#metabrainz

/

      • Jigen joined the channel
      • 2025-08-12 22445, 2025

      • Zenigata has quit
      • 2025-08-12 22458, 2025

      • ApeKattQuest has quit
      • 2025-08-12 22446, 2025

      • Zenigata joined the channel
      • 2025-08-12 22424, 2025

      • pite_ joined the channel
      • 2025-08-12 22416, 2025

      • pite has quit
      • 2025-08-12 22432, 2025

      • Chris[m]12 joined the channel
      • 2025-08-12 22432, 2025

      • Chris[m]12
        Greetings! Does the inclusion of any streaming sevice and social media links is open?
      • 2025-08-12 22432, 2025

      • Chris[m]12
        There are some in my list.
      • 2025-08-12 22433, 2025

      • Chris[m]12
        Like Xiaohongshu, Showroom (showroom-live.com) and Line Music
      • 2025-08-12 22442, 2025

      • d4rkie has quit
      • 2025-08-12 22410, 2025

      • d4rkie joined the channel
      • 2025-08-12 22451, 2025

      • Chris[m]12
        Marshmallow (https://marshmallow-qa.com) and Mixi (https://mixi.social) as well :)
      • 2025-08-12 22441, 2025

      • bitmap[m]
        zas: thanks, I'm happy to move test.mb somewhere else, just not sure where...it used to run on aretha, but was unusable there (most requests would just time out)
      • 2025-08-12 22411, 2025

      • d4rkie has quit
      • 2025-08-12 22443, 2025

      • d4rkie joined the channel
      • 2025-08-12 22442, 2025

      • bitmap[m]
        only these nodes don't have a musicbrainz-website-prod container: aretha, cesaria, gaga, kiss, mingus, quest, rex, rudi
      • 2025-08-12 22439, 2025

      • Maxr1998_ has quit
      • 2025-08-12 22416, 2025

      • Maxr1998 joined the channel
      • 2025-08-12 22401, 2025

      • aerozol[m]
        Hi Chris! You can make tickets to add new social media and streaming service links. It's helpful if you include a bit of a blurb as to why it should be added. You can clone an existing ticket and then edit it to fit, like: https://tickets.metabrainz.org/browse/STYLE-2479
      • 2025-08-12 22401, 2025

      • BrainzBot
        STYLE-2479: Add audioculture.co.nz to the other databases whitelist
      • 2025-08-12 22448, 2025

      • Chris[m]12
        <aerozol[m]> "Hi Chris! You can make tickets..." <- Thanks! I'll do that soon!
      • 2025-08-12 22457, 2025

      • d4rkie has quit
      • 2025-08-12 22427, 2025

      • d4rkie joined the channel
      • 2025-08-12 22417, 2025

      • d4rkie has quit
      • 2025-08-12 22403, 2025

      • d4rkie joined the channel
      • 2025-08-12 22425, 2025

      • __BrainzGit
        [bookbrainz-site] 14dependabot[bot] opened pull request #1177 (03master…dependabot/npm_and_yarn/elastic/elasticsearch-9.1.1): chore(deps): bump @elastic/elasticsearch from 5.6.22 to 9.1.1 https://github.com/metabrainz/bookbrainz-site/pul…
      • 2025-08-12 22427, 2025

      • __BrainzGit
        [bookbrainz-site] 14dependabot[bot] closed pull request #1176 (03master…dependabot/npm_and_yarn/elastic/elasticsearch-9.1.0): chore(deps): bump @elastic/elasticsearch from 5.6.22 to 9.1.0 https://github.com/metabrainz/bookbrainz-site/pul…
      • 2025-08-12 22401, 2025

      • aerozol[m]
        Forums down? Getting HTTP ERROR 500 on https://community.metabrainz.org/
      • 2025-08-12 22418, 2025

      • aerozol[m]
        Guess meeting notes will have to wait :0
      • 2025-08-12 22433, 2025

      • suvid[m] joined the channel
      • 2025-08-12 22433, 2025

      • suvid[m]
        <aerozol[m]> "suvid: question for you btw..." <- Let me look into it in the evening today
      • 2025-08-12 22423, 2025

      • SigHunter has quit
      • 2025-08-12 22440, 2025

      • SigHunter joined the channel
      • 2025-08-12 22453, 2025

      • lusciouslover has quit
      • 2025-08-12 22442, 2025

      • lusciouslover joined the channel
      • 2025-08-12 22434, 2025

      • d4rkie has quit
      • 2025-08-12 22414, 2025

      • d4rkie joined the channel
      • 2025-08-12 22429, 2025

      • Zhele has quit
      • 2025-08-12 22434, 2025

      • Zhele joined the channel
      • 2025-08-12 22434, 2025

      • zas[m]
        mayhem: caa backup on mingus eats all disk space, causing forums to fail
      • 2025-08-12 22459, 2025

