jeez Š§Š°Š¹ŠŗŠ¾Š²ŃŠŗŠøŠ¹ alias page is hella slow :P
2017-05-18 13808, 2017
saifulbkhan has quit
2017-05-18 13836, 2017
CatQuest
chatlog gurues: can I have a quick link to the chatlogs where we start talking about the timeline stuff? (where I say "anyone remember whne we got url" )
2017-05-18 13845, 2017
CatQuest
preferably in a PM since i
2017-05-18 13803, 2017
CatQuest
so I don't have to scroll because I'm looking through the conwo
I was just looking at that. I found a container that I cared about -- how do I find its host?
2017-05-18 13806, 2017
zas
in the cpu graph of it
2017-05-18 13846, 2017
zas
check the legend, list of hosts
2017-05-18 13844, 2017
ruaok
I'm looking to find where metabrainz is running. looking at the metabrainz section, I see no data, no server names.
2017-05-18 13845, 2017
zas
i guess it is because that's not the name of the container (metabrainz being an old name)
2017-05-18 13808, 2017
zas
was it renamed ?
2017-05-18 13840, 2017
zas
if yes, i need to add it to a whitelist (to prevent even more mess with temporary containers)
2017-05-18 13846, 2017
ruaok
unsure. Gentlecat?
2017-05-18 13802, 2017
Gentlecat
yes, it was renamed
2017-05-18 13807, 2017
zas
that's not perfect, because when you launch a container without specifying a name, a random name is created, and it clutters influx db with almost unusable data, so we went to whitelist containers to get stats from
2017-05-18 13822, 2017
zas
what is the new name ? i'll fix that asap
2017-05-18 13815, 2017
Gentlecat
prod.metabrainz iirc
2017-05-18 13834, 2017
ruaok
does that mean the deploy instructions have changed?
Your requests are being throttled by MusicBrainz because the application you are using has not identified itself. Please update your application, and see http://musicbrainz.org/doc/XML_Web_Service/Rate_Lā¦ for more information.
2017-05-18 13836, 2017
Rotab
or does a 403 mean that it is throttled completely?
2017-05-18 13807, 2017
Leo_Verto[m]
and you are setting a UA?
2017-05-18 13827, 2017
Rotab
i'm not setting anything
2017-05-18 13812, 2017
Rotab
it looks like they've fixed it. it just hasn't gotten released yet
I am trying to use the musicBrainz json-interface. However, I keep getting a lot of 503-errors (at least in 80% of the cases). However, pasting the same query to the web browser, is getting a lot more hits. Any ideas?
2017-05-18 13841, 2017
ruaok
ok, meeting done.
2017-05-18 13804, 2017
ruaok
Gentlecat: do I still push new metabrainz containers with
I think this should be a valid user agent header....?
2017-05-18 13858, 2017
ruaok
ok, are you waiting the required 1s between requests?
2017-05-18 13802, 2017
ruaok
yes, seems ok.
2017-05-18 13822, 2017
pere_
Currently I am testing the script manually. I have tried waiting much longer than 1 second. It can of course happen that I have accidentally clicked twice. How long is the ban if I violate 1 sec?
2017-05-18 13856, 2017
pere_
hmmm.... reading through my script here... just realised that when I am paging through multi-page results, I might violate the 1 second rule...
2017-05-18 13847, 2017
Mineo joined the channel
2017-05-18 13837, 2017
pere_
Implemented a quick fix, where I force php to sleep for two seconds after every query made to musicBrainz. To be 100% certain Im not accidentally violating the 1 sec rule. However. I still see mostly 503 errors.
2017-05-18 13840, 2017
arbenina_ joined the channel
2017-05-18 13803, 2017
pere_
Mostly 503-errors really means mostly 503-errors. I did run the script 5 times after the rewrite, and I did get 503-errors every time.
2017-05-18 13855, 2017
zas
pere_: servers are overloaded, you are hitting the global rate limit, nothing you can do. We currrently honor 350 req/s but incoming requests are ~3 times that
2017-05-18 13811, 2017
pere_
I also notice that all the errors I am getting from the php-script is: "Service Temporarily Unavailable". Running the same query from the web-browser, usually works, but when it fails it occasionally fails (less than 1:10) it says that the servers are busy.
Yes. I did look at that. Reading it again, I see that not waiting 1 second when paging through multi-page-results, might have put the ip-address of my server on the block-list.
2017-05-18 13826, 2017
hibiscuskazeneko joined the channel
2017-05-18 13834, 2017
zas
I see no reason why it would fail more with your script than in a browser, on our side requests are handled the same way. You can also read https://community.metabrainz.org/t/what-do-the-x-ā¦ (those may help you to diagnose the issue)
2017-05-18 13833, 2017
pere_
forgot to tell you that the server and the browser do not have same ip-address, so if there is a block on the ip of the server, that might affect it.
