samj1912: i deployed new solr version, the cluster looks ok, i'll wait few hours be concluding anything but it seems it doesn't disk read as much as previous version and response times are a tad better
ruaok
iliekcomputers: in theory we should be able to see the logs from the workers using something like this:
docker service logs yarn-workers
but right now that says that the worker nodes are not available.
do you want to work through this, or do you want me to give a hand with it (I note that it doesn't cover any of the concerns that I wrote in the ticket, so I think it'll need quite a bit of discussion/development still)
reosarevok
Oh god. Trying to properly document an event, it sucks so much. Why did you let me write this.
bitmap: while converting the release editor, can you make sure as much as possible of the stuff is reusable? (if it isn't already). We really need per-performed-work proper data at some point :/
iliekcomputers
alastairp: be really helpful if you review =)
alastairp
ok, I've assigned myself, I'll try and write some notes on it this week :)
do we know this contributor? I see it's their first PR. GCI student? potential SoC? Someone you know?
iliekcomputers
Yeah, good friend of mine, potential summer of Code
:)
alastairp
ah! great :)
ruaok
good to know. I had the same sort of reaction. :)
iliekcomputers
Well, soc is a long way away still, let's see.
Slurpee has quit
Slurpee joined the channel
c1e0 has quit
D4RK-PH0ENiX has quit
george
thefar8[m]: That report is sweet
D4RK-PH0ENiX joined the channel
slurpee- joined the channel
c1e0 joined the channel
Slurpee has quit
thefar8[m]
Okay Thank you Freso (sorry for the late response, got a lot things to do today)
george: Thanks !
bitmap
reosarevok: which parts of the release editor would apply to works?
reosarevok
I was thinking the UI bits would be fairly relevant (since "artist, work" seems to be reasonably close to "artist, track name") - only that I guess with an option to find the works in MB rather than just a string name
Freso
thefar8[m]: Well, my response was even later. :)
reosarevok
But we would need to think more specifically about how to store the info in a way that can be reused and queried, because that's the worst part about the current setlists
You can't really get "this song was played on these concerts" at all
(we could just use normal relationships for that, maybe ordered, but then that would just duplicate half the event setlist info)
bitmap
right, missed the previous line where you mentioned events :)
TOPIC: MetaBrainz Community and Development channel | MusicBrainz non-development: #musicbrainz | GSoC https://goo.gl/7jsjG2 | Meeting agenda (17th is last meeting of 2018!): Reviews, Google Code-in (Freso), mini summit plans (iliekcomputers)
TOPIC: MetaBrainz Community and Development channel | MusicBrainz non-development: #musicbrainz | GSoC https://goo.gl/7jsjG2 | Meeting agenda (17th is last meeting of 2018!): Reviews, Google Code-in (Freso), mini summit plans (iliekcomputers), Holidays/next meeting (Freso)
reosarevok
bitmap: do you have any good ideas about storing setlists in a way that allows for them to be integrated properly, like the rest of MB is?
Without adding extra entities, that is :D adding them it's easier
bitmap
add a setlist table? or is that cheating
all it needs is (event, work, artist_credit, name) right?
, position
if we had n-ary relationships we might be able to abuse those instead, but
c1e0 has quit
c1e0 joined the channel
yvanzo
reosarevok: an idea mentionned earlier is to handle setlist similarly to tracklist
slurpee- has quit
c1e0 has quit
reosarevok
Ideally we'd be able to also add relationships (orchestra, conductor, etc) to each entry, but I guess that'd be a good start that could be made into something deeper more easily later too if desired
Also would need to allow for string-only comments and whatnot for the equivalent of the current # but that's probably not too hard either
iliekcomputers
ruaok: `docker service logs yarn-workers -f` works 😅
probably a bug in docker...
ruaok
wait, what?
-f in the wrong place?
sorry, that was going to be the next thing on my list. still chasing people to pay us.
iliekcomputers
yeah, no worries =)
yvanzo
reosarevok: by "string-only comments", you mean annotation :)
we crossed the $500,000 income line sometime mid december.
iliekcomputers
i don't think the placement of the `-f` matters, `docker service logs yarn-workers` doesn't work, adding a -f does.
>we crossed the $500,000 income line sometime mid december.
ruaok
now back to real business!
iliekcomputers
noiiice
ruaok
that does sound like a docker bug
CatQuest
[14:48] <reosarevok> I was thinking the UI bits would be fairly relevant (since "artist, work" seems to be reasonably close to "artist, track name") - only that I guess with an option to find the works in MB rather than just a string name
I'd except that the *relationship* editor would make more sence for work?
i mena i'd love to see an extention of it basiclaly with a separate column with all the stuff for works as well (but for every work thing o an artist)
well i guess for a release
thefar8[m]
Freso: :) I've sent my response on that
ruaok
iliekcomputers: so, need anything from me now?
iliekcomputers
no, i'll go back to seeing why the dataframes aren't being written.
CatQuest
oh this is for setlists
ruaok
let me know if you want a second set of eyes, iliekcomputers
iliekcomputers
ok.
CatQuest
yea probably liek a tracklsit would work then
iliekcomputers
so remember we talked about how there's a RPC interface and an HTTP interface for HDFS
but a setlist doesnt have to be a list of recordings (thy might not be recorded or even if not in the database (but be added) or it *can* be recordings)
iliekcomputers
and the RPC interface works from inside the hadoop-master container.
ruaok
but not outside.
iliekcomputers
but not from inside other containers for some reason.
yeah.
CatQuest
or sohuld they always be as recordings becasue they *existed* see like radio-brodcatsts
ruaok
do you know what port numbers need to be open for this?