#metabrainz

/

      • zas
        ahah, it was less obvious but static files are slow too on critiquebrainz
      • 2015-12-06 34045, 2015

      • zas
      • 2015-12-06 34054, 2015

      • zas
        on chrome
      • 2015-12-06 34045, 2015

      • zas
        https://critiquebrainz.org/static/js/jquery-2.1.0… 29k in 1.21s , not that good either
      • 2015-12-06 34010, 2015

      • bitmap
        does the HTML download quickly though?
      • 2015-12-06 34038, 2015

      • creature has quit
      • 2015-12-06 34017, 2015

      • bitmap
        wget https://critiquebrainz.org/ -> 20.39K --.-KB/s in 0.003s
      • 2015-12-06 34034, 2015

      • bitmap
        it'd be more strange if it were only certain files
      • 2015-12-06 34041, 2015

      • zas
        144 MB/s for this one
      • 2015-12-06 34008, 2015

      • zas
        it seems to me that's only concerning static files
      • 2015-12-06 34018, 2015

      • zas
      • 2015-12-06 34036, 2015

      • zas
        215 711 6,83KB/s ds 23s
      • 2015-12-06 34000, 2015

      • zas
        good old modem speed
      • 2015-12-06 34028, 2015

      • bitmap
        let's see what happens if I change the file type, to rule that out
      • 2015-12-06 34007, 2015

      • chirlu`
        It’s faster for me. I get the Picard image in 1.5 s, the MB static.js in 2.2 s.
      • 2015-12-06 34031, 2015

      • chirlu`
        *common.js
      • 2015-12-06 34051, 2015

      • chirlu`
        The js is four times the size of the image, though.
      • 2015-12-06 34022, 2015

      • zas
      • 2015-12-06 34030, 2015

      • zas
        7 769 --.-K/s ds 0,002s
      • 2015-12-06 34030, 2015

      • Leo_Verto joined the channel
      • 2015-12-06 34042, 2015

      • mildused
        Can someone summarize what this conversation is about?
      • 2015-12-06 34049, 2015

      • bitmap
      • 2015-12-06 34008, 2015

      • bitmap
        got 1.00M 156KB/s in 5.9s
      • 2015-12-06 34024, 2015

      • zas
        lol ...
      • 2015-12-06 34032, 2015

      • zas
        eta 84s
      • 2015-12-06 34038, 2015

      • chirlu`
        9.6 s
      • 2015-12-06 34059, 2015

      • mildused
        I got 155kb/s at starbucks
      • 2015-12-06 34013, 2015

      • zas
        not finished yet
      • 2015-12-06 34029, 2015

      • zas
        max 13K/s
      • 2015-12-06 34034, 2015

      • chirlu`
        zas: Do you see packet loss when pinging picard.musicbrainz.org?
      • 2015-12-06 34053, 2015

      • zas
        1 048 576 12,7KB/s ds 91s
      • 2015-12-06 34029, 2015

      • mildused
        ping averaging ~70ms for me
      • 2015-12-06 34035, 2015

      • mildused
        50 bytes
      • 2015-12-06 34002, 2015

      • chirlu`
        Ping 188 ms for me, no packet loss.
      • 2015-12-06 34019, 2015

      • zas
      • 2015-12-06 34028, 2015

      • zas
        packet loss
      • 2015-12-06 34035, 2015

      • zas
        here is the issue
      • 2015-12-06 34057, 2015

      • zas
        but this isn't a temporary one
      • 2015-12-06 34004, 2015

      • bitmap
        I wonder if other websites located at DWNI would exhibit similar issues
      • 2015-12-06 34015, 2015

      • mildused
        Why not try their own site?
      • 2015-12-06 34014, 2015

      • mildused
        What host is this?
      • 2015-12-06 34017, 2015

      • bitmap
      • 2015-12-06 34058, 2015

      • zas
      • 2015-12-06 34029, 2015

      • mildused
        86 packets transmitted, 86 received, 0% packet loss, time 85127ms
      • 2015-12-06 34029, 2015

      • mildused
        rtt min/avg/max/mdev = 64.654/70.662/181.267/17.441 ms
      • 2015-12-06 34045, 2015

      • bitmap
        I think you should provide some traceroutes to rob so he can send it to them
      • 2015-12-06 34013, 2015

      • chirlu`
        From both ends, IIRC.
      • 2015-12-06 34056, 2015

      • zas
        i tested few other sites, no ping packet loss, much better rtt <25ms)
      • 2015-12-06 34058, 2015

      • mildused
        Any packet loss on an independent site?
      • 2015-12-06 34002, 2015

      • mildused
        Sniped me
      • 2015-12-06 34007, 2015

      • mildused
        zas, What about on a different network/ computer
      • 2015-12-06 34023, 2015

      • chirlu`
        bitmap: Could you check if the production DB has a primary key for the place table?
      • 2015-12-06 34012, 2015

      • bitmap
        chirlu`: "place_pkey" PRIMARY KEY, btree (id)
      • 2015-12-06 34048, 2015

