#metabrainz

/

      • jesus2099
        Freso bitmap if someone is here, when I load a page, the expires of each image is slightly in the past (Expires: Fri, 25 Sep 2015 14:26:43 GMT) or is the now second… this is why there is no cache at all
      • 2015-09-24 26719, 2015

      • jesus2099
        (i think)
      • 2015-09-24 26715, 2015

      • jesus2099
        it’s now 16:29 here and it says :// —  — Expires: Fri, 25 Sep 2015 14:29:30 GMT Cache-Control: max-age=86400
      • 2015-09-24 26738, 2015

      • jesus2099
        i guess my time zone 16:29 equals that 14:29 GMT
      • 2015-09-24 26753, 2015

      • jesus2099
        expires is supposed to be in the FUTURE in order to cache work properly
      • 2015-09-24 26712, 2015

      • jesus2099
        but maybe it’s my company firewall doing some funny stuff (only on mb.org though)
      • 2015-09-24 26719, 2015

      • jesus2099
        i’ll check from home and open a ticket if same
      • 2015-09-24 26716, 2015

      • zas joined the channel
      • 2015-09-24 26725, 2015

      • CatQuest
        I've got 644 mb memory left. I've got to quit this process, sorry guys
      • 2015-09-24 26713, 2015

      • alastairp
        whoops
      • 2015-09-24 26737, 2015

      • alastairp
        our next step is to make it continue an import where you finished the last one
      • 2015-09-24 26743, 2015

      • alastairp
        so that you can stop and start again
      • 2015-09-24 26713, 2015

      • ruaok
        any clue why it leaks so much ram, alastairp ?
      • 2015-09-24 26737, 2015

      • alastairp
        no
      • 2015-09-24 26755, 2015

      • alastairp
        at a guess we’re keeping hold of a handle and it’s not GCing
      • 2015-09-24 26730, 2015

      • alastairp
        but, you know. javascript
      • 2015-09-24 26710, 2015

      • marc2k3 has quit
      • 2015-09-24 26736, 2015

      • jesus2099 has quit
      • 2015-09-24 26734, 2015

      • JesseW joined the channel
      • 2015-09-24 26749, 2015

      • ruaok
        with this silly log of mine, I've learned a few interesting things, alastairp
      • 2015-09-24 26757, 2015

      • ruaok
        it makes lots of log files. \ø/
      • 2015-09-24 26713, 2015

      • ruaok
        in more useful knowledge, it favors a very few pids.
      • 2015-09-24 26723, 2015

      • ruaok
        in some cases only one ever.
      • 2015-09-24 26725, 2015

      • alastairp
        that’s like a koan
      • 2015-09-24 26736, 2015

      • alastairp
        yeah, which is what uwsgitop seemed to indicate
      • 2015-09-24 26758, 2015

      • ruaok
        those few, are using cpu, regardless if there are requests or not.
      • 2015-09-24 26703, 2015

      • alastairp
        right
      • 2015-09-24 26716, 2015

      • alastairp
        do we want to go down from 50-20 again?
      • 2015-09-24 26719, 2015

      • alastairp
        -> 20
      • 2015-09-24 26727, 2015

      • ruaok
        so they are not stuck procs as we had considered.
      • 2015-09-24 26738, 2015

      • ruaok
        sure, I just dont know that that would help
      • 2015-09-24 26759, 2015

      • alastairp
        we weren’t seeing highly loaded processes when we only had 20, right?
      • 2015-09-24 26724, 2015

      • ruaok
        I dont think we could tell them apart.
      • 2015-09-24 26728, 2015

      • ruaok
        but sure, lets try it.
      • 2015-09-24 26721, 2015

      • github joined the channel
      • 2015-09-24 26721, 2015

      • github
        [listenbrainz-server] mayhem pushed 1 new commit to master: https://github.com/metabrainz/listenbrainz-server…
      • 2015-09-24 26721, 2015

      • github
        listenbrainz-server/master d316af7 Robert Kaye: Update listenbrainz.ini.uwsgi...
      • 2015-09-24 26721, 2015

      • github has left the channel
      • 2015-09-24 26747, 2015

      • ruaok
        ps aux | grep uwsgi | wc -l
      • 2015-09-24 26747, 2015

      • ruaok
        43
      • 2015-09-24 26730, 2015

      • diana_olhovik has quit
      • 2015-09-24 26701, 2015

      • kepstin-laptop joined the channel
      • 2015-09-24 26702, 2015

      • ruaok
        alastairp: first observation. the requests are spread over a greater number of backends
      • 2015-09-24 26746, 2015

