#metabrainz

bitmap: ping

zas: pong

i just noticed something in ernie nginx ocnf

since rob isn't around, i wonder if you can confirm ;)

here we go

On Ernie (nginx conf):

conf/sites-available/critiquebrainz: listen 72.29.166.148:443;

conf/sites-available/metabrainz-full: listen 72.29.166.148:443;

conf/sites-available/metabrainz-test: listen 72.29.166.148:443;

But the ip block is 72.29.166.152/29, so i guess those are errors and should be replaced with 72.29.167.148:443

how come this wasn't detected before ? do i miss smt ?

hmm

i'm currently working on the gateway chef, and i discovered that after nginx complained about it, and prod servers are using the same setup i'm using on my VMs to develop failover gateways chef cookbooks

I think you're right that they're typos...

it means https is half broken (since nginx cannot listen on this address), it seems weird no one complain

is it actually broken, or does it just ignore the bad address?

it ignores it, since this ip isn't configured on any address

s/address$/interface/

so it only listens on 72.29.166.157:443

i wonder in which cases the server is hit with the 167.148 address for those hostnames, ie. critiquebrainz.org points to 72.29.166.157 so it doesnt happen for conf/sites-available/critiquebrainz

right, that's what I was wondering

apparently beta.musicbrainz.org is one

yes, https://lb1.musicbrainz.org/

but it works...

telnet 72.29.167.148 443 is answering, so something is handling it...

i don't think we have a case where it actually fails

isn't that correct though? 72.29.167.148 is in our public range, as opposed to 72.29.166.148

I think we can just remove 72.29.166.148 from those 3 files...

it is what i'm thinking about atm, it makes no sense to listen on those addresses for those hostnames anyway, and since it is broken since ages without any visible impact it can hadly hurt

but i wonder what was the initial purpose

i'll wait for ruaok to give his own view on that before making changes

ok, sounds good

Leftmost: ping!

...

Freso: I don't think he's here

I'll use the time to drop Leo and email and see if he's doing OK

Let's go!

Hey

o/

Just trying to find Rob's mail so he can go first.

Though I guess someone else can go while I'm looking.

zas, make it so.

Well, it will be quite short

I mainly worked on getting gateway chef functionnal, and it is more complicated i initially though

thought*

That's what you say every week, before you rant on for three hours. ;)

Basically i have to rewrite each cookbook, and test the whole thing on VMs using a special setup to emulate all aspects of the production gateways

i made significant progress but more work is needed to have something production-ready

I also fought once with search servers this week

they went mad for some yet-unknown reason, and when it happens this is a real mess to get them back up

i have few questions for bitmap about that

Did you see my question about browser stats?

yes, i didn't answer since i don't know what to do yet

:)

I guess that also means delaying agenda item #2 for next week then. ;)

i guess i can set up a webalizer-like thing and process logs

what was the question about browser stats?

I think we have Google Analytics somewhere though, maybe.

about IE9 ... well i would just ignore this shit ;)

Alright, let's revisit the browser talk when we're through the reviews then.

If we can avoid to process logs and GA can do the job ... i would say let's use GA

bitmap: ^

zas: fin?

Yes, and sadness all the weekend

Mhm. :/

Rob says:

On monday I met up with Lego (the Wikimedia fellow who upgraded our wiki) and generally chatted with him and also chatted about the Taio Cruz situation with other wikimedians. Then I went schmoozing at the SF Music Tech pre-event.

On tuesday I attended SF Music Tech and banged the MetaBrainz drum. (not a hard job these days)

On wednesday I met with Google and realized a long term contract is moving from the knowledge graph to youtube. That’s generally a win for us. I also met with Pandora and they will start using our data. Sadly this doesn’t translate to more money since Next Big Sound already belongs to them. But it is another big name for us to put on our site. In the evening I met up with Caller and ViaPanda for

beers.

Thursday was mostly a travel day and friday was spent doing some online work, but mostly pounding out a presentation for the conference tomorrow.

/Rob says

Between travelling the MA/Boston wastes and doing a lot of Google Code-in work, I haven't gotten much else done in the last week.

But hey, we got accepted into the GCI! \o/

We could really use some mentors familiar with the MBS code if anyone is out there... :)

fin; go bitmap.

ok, so monday & tuesday I was still in SF for the gsoc thing and then traveling back

once I got back I prioritized fixing the remaining issues the Google people had with our JSON-LD markup

bitmap: I would love to have you, but I'm not sure if Rob wants to put hours towards towards this (or how many), and I have no idea how much of a time sink this will be, so I just assumed there were no one to start with and worked from there. :) I would love to have you in on the action though, go ahead and ping me your Google account if you're game.

based on the most recent emails from them it looks like they're ready to sign off on things

Yay! \o/

the other big thing I worked on after that was a UI for reporting users to admins

I saw that one. It's looking spiffy from the screenshots. Will be good to see it in "RL" action.

the PR is at https://github.com/metabrainz/musicbrainz-serve... if anyone wants to look

It won't actually :p

It'd be nicer to never have to ;)

But yeah, it looked useful :)

other than that I accidentally fixed a GCI ticket and some small beta-only issues

fin.

This week I've been doing more work on the relationship editor. I really wanted to get that done today, but it looks like I won't be able to :( However, I hope to finish it off on Wednesday, and do a release containing that and the refactored code then. I'll also aim to update our dependencies to the latest versions.

Plan then is to finish off the remaining feature for the November release, the reputation interface, and get that released in a mini-release next Monday.

I've also added some GCI tasks, which I'll be mentoring with Leftmost - these cover tweaks to our UI, plugins for Calibre and starting the style guidelines. I've also actually installed Calibre now, just to make sure my demands weren't completely unreasonable ;)

I won't have much time for the December release, so we may have to throw out some stuff, but we'll see about that...

fin.

reosarevok: It's definitely nicer to have it and have all the ugly things surface, than have a lot of vitriol and guile hidden away and just continually seeding more of them same.

LordSputnik: Who goes next?

Gentlecat!

(or reosarevok)

hello

Gentlecat, then :P

Freso: that reminds me, should we look into "moving style and the like to templates" being GCI tasks?

didn't have time to do much last week, unfortunatly

I don't remember whether we agreed we wanted to do it or not :)

primarily helped Freso with GCI stuff

fixed issue with dataset validation in AcousticBrainz that was causing all kinds of postgres errors on the whole website

(this is still not merged though, waiting for alastairp's review)

and reviewed a couple of pull requests for MBS

fin. reosarevok 💨

I didn't do a huge lot. I found some GCI tickets, I added some more instruments, and I edited quite a bit (but somehow still didn't catch up to drsaunde again)

reosarevok: Ask me under agenda item #3 ;)

:D

Ok

That was about it for this week really

Who's left? chirlu`? alastairp?