuration?
      • 2022-05-03 12350, 2022

      • outsidecontext
        rdswift: Probably just a GitHub issue. I know it takes a few minutes for changes to be available. But this time it is unusually long, so I wondered if maybe made some mistake
      • 2022-05-03 12317, 2022

      • HenryG has quit
      • 2022-05-03 12324, 2022

      • outsidecontext
        But if it looks ok to you maybe just wait a bit longer. I need to wait for the snap builds to finish anyway :)
      • 2022-05-03 12308, 2022

      • rdswift
        Actually, it looks like it might be something with the 404 remapping code. It doesn't seem to be doing the version check properly. I'm looking at it now.
      • 2022-05-03 12303, 2022

      • HenryG joined the channel
      • 2022-05-03 12322, 2022

      • reosarevok
        yvanzo: pushed that. next issue: ERROR: trigger "restore_collection_sub_on_public" for relation "editor_collection" already exists
      • 2022-05-03 12343, 2022

      • reosarevok
        We've already added that on prod, so I guess I need to DROP TRIGGER IF EXISTS it first :)
      • 2022-05-03 12300, 2022

      • yvanzo
        Yes
      • 2022-05-03 12320, 2022

      • yvanzo
        This is what Iā€˜m changing in sir triggers too by the way
      • 2022-05-03 12337, 2022

      • reosarevok
        I wasn't even asking, I just keep talking in a rubber-ducky way for some reason :D But I'll send a PR with that + changing schema to 27 in upgrade.sh, then we can merge that already
      • 2022-05-03 12344, 2022

      • BrainzGit
        [musicbrainz-server] 14reosarevok opened pull request #2508 (03schema-change-2022-q2…schema-change-2022-q2-amend-1): Small fixes for schema change https://github.com/metabrainz/musicbrainz-server/…
      • 2022-05-03 12347, 2022

      • reosarevok
        yvanzo: ^
      • 2022-05-03 12334, 2022

      • reosarevok
        I'm sure there'll be more, but
      • 2022-05-03 12309, 2022

      • zas
        atj: nope
      • 2022-05-03 12307, 2022

      • reosarevok
        Heh
      • 2022-05-03 12314, 2022

      • reosarevok
        CREATE OR REPLACE TRIGGER only works on psql 14
      • 2022-05-03 12316, 2022

      • reosarevok
        Whoops
      • 2022-05-03 12345, 2022

      • reosarevok
      • 2022-05-03 12342, 2022

      • yvanzo
        reosarevok: DROP TRIGGER IF EXISTS is what I use for sir: https://www.postgresql.org/docs/12/sql-droptrigge…
      • 2022-05-03 12354, 2022

      • reosarevok
        Yeah, we should use that there too
      • 2022-05-03 12307, 2022

      • reosarevok
        Once we start requiring PG 14, CREATE OR REPLACE TRIGGER looks neat, but
      • 2022-05-03 12355, 2022

      • elomatreb[m]
        Question about the redesign: Is there a central place where one can view mockups (if there are any)? I'm interested especially in the release page
      • 2022-05-03 12328, 2022

      • rdswift
        outsidecontext, I just downloaded the repo and loaded it into a test site https://picard-docs-dev.rsds.ca site and it seems to work as expected.
      • 2022-05-03 12314, 2022

      • rdswift
        Looks indeed like it is a problem with GitHub pages.
      • 2022-05-03 12303, 2022

      • rdswift
        I wonder if adding the new batch of files pushed it over some sort of limit?
      • 2022-05-03 12351, 2022

      • zas
      • 2022-05-03 12352, 2022

      • rdswift
        Looks like we might have hit their 1GB limit for the site.
      • 2022-05-03 12351, 2022

      • v6lur joined the channel
      • 2022-05-03 12351, 2022

      • rdswift
        Time to find a different hosting option?
      • 2022-05-03 12346, 2022

      • akshaaatt
        Hi elomatreb[m]! https://www.figma.com/file/ln8XiFLit634KC3YkUW2RS… serves as the link for the firm redesign. As of now the release page redesign was initiated by aerozol on the community forum but we'll look to have it added to figma soon.
      • 2022-05-03 12308, 2022

      • elomatreb[m]
        thanks!
      • 2022-05-03 12325, 2022

      • elomatreb[m]
        I was thinking maybe we could show a colored world map (like the one you get in the statistics on Listenbrainz) instead of the huge list of flags and dates we currently show for imported digital releases
      • 2022-05-03 12345, 2022

      • elomatreb[m]
        (I'm resigned to this style being accepted by the community)
      • 2022-05-03 12349, 2022