      • JesseW has quit
      • 2015-09-24 26716, 2015

      • reoafk is now known as reosarevok
      • 2015-09-24 26734, 2015

      • ruaok
        now we're idle and we have 5 uwsgi procs taking cpu time.
      • 2015-09-24 26726, 2015

      • reosarevok
        ruaok: did you see my question about subscriptions?
      • 2015-09-24 26757, 2015

      • ruaok
        I didn't sorry
      • 2015-09-24 26718, 2015

      • reosarevok
        Just wondering if we know why subscription emails came several hours late today and yesterday
      • 2015-09-24 26731, 2015

      • reosarevok
        (I got it at almost 11 AM today)
      • 2015-09-24 26753, 2015

      • ruaok
        I have no idea, sorry.
      • 2015-09-24 26723, 2015

      • reosarevok
        Ok
      • 2015-09-24 26730, 2015

      • reosarevok
        bitmap: ^ when you're around
      • 2015-09-24 26717, 2015

      • reosarevok
        Oh. Someone mentioned http://www.oreilly.com/data/free/data-and-music.c… in the forum (apparently talks about MB too)
      • 2015-09-24 26721, 2015

      • reosarevok
        Anyone know if it's worth a read?
      • 2015-09-24 26756, 2015

      • ruaok
        ah, he was supposed to let me know when it was published.
      • 2015-09-24 26705, 2015

      • Mineo
        somehow that question made me think of the german intro to the pink panther: https://www.youtube.com/watch?v=JRL5Z1k60tg (the title translates to something like "who played around with the clock?")
      • 2015-09-24 26732, 2015

      • reosarevok
        Maybe it just came out, the forum person posted today
      • 2015-09-24 26740, 2015

      • madmouser1
        oh man i mucked up my github repository :-(
      • 2015-09-24 26701, 2015

      • reosarevok
        (oh no, a tweet from Sep 10 mentions it already)
      • 2015-09-24 26755, 2015

      • ruaok
        feh. I wanted to add spotify links to our listens page and I just looked at what renders that page.
      • 2015-09-24 26758, 2015

      • ruaok
        FFS.
      • 2015-09-24 26712, 2015

      • ruaok closes tab and gives up on LB/MB/MeB for today
      • 2015-09-24 26730, 2015

      • reosarevok
        Is the code so bad or?
      • 2015-09-24 26700, 2015

      • ruaok
        no, I'd need to learn another three js frameworks or something. just to add one link.
      • 2015-09-24 26729, 2015

      • reosarevok
        Oh heh
      • 2015-09-24 26718, 2015

      • outsidecontext has quit
      • 2015-09-24 26721, 2015

      • zas
        madmouser1: just git reset --hard upstream/master (where upstream is the remote for musicbrainz/picard-plugins) and git cherry-pick 2da0673, then push the branch on your pr branch (PR/37), it should be clean the mess you have ;)
      • 2015-09-24 26705, 2015

      • reosarevok
        ruaok: is the main LB repo the github or is it BB like for MBS?
      • 2015-09-24 26736, 2015

      • madmouser1
        zas: thanks let me read and try to understand the logic before I mess up further
      • 2015-09-24 26737, 2015

      • yeeeargh joined the channel
      • 2015-09-24 26721, 2015

      • Lotheric has quit
      • 2015-09-24 26745, 2015

      • samphippen has quit
      • 2015-09-24 26747, 2015

      • Lotheric joined the channel
      • 2015-09-24 26713, 2015

      • regagain joined the channel
      • 2015-09-24 26739, 2015

      • UmkaDK has quit
      • 2015-09-24 26756, 2015

      • Gentlecat
        ruaok: there's a pull request that converts it into jinja2 template
      • 2015-09-24 26705, 2015

      • Gentlecat
        you can add it there if you want
      • 2015-09-24 26727, 2015

      • Gentlecat
        I can try to finish it in a few days
      • 2015-09-24 26710, 2015

      • Freso
        alastairp: What is the reasoning behind the MBIDs going into that huge pile of "whatever data you want to feed us!" instead of elevating them a level and designate them specifically (even as optionals), just to ensure consistency in "tag naming"? (E.g., to avoid confusion between spaces/non-spaces, sending as lists vs. sending as strings, etc.)
      • 2015-09-24 26753, 2015

