So.. one more issue from my side and we're good to go for the next CB release!
Done!
Who's next?
Freso
🙌
!m code_master5
BrainzBot
You're doing good work, code_master5!
Freso
I'm next!
code_master5
Freso: Thanks!
Freso
Thank you everyone for reviews. No more people are up for reviews tonight, so we're moving on to the rest of the agenda items. First up:
Freso: Google Code-in
Mostly just a heads-up/announcement:
This is the last week for students to claim tasks in Google Code-in. Next Monday students will have two days to finish up any tasks, and then from Wednesday next week, we mentors have to pick the winners and finalists.
As in previous years, we have one task we'd like all students to do before the contest ends:
Any students reading here, please try and fit this task into your schedule. Shouldn't take more than 10-ish minutes, probably. :)
Finally: Good luck for the last stretch to all our students, and thanks for all your hard work so far!
fin.
ruaok: private data in BB dumps
TOPIC: MetaBrainz Community and Development channel | MusicBrainz non-development: #musicbrainz | GSoC https://goo.gl/7jsjG2 | Meeting agenda: private data in BB dumps (ruaok), Upgrade MB to Postgres 10? (ruaok), kinda-sorta libellous tags (reo)
george
👍👍👍
ruaok
as I examined the BB data dumping code, I noticed that the database is dumped with the exception of some usage stats tables.
those don't belong in PG, but that wasn't the issue.
BB tracks user birthdates, makes them public on the use profile page and then includes them in the dump.
since all the info is public, the dumps don't really create a problem.
but we collect birthdates which MB does as well. and MB shows only age, not birthdate to other users.
any thoughts on what we should do?
I'm for nuking birthdates very soon.
Leo_Verto
this sounds like it's a pretty big GDPR violation, sharing PID with basically every third party that downloads the dumps
ruaok
and then finding a way to fetch this data (or just age) from MB so we don't duplicate the data.
Mr_Monkey
How easy is it to get the age from MB profile?
Freso
Leo_Verto: Only if they've entered it into their public profile themselves.
The bigger issue, IMHO, is that it apparently isn't possible to delete it again.
ruaok
Leo_Verto: I dunno if it is or not.
Leo_Verto
Yes but even then they need to explicitly consent to that data being shared.
ruaok
yes.
and there was ZERO talk about GDPR related thinking on the DOB stuff.
can we just nuke it and be done?
bitmap
I don't see any reason for us to store DOB
ruaok
who strongly objects to this?
CatCat
what is DOB?
iliekcomputers
+1 to nuking.
reosarevok
Date of birth
ruaok
date of birth.
michelv joined the channel
reosarevok
I think having user age is great
For users who are willing to share that info
ruaok
reosarevok: +1 to that. and we already have that in MB.
Freso
Yeah.
CatCat
but yea date of birth . age is fine from mb
ruaok
let's not duplicate data.
reosarevok
If we can take it from MB, good
CatCat
yes
Freso
I do agree that we should try and centralise the user data somewhat.
CatCat
what is PID?
ruaok
ok, I hear no objections.
Freso
So I'm happy with dropping area, gender, and dob from BB.
reosarevok
I mean, we should have one MeB profile, that's the main issue :p
ruaok
Mr_Monkey: please expedite nuking DOB from BB.
Mr_Monkey
Does anyone know if it can be retreived from the MB profile?
CatCat
+1 reosarevok
reosarevok
In the meantime, can we actually get it from MB?
ruaok
Mr_Monkey: not yet.
but we can make reosarevok code us an internal API endpoint. :)
reosarevok
Because we shouldn't just drop it from BB if we have no clear plans how to get it back, that seems backwards
Drop it from the dumps instead
ruaok
I have a clear plan to get rid of a possible GDPR violation.
Freso
reosarevok: It's still (potentially) duplicated information.