      • zas[m]
        I reclaimed a bit of disk space and killed caa_downloader.py
      • 2025-08-12 22448, 2025

      • zas[m]
        I'm thinking about getting a new machine to run musicbrainz-website-test (to isolate it), we could also use it for caa backup (which could be on a storage)
      • 2025-08-12 22426, 2025

      • zas[m]
        ping me when caffeinated enough
      • 2025-08-12 22415, 2025

      • d4rkie has quit
      • 2025-08-12 22441, 2025

      • d4rkie joined the channel
      • 2025-08-12 22406, 2025

      • mayhem[m]
        Shit, already? Sorry!
      • 2025-08-12 22422, 2025

      • mayhem[m]
        Coffee soon, just waking up
      • 2025-08-12 22456, 2025

      • mayhem[m]
        coffee in hand, not in veins. :)
      • 2025-08-12 22451, 2025

      • mayhem[m]
        CAA backup is going to need a LOT of disk over 10TB, IIRC. I think a hetzner storage box is required. so, if we connect that to mingus and never write to local disk the forums will never know.
      • 2025-08-12 22403, 2025

      • zas[m]
        It can work, getting a storage box is the way to go, I can set it up, and link it to mingus if you're ok
      • 2025-08-12 22434, 2025

      • mayhem[m]
        yes, but. Samba connection are unreliable. I've always had problems with them.
      • 2025-08-12 22443, 2025

      • mayhem[m]
        let me review what other options we have.
      • 2025-08-12 22456, 2025

      • mayhem[m]
        (for storage box conn)
      • 2025-08-12 22412, 2025

      • zas[m]
        ok, I was also thinking about getting a new machine
      • 2025-08-12 22428, 2025

      • mayhem[m]
        Bx41 is what we need
      • 2025-08-12 22429, 2025

      • zas[m]
        to separate mb test containers from prod servers
      • 2025-08-12 22409, 2025

      • mayhem[m]
        storage boxes have an HTTPS interface. I wonder if that is for storage as well as fetching.
      • 2025-08-12 22441, 2025

      • mayhem[m]
        go ahead and get the BX41 storage box.
      • 2025-08-12 22403, 2025

      • mayhem[m]
        if we do a SAMBA connection, is there a way to monitor it?
      • 2025-08-12 22425, 2025

      • mayhem[m]
        if hetzner does maintenance, it disconnects and then needs to be reconnected.
      • 2025-08-12 22430, 2025

      • mayhem[m]
        it feels brittle.
      • 2025-08-12 22449, 2025

      • zas[m]
        mingus out of disk space again, can you delete some stuff so at least forums still work
      • 2025-08-12 22421, 2025

      • mayhem[m]
        85% space now.
      • 2025-08-12 22436, 2025

      • d4rkie has quit
      • 2025-08-12 22458, 2025

      • lucifer[m]
        mayhem: it seems https access is only through webdav.
      • 2025-08-12 22404, 2025

      • d4rkie joined the channel
      • 2025-08-12 22414, 2025

      • mayhem[m]
        yeah, not great.
      • 2025-08-12 22425, 2025

      • lucifer[m]
        but there's also scp and rsync.
      • 2025-08-12 22431, 2025

      • mayhem[m]
        I need to periodically scan the collection and webdav would be PAINFUL
      • 2025-08-12 22453, 2025

      • mayhem[m]
        lucifer[m]: even painfuller.
      • 2025-08-12 22427, 2025

      • mayhem[m]
        I think SMB with some monitoring and we could be ok
      • 2025-08-12 22428, 2025

      • lucifer[m]
        should be possible to download a 1GB/1000 items on mingus then scp/rsync it. delete local dir and loop again.
      • 2025-08-12 22400, 2025

      • mayhem[m]
        yes. but doing a whole collection scan? ick.
      • 2025-08-12 22437, 2025

      • lucifer[m]
        but you wouldn't need to do a whole scan though?
      • 2025-08-12 22407, 2025

      • mayhem[m]
        I do. with the current design, yes.
      • 2025-08-12 22441, 2025

      • mayhem[m]
      • 2025-08-12 22446, 2025

      • mayhem[m]
        read the last two sections of the readme
      • 2025-08-12 22417, 2025

      • mayhem[m]
        rather than have a diff between MB DB and local cache DB, I replace the local DB, scan, continue.
      • 2025-08-12 22402, 2025