      • Freso
        jesus2099: Open a ticket about it?
      • 2015-09-24 26726, 2015

      • kepstin-laptop
        we really should have a list of standard fields so that apps are consistent, yeah
      • 2015-09-24 26734, 2015

      • kepstin-laptop should help out on the audioscrobbler-compatible listen api
      • 2015-09-24 26742, 2015

      • Freso
        kepstin-laptop: Yes. :)
      • 2015-09-24 26757, 2015

      • marc2k3 joined the channel
      • 2015-09-24 26734, 2015

      • zas has quit
      • 2015-09-24 26721, 2015

      • kepstin-laptop
        I think we'd probably want to have something like three levels of fields, 1) a minimum set of data that's required or the listen is just discarded, 2) "standard" optional data fields with consistent names that are used by listenbrainz in some way, 3) other fields that are just stored but not used.
      • 2015-09-24 26744, 2015

      • lenniepaine joined the channel
      • 2015-09-24 26704, 2015

      • kepstin-laptop loves that you still can't scrobble (to last.fm) a track by the artist 'MOSAIC.WAV' unless your scrobble includes an mbid.
      • 2015-09-24 26752, 2015

      • kepstin-laptop
        since it would hit the spam filter according to the artist name (field matches "*.wav" which looks like a filename), but that's overridden by the mbid presence :)
      • 2015-09-24 26707, 2015

      • jesus2099 joined the channel
      • 2015-09-24 26710, 2015

      • jesus2099
        Freso: you’re here ?
      • 2015-09-24 26730, 2015

      • jesus2099
        can you confirm if you see the MB banner reloading each time you load an MB page ?
      • 2015-09-24 26736, 2015

      • Mineo
        jesus2099: if you read the expires headers you posted earlier correctly then you'll see that the data expires tomorrow, not today :)
      • 2015-09-24 26747, 2015

      • jesus2099
        and the EXPIRE property of the RESPONSE for let’s say the header logo png ?
      • 2015-09-24 26757, 2015

      • jesus2099
        tomorrow ?
      • 2015-09-24 26728, 2015

      • Mineo
        yes: "jesus2099| it’s now 16:29 here and it says :// —  — Expires: Fri, 25 Sep 2015 14:29:30 GMT Cache-Control: max-age=86400" today is the 24th
      • 2015-09-24 26731, 2015

      • jesus2099
        oh yes !
      • 2015-09-24 26734, 2015

      • jesus2099
        thanks !
      • 2015-09-24 26747, 2015

      • jesus2099
        so it definitely comes from my new company crap called zscaler
      • 2015-09-24 26707, 2015

      • jesus2099
        all HTTPS responses are denied from being cached or somethign…
      • 2015-09-24 26726, 2015

      • jesus2099
        (for instance now no HTTPS/MB links are marked as visited any more)
      • 2015-09-24 26727, 2015

      • kepstin-laptop
        ouch, your company is intercepting https? :(
      • 2015-09-24 26739, 2015

      • jesus2099
        thanks !
      • 2015-09-24 26750, 2015

      • jesus2099
        yes, why not ? ;)
      • 2015-09-24 26751, 2015

      • jesus2099 has quit
      • 2015-09-24 26750, 2015

      • kepstin-laptop
        https://www.kepstin.ca/dump/abzterm.png that time I ran abzsubmit
      • 2015-09-24 26718, 2015

      • zas joined the channel
      • 2015-09-24 26704, 2015

      • lenniepaine has quit
      • 2015-09-24 26731, 2015

      • regagain has quit
      • 2015-09-24 26752, 2015

      • oliverl joined the channel
      • 2015-09-24 26726, 2015

      • Freso
        jesus2099: I'm always here, even if I'm not.
      • 2015-09-24 26711, 2015

      • Freso
        jesus2099: Also, you do realise that while I'm employed by MeB now, I'm not employed as a coder/programmer, right? I wouldn't even know where to begin poking at that header in the MBS codebase. :)
      • 2015-09-24 26709, 2015

      • Freso
        kepstin-laptop: I don't get it?
      • 2015-09-24 26700, 2015

      • reosarevok has quit
      • 2015-09-24 26704, 2015

      • kepstin-laptop
        Freso, it printed out really really long lines when it ran on my system, so I made the window wide enough to avoid wrapping
      • 2015-09-24 26716, 2015