2017-05-18 13831, 2017
pere_
Are there any way to check if the server is added to a block list? Last half hour I have gotten zero successfull responses from the server script. Earlier today, I was able to get 3-4 valid responses, but then the drop rate was really high (80%++)
idk if it's been fully updated for the latest schema release
2017-05-18 13822, 2017
bitmap
looks like it was 11 hours ago, if you have the latest commit
2017-05-18 13840, 2017
bitmap
and yes, mbslave is fine & simpler if you don't need the entire musicbrainz-server running, just db access
2017-05-18 13808, 2017
boneskull
hmm, well, I want the API, so I need the server then, right? if I have setup REPLICATION_TYPE as RT_SLAVE, do I need to do anything?
2017-05-18 13840, 2017
boneskull
basically I just want a local version of this thing I can hit
2017-05-18 13812, 2017
arbenina_ has quit
2017-05-18 13813, 2017
boneskull
don't need to edit it, so sounds like a "mirror server" is what I want
2017-05-18 13842, 2017
bitmap
ok, then you might as well use the replication scripts that come with musicbrainz-server
2017-05-18 13856, 2017
bitmap
there's a admin/cron/slave.sh file in the repo that you can add to cron
2017-05-18 13807, 2017
boneskull
excellent, that's what I'm looking for, ty
2017-05-18 13821, 2017
samj1912 has quit
2017-05-18 13837, 2017
pere_
bitmap/boneskull: Maybe that is what I need as well...;) Im heavily debugging here, and reducing/waiting, but I keep ending up with 80% of the requests being dropped. Will the mbslave allow me to use the same php/json scripts for accessing a local version of the db?
2017-05-18 13859, 2017
bitmap
if you need a local version of the webservice/api, you need musicbrainz-server, not mbslave
2017-05-18 13810, 2017
bitmap
but yeah, you could then point your scripts to your local instance
2017-05-18 13839, 2017
pere_
Thanks Bitmap. Seems like a solution, but also a bit overkill for my use. I have a guitar website where I need to verify artistname/songnames. My guess is that I need 1000-2000 requests a day. Is there anywhere I can buy such access?
2017-05-18 13828, 2017
bitmap
there's no paid access to the webservice right now
2017-05-18 13853, 2017
bitmap
2000 requests a day doesn't seem like a lot though
2017-05-18 13807, 2017
SothoTalker_
10 cent per request? :D
2017-05-18 13854, 2017
SothoTalker_
the thing is, with all the userscripts that inline stuff.. they do use the webservice
we really need to handle music metadata education......
2017-05-18 13817, 2017
ricflomag joined the channel
2017-05-18 13821, 2017
ricflomag
Hi there ! Right now we have 288,096 CD stubs
2017-05-18 13801, 2017
ricflomag
Some of them have been waiting for approval since many years
2017-05-18 13838, 2017
ricflomag
Just one example : Only ! Nina Simone, published in 2010, submitted 6 years ago
2017-05-18 13830, 2017
ricflomag
Why not make CD stubs public ? They would still have a "draft" status, but people would be able to fetch them (and who knows, would be keen to edit them if neede)
2017-05-18 13856, 2017
ricflomag
Right now, the situation is a bit frustrating : you insert your CD, no data is fetched from musicbrainz but it *is* there, and you just have to copy and paste the date by hand :<
2017-05-18 13812, 2017
ricflomag
date -> data
2017-05-18 13808, 2017
SothoTalker_
i hate CD stubs :D
2017-05-18 13827, 2017
Quesito
ruaok: I don't want to scare you, but I feel like a warning is appropriate. I made an excel status sheet for the accounting transition status'. We should discuss maƱana.
2017-05-18 13844, 2017
Quesito
:D
2017-05-18 13808, 2017
CatQuest
ricflomag: you can actually use cdstubs when adding releases? use "add medium" and "use cd stub" it's a bit buried in menues
2017-05-18 13831, 2017
CatQuest
but I agree, it would be really neat f you could access cdstubs when adding via the CDTOC
2017-05-18 13838, 2017
pere_
one hour heavy debugging. Frustrating. Seem to be getting much more 503's querying 'release-groups' than 'artists'. Might be a coincident. Still only 10-20% goes trough, and thats is just too low to be useful.
2017-05-18 13834, 2017
ricflomag
CatQuest: I'm not very familiar with the submitting process and not sure I understand well :)
2017-05-18 13855, 2017
CatQuest
itsa steep learning cure, yes
2017-05-18 13800, 2017
SothoTalker_
CD stubs are public, i think
2017-05-18 13828, 2017
CatQuest
can't edit witotuh an mb user thoghu.. you have to be logged in to use it?
2017-05-18 13837, 2017
pere_
Something different: Have you guys ever tried putting Varnish in front of the web-service-server? It really rocks for stuff like this..;)
2017-05-18 13813, 2017
CatQuest
no, but we used shoepolish once :D
2017-05-18 13819, 2017
ricflomag
My point is : we have tens of thoushand of CD stubs which are not fetched by applications that connect to the database when looking for the metadata of the corresponding CDs
2017-05-18 13821, 2017
CatQuest
/reosarevokjoking
2017-05-18 13852, 2017
boneskull
bitmap I think it worked :D
2017-05-18 13805, 2017
bitmap
great!
2017-05-18 13834, 2017
SothoTalker_
CD stubs are not taken into consideration when you add releases the normal way, though
2017-05-18 13835, 2017
ricflomag
I guess this is because we want to have only approved quality metada fetched by end users