      • lucifer[m]
        ah okay that.
      • 2025-08-12 22433, 2025

      • lucifer[m]
        i think it might be better to use some other storage solution then.
      • 2025-08-12 22454, 2025

      • mayhem[m]
        what do you have in mind?
      • 2025-08-12 22430, 2025

      • lucifer[m]
        google cloud bucket is $40 a month for 10TB as well if you use coldline storage - good for infrequently accessed items.
      • 2025-08-12 22410, 2025

      • lucifer[m]
        and the interface would be much much nicer, plenty of dev tools around s3 compatible storages.
      • 2025-08-12 22450, 2025

      • lucifer[m]
        archive storage is even cheaper - $15 but slower to access.
      • 2025-08-12 22409, 2025

      • lucifer[m]
        ah but this is only storage cost, there will be retrieval costs when if actually want to use this backup too :/
      • 2025-08-12 22442, 2025

      • lucifer[m]
        what do you think about getting a couple of 10TB hard disks for the backup servers at office?
      • 2025-08-12 22415, 2025

      • mayhem[m]
        lucifer[m]: we already have 4 of those.
      • 2025-08-12 22418, 2025

      • mayhem[m]
        would it cost $400 to retrieve the 10TB with google cloud?
      • 2025-08-12 22439, 2025

      • mayhem[m]
        20TB at .02? yeah, no.
      • 2025-08-12 22452, 2025

      • mayhem[m]
        10TB flash cards are not really available yet.
      • 2025-08-12 22454, 2025

      • lucifer[m]
      • 2025-08-12 22426, 2025

      • lucifer[m]
        no egress fees.
      • 2025-08-12 22403, 2025

      • mayhem[m]
        but with per operation fees.
      • 2025-08-12 22442, 2025

      • mayhem[m]
        that sounds like an opportunity to get screwed.
      • 2025-08-12 22456, 2025

      • mayhem[m]
        only the storage box offers no hidden fees.
      • 2025-08-12 22442, 2025

      • lucifer[m]
        fair point, but with r2 we can actually serve the data on MB too if the time comes. would be harder to do with storage box.
      • 2025-08-12 22450, 2025

      • mayhem[m]
        ok, if I build a proper sync, that eliminates the re-scan, then S3 like storage becomes better. but if we ever need to verify the integrity, ouch.
      • 2025-08-12 22417, 2025

      • mayhem[m]
        lucifer[m]: the board of directors would not allow that until we have litigation insurance.
      • 2025-08-12 22427, 2025

      • lucifer[m]
        i see. makes sense
      • 2025-08-12 22428, 2025

      • mayhem[m]
        which sounds fuuuucking $$$$$$$$$$.
      • 2025-08-12 22449, 2025

      • mayhem[m]
        so, the worst case example is exporting it to a new server hosted by.... some volunteer.
      • 2025-08-12 22404, 2025

      • mayhem[m]
        or other org who can deal with this in a way we can't.
      • 2025-08-12 22454, 2025

      • mayhem[m]
        I think a properly monitored storage box is the right solution -- at least for now.
      • 2025-08-12 22403, 2025

      • lucifer[m]
        makes sense
      • 2025-08-12 22404, 2025

      • lucifer[m]
        btw just saw that hetzner also offers s3 compatible storage. but more expensive than storage box and egress costs.
      • 2025-08-12 22438, 2025

      • mayhem[m]
        yeah, S3 is not a great paradigm for this.
      • 2025-08-12 22411, 2025

      • lucifer[m]
        IA uses s3 for this so I do think they have the tools/etc. to do this properly.
      • 2025-08-12 22458, 2025

      • mayhem[m]
        Yeah, Brewster told me about an rsync script for this purpose, but he hasn't answered my emails since the hack.
      • 2025-08-12 22458, 2025

      • lucifer[m]
        and one final option, 22TB HDDs are 26eur so we could hook that up to one of the servers too.
      • 2025-08-12 22424, 2025

      • mayhem[m]
        I would not trust a 22TB for 26eur. temu special?
      • 2025-08-12 22432, 2025

      • lucifer[m]
        hetzner
      • 2025-08-12 22454, 2025

      • mayhem[m]
        oh. huh. raid-1 of that, I would.
      • 2025-08-12 22416, 2025

      • lucifer[m]
        makes sense
      • 2025-08-12 22422, 2025

      • mayhem[m]
        zas: the machine you want to get, can we get two 22TB drives on it?
      • 2025-08-12 22427, 2025

      • mayhem[m]
        that would be ideal!
      • 2025-08-12 22442, 2025

      • mayhem[m]
        it would still need to be able to connect to an MB mirror periodically.
      • 2025-08-12 22454, 2025