      • chirlu`
        OK, so it got added manually at some point, I guess.
      • 2015-12-06 34016, 2015

      • chirlu`
      • 2015-12-06 34023, 2015

      • bitmap
        ah...
      • 2015-12-06 34021, 2015

      • zas
        ruaok: traceroutes & ping to send to DWNI : https://gist.github.com/zas/c16530732d81e35a8487
      • 2015-12-06 34009, 2015

      • zas
        ah search servers ....
      • 2015-12-06 34009, 2015

      • zas
        i do it
      • 2015-12-06 34028, 2015

      • ruaok
        Send then to
      • 2015-12-06 34011, 2015

      • ruaok is out and about.
      • 2015-12-06 34011, 2015

      • zas
        k
      • 2015-12-06 34027, 2015

      • zas
        search servers freakinbg out again
      • 2015-12-06 34033, 2015

      • YatharthROCK_ has quit
      • 2015-12-06 34029, 2015

      • Bookzombie has quit
      • 2015-12-06 34043, 2015

      • Bookzombie joined the channel
      • 2015-12-06 34034, 2015

      • YatharthROCK joined the channel
      • 2015-12-06 34052, 2015

      • YatharthROCK_ joined the channel
      • 2015-12-06 34010, 2015

      • YatharthROCK has quit
      • 2015-12-06 34006, 2015

      • zas
        search servers are back
      • 2015-12-06 34019, 2015

      • Leo_Verto has quit
      • 2015-12-06 34007, 2015

      • zas
        down again :(-
      • 2015-12-06 34022, 2015

      • mildused
        zas, What has been causing this recurring issue?
      • 2015-12-06 34018, 2015

      • zas
        crappy code, and no one to fix it + unfinished work on a new solution
      • 2015-12-06 34049, 2015

      • mildused
        zas, What is the new solution?
      • 2015-12-06 34016, 2015

      • zas
        SOLR, though it will be a solution after it was finished and uber tested
      • 2015-12-06 34047, 2015

      • mildused
        What is being used now?
      • 2015-12-06 34055, 2015

      • mildused
        Elastic?
      • 2015-12-06 34059, 2015

      • zas
        lucene
      • 2015-12-06 34000, 2015

      • mildused
        Oh no wonder. Lucene is supposed to be used for embedding search in desktop applications
      • 2015-12-06 34053, 2015

      • zas
        well, i dunno much about the history and reasons of that, but i'm very aware of the result atm ;)
      • 2015-12-06 34030, 2015

      • mildused
        Is the SOLR implementation on github?
      • 2015-12-06 34018, 2015

      • mildused
        Have you looked into ElasticSearch?
      • 2015-12-06 34042, 2015

      • ruaok
        mildused: elastic search is a horrible technology that will fail on you the second you rely on it.
      • 2015-12-06 34049, 2015

      • ruaok
        I strongly advise steering clear.
      • 2015-12-06 34056, 2015

      • ruaok
        zas: how are things? can I help?
      • 2015-12-06 34051, 2015

      • zas
        ruaok: ok for now, restarted
      • 2015-12-06 34056, 2015

      • zas
        nothing new, hard to say if stopping indexer was of any help, it could have been worse with (or no) ;)
      • 2015-12-06 34024, 2015

      • zas
        i'll restart it when load will start to decrease
      • 2015-12-06 34016, 2015

      • ruaok
        I want to believe it was better. :)
      • 2015-12-06 34033, 2015

      • zas
        :)) let's say it was !
      • 2015-12-06 34010, 2015

      • legoktm is now known as legoktm[NE]
      • 2015-12-06 34028, 2015

      • YatharthROCK_ has quit
      • 2015-12-06 34006, 2015

      • YatharthROCK joined the channel
      • 2015-12-06 34027, 2015

      • YatharthROCK has quit
      • 2015-12-06 34039, 2015

      • ruaok
        search again, I'll go
      • 2015-12-06 34013, 2015

      • MajorLurker joined the channel
      • 2015-12-06 34040, 2015

      • sibilant joined the channel
      • 2015-12-06 34007, 2015

      • sibilant has quit
      • 2015-12-06 34040, 2015

      • YatharthROCK_ joined the channel
      • 2015-12-06 34040, 2015

      • YatharthROCK_ has quit
      • 2015-12-06 34040, 2015

      • YatharthROCK_ joined the channel
      • 2015-12-06 34046, 2015

      • YatharthROCK_ has quit
      • 2015-12-06 34025, 2015

      • CallerNo6
        mildused, the solr setup is split into two GSoC projects.
      • 2015-12-06 34050, 2015

      • mildused
        CallerNo6, great..
      • 2015-12-06 34054, 2015

      • CallerNo6
      • 2015-12-06 34027, 2015

      • CallerNo6
      • 2015-12-06 34032, 2015

      • ruaok
        mildused: the assessment that lucene should be used for desktop apps is incorrect.
      • 2015-12-06 34035, 2015

      • CallerNo6
        er, I might have that wrong, but that's the idea
      • 2015-12-06 34044, 2015