CatCat
true
ruaok
and that should and does overrule everything else, reosarevok.
iliekcomputers
if it is a GDPR violation, we should remove it everywhere.
reosarevok sighs
CatCat
but tere are proably also peopel who have fileld out the bb ting but not the mb thing
Freso
And as ruaok says, the fewer places we have such potentially sensitive information, the better.
iliekcomputers
considering getting it back should be pretty easy anyways.
reosarevok
Whatever, get rid of it
ruaok
the current implementation is a GDPR issue, if not a violation.
Mr_Monkey: bake it so.
reosarevok
Sure, because it's in the dumps
ruaok
onward.
reosarevok
Which is stupid
But whatever
Your choice
Next
ruaok
postgres 10.
Freso
reosarevok: It's also displayed on their public profile.
ruaok
there is a requirement in BB to use PG 10.
reosarevok
Freso: obviously, because they entered it? Anyway, I'm not going to argue about this, the decision is made
ruaok
but that prevents us from moving BB to our infrastructure because in production we run 9.x. 9.5, right?
bitmap
yep
ruaok
so, I am strongly urging Mr_Monkey to find a workaround for the need of PG10.
but, we should consider when we're going to upgrade ourselves.
iliekcomputers
you can just use the pg10 instance on frank ?
Leo_Verto
Are you planning on using the same PG servers for MB and BB?
CatCat
how fast can mb get to PG 10 tho
ruaok
iliekcomputers: Leo_Verto: I would like to use only one set of DB servers if at all possible.
iliekcomputers
but idk, maybe we wanna keep the data consolidated, all other projects (except for AB) have data on bowie.
ruaok
AB it isn't possible.
iliekcomputers
yeah, makes sense.
ruaok
but iliekcomputers that is a decent compromise if we MUST got to PG10.
but pg10 is a single point of failure.
CatCat
I'd instead try to work towards MB PG 10
ruaok
more so than queen/bowie.
bitmap: what do you think of planning to do a PG upgrade in the fall?
CatCat
or atleast. like have an eta
ruaok
no schema change release, but a migration to PG10 and give people 6 months notice.
bitmap
MB already works run with pg 10 (though we should test replication/dumping more)
CatCat
can BB wait unti lthne?
bitmap
that would be great
ruaok
ok, this is mostly dealing with the coorindation of everyone downstream.
yvanzo
CatCat: you're welcome to do so ;)
ruaok
ok, lets put that into the plans then.
Freso
fin?
iliekcomputers
everyone downstream will have to move to pg10 too, right?
ruaok
I'll kick the process off and we should plan to make the move on Oct 15 (or close to it).
Mr_Monkey
For BB to wait means renaming entities on the ORM and web facing parts, but not renaming entities in the database, which would suck
ruaok
iliekcomputers: yes.
bitmap
pg11 would be even better :)
CatCat
yvanzo: huh?
ruaok
bitmap: ugh. hmmm.
iliekcomputers
is pg11 released yet?
ruaok
is there an ETA for pg11?
bitmap
it's released already, I thought?
Leo_Verto
it is
ruaok
oh. cool.
iliekcomputers
ugh, should have used it for frank then...
ruaok
might as well then.
lol
it should be nice and stable come fall.
fin.
thanks!
Freso
reo: kinda-sorta libellous tags
reosarevok
So
CatCat
hva skjer?
reosarevok
We got a case at support, a band who has been tagged as "nazi", "rac" and "twats"
It seems like there is a band of the same name which is *indeed* nazi and rac (and, I assume, also twats)
But that's not the one the tags ended up on
iliekcomputers
rac?
okie.
reosarevok
The band was asking us to remove the tags - they're now downvoted, but they still feel the fact someone can come to their page from the BBC and see a "nazi" tag, even if hidden, is damaging
ruaok
reosarevok: is there a threshold at which tags get hidden?
yvanzo
Freso: We keep improving!
reosarevok
ruaok: 0 or less votes
Freso
ruaok: 0
reosarevok
But you can still see them if you ask to, and they'll still be sent via the ws