      • mayhem[m]
        so, depending on how isolated you want to make the new machine...
      • 2025-08-12 22404, 2025

      • mayhem[m]
        good catch, lucifer.
      • 2025-08-12 22453, 2025

      • zas[m] uploaded an image: (54KiB) < https://matrix.chatbrainz.org/_matrix/media/v3/download/matrix.org/YcSdInvUQBJuqWSAFzZJXrzZ/image.png >
      • 2025-08-12 22413, 2025

      • mayhem[m]
        bake it so, please
      • 2025-08-12 22450, 2025

      • zas[m]
        If ok to order, I will do it after lunch
      • 2025-08-12 22418, 2025

      • mayhem[m]
        despite me hating spinning disks, this is a very good solution.
      • 2025-08-12 22400, 2025

      • mayhem[m]
        can we monitor running services now?
      • 2025-08-12 22443, 2025

      • mayhem[m]
        I could have caa-backup put the latest fetched timestamp of images into a redis item. can we monitor redis items easily via our monitoring?
      • 2025-08-12 22430, 2025

      • mayhem[m]
        worst case, I'll start a thread that runs a simple HTTP server that serves that timestamp.
      • 2025-08-12 22416, 2025

      • zas[m]
        https://github.com/influxdata/telegraf/blob/maste… -> check redis.commands in the example
      • 2025-08-12 22427, 2025

      • zas[m]
        ## Optional. Specify redis commands to retrieve values... (full message at <https://matrix.chatbrainz.org/_matrix/media/v3/download/chatbrainz.org/xQWANiblNGDEJpgAAeHpVbEQ>)
      • 2025-08-12 22440, 2025

      • mayhem[m]
        fabu, will do.
      • 2025-08-12 22447, 2025

      • zas[m]
        So I guess we can manage to do that
      • 2025-08-12 22413, 2025

      • mayhem[m]
        yes. can string be a date?
      • 2025-08-12 22421, 2025

      • mayhem[m]
        which format is easiest?
      • 2025-08-12 22454, 2025

      • zas[m]
        What do you want to monitor exactly?
      • 2025-08-12 22427, 2025

      • mayhem[m]
        most important: most recently fetched timestamp. reason: can a died process or disconnceted storage box.
      • 2025-08-12 22437, 2025

      • mayhem[m]
        s/can/find/
      • 2025-08-12 22451, 2025

      • mayhem[m]
        for grafana monitoring, total count and fetch rate would be nice.
      • 2025-08-12 22403, 2025

      • mayhem[m]
        I can make an endpoint that covers all three datapoints.
      • 2025-08-12 22425, 2025

      • __BrainzGit
        [musicbrainz-server] 14reosarevok opened pull request #3606 (03master…MBS-14124): MBS-14124: Only try to merge each RG once https://github.com/metabrainz/musicbrainz-server/…
      • 2025-08-12 22426, 2025

      • BrainzBot
        MBS-14124: RG listed multiple times in merge-from-release if merging several releases from it https://tickets.metabrainz.org/browse/MBS-14124
      • 2025-08-12 22430, 2025

      • mayhem[m] heads to the office.
      • 2025-08-12 22458, 2025

      • zas[m]
        It isn't very convenient to work with dates as values, but we can have number of items fetched (absolute since start), so we can easily have a rate and check if no fetched items since a while (rate = 0 for a certain time -> alert)
      • 2025-08-12 22406, 2025

      • zas[m]
        feasible?
      • 2025-08-12 22448, 2025

      • zas[m]
        redis command count can work perhaps
      • 2025-08-12 22440, 2025

      • zas[m]
        mayhem: such config would take few days to set up, would 2x16Tb be enough (one can be set up in few minutes)?
      • 2025-08-12 22401, 2025

      • zas[m] uploaded an image: (80KiB) < https://matrix.chatbrainz.org/_matrix/media/v3/download/matrix.org/OhDbqkHOzeFwgYugdseYbqCj/image.png >
      • 2025-08-12 22406, 2025

      • zas[m]
        "Processing time: usually several working days"
      • 2025-08-12 22430, 2025

      • zas[m] uploaded an image: (82KiB) < https://matrix.chatbrainz.org/_matrix/media/v3/download/matrix.org/qtDkNLeXLGyURjmpfOSvFZgy/image.png >
      • 2025-08-12 22404, 2025

      • zas[m]
        If not urgent, I guess few days isn't an issue
      • 2025-08-12 22445, 2025

      • mayhem[m]
        how comfortable are we having the CAA backup without redudancy?
      • 2025-08-12 22457, 2025

      • mayhem[m]
        not urgent for me at all.