      • ruaok
        solr and elastic search both use lucene
      • 2015-12-06 34051, 2015

      • mildused
        I meant that Lucene would have better time handling custom requirements for low level access to Lucene APIs
      • 2015-12-06 34021, 2015

      • YatharthROCK joined the channel
      • 2015-12-06 34021, 2015

      • YatharthROCK has quit
      • 2015-12-06 34021, 2015

      • YatharthROCK joined the channel
      • 2015-12-06 34030, 2015

      • ruaok
        the thing is that our use of lucense pre-dates SOLR and ES by years.
      • 2015-12-06 34017, 2015

      • Leftmost
        weeksio has been busy for the past while now, right? Is he still contactable for questions on it?
      • 2015-12-06 34008, 2015

      • ruaok
        he just dropped me an email that I haven't gotten to. lemme see
      • 2015-12-06 34007, 2015

      • ruaok
        > I should have done a B.A. :-/
      • 2015-12-06 34014, 2015

      • ruaok
        As opposed a to B.S.
      • 2015-12-06 34027, 2015

      • ruaok
        that was essentially the gist of it. :(
      • 2015-12-06 34037, 2015

      • ruaok puts out the proposal to hire weeksio for a semester.
      • 2015-12-06 34057, 2015

      • Freso
        bitmap | do static resources on other MeB sites also load slowly? -- http://tickets.musicbrainz.org/browse/MBH-413 :D
      • 2015-12-06 34009, 2015

      • Leftmost
        I'd like to get a sense of what's missing still and try to dig in to get it working. Gonna need to do it for BB too.
      • 2015-12-06 34015, 2015

      • ruaok
        Leftmost: I'd love that.
      • 2015-12-06 34023, 2015

      • ruaok
        let me know where to send choco/beer to!
      • 2015-12-06 34051, 2015

      • YatharthROCK has quit
      • 2015-12-06 34052, 2015

      • Leftmost
        My mouth, duh.
      • 2015-12-06 34058, 2015

      • YatharthROCK_ joined the channel
      • 2015-12-06 34017, 2015

      • ruaok wonders if the postoffice knows how to route that
      • 2015-12-06 34016, 2015

      • Leftmost
        Leftmost's Mouth, Montana I Think, United States
      • 2015-12-06 34037, 2015

      • LordSputnik
        Leftmost: pong
      • 2015-12-06 34036, 2015

      • CallerNo6 would like to shadow Leftmost if he's looking into solr.
      • 2015-12-06 34000, 2015

      • Leftmost
        LordSputnik, hey. Wanted to ask you about the button layout task. It's not clear from the description what you're looking for, and I was going to revise it and add some design tasks.
      • 2015-12-06 34018, 2015

      • LordSputnik
        bitmap: also, static stuff for BB is incredibly slow, I managed 4-10 seconds by eliminating a 1 MB SVG
      • 2015-12-06 34033, 2015

      • LordSputnik
        But our JS bundle is still taking 3-6 seconds to download
      • 2015-12-06 34048, 2015

      • LordSputnik
        Leftmost: go for it, it's the task I'm least happy with
      • 2015-12-06 34028, 2015

      • Freso
        drsaunders D4RK-PH0ENiX, please verify that you've gotten into the GCI mentor hangout. :)
      • 2015-12-06 34048, 2015

      • Leftmost
        What were you thinking for it? Would it make sense to turn it into "revise layout of entity display pages"?
      • 2015-12-06 34008, 2015

      • D4RK-PH0ENiX
        Freso: I got a lot of GCI mail list spam >_>
      • 2015-12-06 34018, 2015

      • Leftmost
        CallerNo6, I'd be happy for it. I suspect it's something I won't have a chance to dig into until later in the week, but I can ping you.
      • 2015-12-06 34022, 2015

      • LordSputnik
        I was going for "look at how users can perform actions in BB at the moment, and see if we could make it clearer what buttons do by moving them about a bit"
      • 2015-12-06 34003, 2015

      • Freso
        D4RK-PH0ENiX: The hangout is not a maillist though. :)
      • 2015-12-06 34012, 2015

      • LordSputnik is going to add some testing tasks
      • 2015-12-06 34019, 2015

      • Freso
        🙌
      • 2015-12-06 34026, 2015

      • Leftmost
        Okay. Well, I'd like three main design tasks for us, each with at least two instances:
      • 2015-12-06 34037, 2015

      • Leftmost
        Layout work on the main page (with specific criteria, such as displaying the requisite menus and branding, as well as providing discoverable actions and dynamic content based on what's happening in the database).
      • 2015-12-06 34056, 2015

      • LordSputnik
        the home page?
      • 2015-12-06 34017, 2015

      • ruaok registered staticbrainz.org
      • 2015-12-06 34033, 2015

      • Leftmost
        Layout work on entity display pages, with attention to clear presentation of data and indication of available interactions, especially editing.
      • 2015-12-06 34011, 2015

      • ruaok
        brainz.io is available.