      • kepstin-laptop
        required 3 monitors side by side to display that ;)
      • 2015-09-24 26750, 2015

      • CatQuest would have just taken 3 screen caps and pasted thme together
      • 2015-09-24 26703, 2015

      • diana_olhovik joined the channel
      • 2015-09-24 26704, 2015

      • diana_olhovik joined the channel
      • 2015-09-24 26721, 2015

      • alastairp
        Freso: musicbrainz ids are required to be with a specific key. it’s just a “whatever you want to feed us” because it’s a dictionary. We could have had 2 dictionaries, but we just have one with a set of known keys
      • 2015-09-24 26722, 2015

      • alastairp
        kepstin-laptop: minimum set of data that is required is artist and title (separate fields), then you can give us mbids, and we’ll do stuff with those if they’re there (second level) and finally you can add any other fields you want
      • 2015-09-24 26724, 2015

      • diana_olhovik_ joined the channel
      • 2015-09-24 26730, 2015

      • diana_olhovik has quit
      • 2015-09-24 26728, 2015

      • MBJenkins
        Project musicbrainz-server_master build #193: SUCCESS in 19 min: http://ci.musicbrainz.org/job/musicbrainz-server_…
      • 2015-09-24 26728, 2015

      • MBJenkins
        Michael Wiencek: Fix call to generate_table_list in DeprecateRelationship* reports
      • 2015-09-24 26748, 2015

      • oliverl has quit
      • 2015-09-24 26739, 2015

      • github joined the channel
      • 2015-09-24 26739, 2015

      • github
        [listenbrainz-server] mwiencek opened pull request #41: Move Scrobble methods to prototype (master...scrobble-prototype) https://github.com/metabrainz/listenbrainz-server…
      • 2015-09-24 26739, 2015

      • github has left the channel
      • 2015-09-24 26750, 2015

      • alastairp
        bitmap: cool
      • 2015-09-24 26718, 2015

      • samphippen joined the channel
      • 2015-09-24 26730, 2015

      • MBJenkins
        Project musicbrainz-server_beta build #667: FAILURE in 18 min: http://ci.musicbrainz.org/job/musicbrainz-server_…
      • 2015-09-24 26730, 2015

      • MBJenkins
        * Michael Wiencek: MBS-8553: Internal server error when an automatic series reordering involves multiple of the same entity
      • 2015-09-24 26731, 2015

      • MBJenkins
        * Michael Wiencek: Cache area containments in memcached
      • 2015-09-24 26731, 2015

      • MBJenkins
        * Michael Wiencek: Use more descriptive variable name
      • 2015-09-24 26732, 2015

      • MBJenkins
        * Michael Wiencek: Fix call to generate_table_list in DeprecateRelationship* reports
      • 2015-09-24 26708, 2015

      • bitmap
        alastairp: idk if it'll solve the problems with ram usage, but can't hurt :)
      • 2015-09-24 26748, 2015

      • alastairp
        it was going up megabytes per lastfm page. I really hope fixing it isn’t that easy :)
      • 2015-09-24 26714, 2015

      • alastairp
        it’d be something like 10k per function if that was the case
      • 2015-09-24 26717, 2015

      • diana_olhovik joined the channel
      • 2015-09-24 26750, 2015

      • diana_olhovik_ has quit
      • 2015-09-24 26725, 2015

      • samphippen has quit
      • 2015-09-24 26748, 2015

      • kepstin-laptop
        if it's going that much per page, there might be some hanging references to data from the html pages loaded that can't be garbage collected :/
      • 2015-09-24 26725, 2015

      • kepstin-laptop
        I'd try reimporting my library, but I don't want to do that until I can get mbids ;)
      • 2015-09-24 26757, 2015

      • MBJenkins
        Yippee, build fixed!
      • 2015-09-24 26757, 2015

      • MBJenkins
        Project musicbrainz-server_beta build #668: FIXED in 19 min: http://ci.musicbrainz.org/job/musicbrainz-server_…
      • 2015-09-24 26757, 2015

      • MBJenkins
        Michael Wiencek: Fix table alias in query
      • 2015-09-24 26729, 2015

      • fractalizator has quit
      • 2015-09-24 26701, 2015

      • marc2k3 has quit