      • yvanzo
        rdswift: How does it come this repository takes 1GB in the first place?
      • 2022-05-03 12325, 2022

      • yvanzo
        zas: answered, thank you
      • 2022-05-03 12358, 2022

      • rdswift
        Mainly because of the images in the documentation and it saves each version number separately for each supported language (readthe docs type format).
      • 2022-05-03 12339, 2022

      • rdswift
        I just checked the size on my local copy and it's just over 2GB
      • 2022-05-03 12335, 2022

      • outsidecontext
        but the pages themselves shouldn't be that large really
      • 2022-05-03 12349, 2022

      • outsidecontext
        maybe the PDF / epub downloads
      • 2022-05-03 12327, 2022

      • outsidecontext
        I'd expect to see some message in github preferences or such, but don't see anything. really strange
      • 2022-05-03 12328, 2022

      • yvanzo
        maybe optimizing images (color palette, zopflipng) could also help with moderating the overall size?
      • 2022-05-03 12320, 2022

      • rdswift
        The problem is that each image is stored for each version and language, so reducing image size should buy us some time.
      • 2022-05-03 12300, 2022

      • rdswift
        What would be really nice would be to find a way for sphinx to only save one copy of an image.
      • 2022-05-03 12302, 2022

      • outsidecontext
        overall image size for the latest version is 8 MB, don't think this will gain us much. we have 6 versions with 2 languages
      • 2022-05-03 12344, 2022

      • atj
        elomatreb[m]: not sure how that would work to be honest. However I think aerozol has taken it upon himself to try and come up with a solution, so it might be worth discussing it with him.
      • 2022-05-03 12346, 2022

      • outsidecontext
        maybe time to consider moving to readthedocs, don't know
      • 2022-05-03 12349, 2022

      • yvanzo
        It can probably be automated though (Git pre-commit hooks, or GitHub Actions).
      • 2022-05-03 12321, 2022

      • elomatreb[m]
        atj: I posted in the release redesign thread on the forums
      • 2022-05-03 12336, 2022

      • yvanzo
        8MB to 1GB, if you reduce image size by half, it saves a lot overall.
      • 2022-05-03 12338, 2022

      • elomatreb[m]
        Although I don't see what would be hard about technically, since you can just map country codes to some specific SVG element?
      • 2022-05-03 12318, 2022

      • yvanzo
        Obviously it won’t solve the hosting issue in the longer term.
      • 2022-05-03 12329, 2022

      • rdswift
        outsidecontext, you're right. Maybe images aren't the problem. I'll do some more digging.
      • 2022-05-03 12304, 2022

      • outsidecontext
        rdswift: if it is about repository size we might be out of luck with GH pages without purging e.g. commit history
      • 2022-05-03 12317, 2022

      • Shubh has quit
      • 2022-05-03 12320, 2022

      • outsidecontext
        because it might be the actual repository size that is at fault https://stackoverflow.com/questions/28706454/gith…
      • 2022-05-03 12321, 2022

      • outsidecontext
        ok, but we probably don't get a quick solution today, so let's skip this for the 2.8rc1 release. not absolutely mandatory for that
      • 2022-05-03 12337, 2022

      • outsidecontext
        zas: could you publish the website?
      • 2022-05-03 12337, 2022

      • outsidecontext
        rdswift: what we could try is completely squashing the gh-pages branch into a single commit
      • 2022-05-03 12359, 2022

      • outsidecontext
        I wouldn't care about it's history much
      • 2022-05-03 12318, 2022

      • rdswift
        Agreed that might make the difference. Of the 2G in the directory, the .git directory eats up 1.5G
      • 2022-05-03 12343, 2022

      • rdswift
        Want to give it a try?
      • 2022-05-03 12346, 2022

      • outsidecontext
        rdswift: shall I try that?
      • 2022-05-03 12352, 2022

      • outsidecontext
        yes
      • 2022-05-03 12308, 2022

      • alastairp
        is this from a github action?
      • 2022-05-03 12318, 2022

      • rdswift
        Go for it. If we need to recover, I have the full history in my clone of the repo.
      • 2022-05-03 12350, 2022

      • alastairp
        there might also be ways to tell it to do a sparse checkout? (I wasn't following and am not sure if the problem is during build, or when trying to push your existing repo)
      • 2022-05-03 12304, 2022

      • outsidecontext
        alastairp: it's likely the actual repository size on github that hits quota limits for Github pages
      • 2022-05-03 12338, 2022

      • alastairp
        right, I just read your logs a bit more and saw that