ruaok: can you setup the tunnel again? So that we can have a look at the UI. Maybe it can help, not sure though. I guess you took it down while testing spark networks
sarthak_jain has quit
ruaok
you do know that I documented how to put up/take down the tunnel, no?
docker/tunnel/README.md
I'll start it now, but know that you can do it yourself in the future.
tunnel is up and running.
Nyanko-sensei has quit
deltafarce joined the channel
deltafarce
Hi, not to bug, but is #musicbrainz the right channel to ask musicbrainz-related code queries in?
Or is it here? I'm confused now lol
ruaok
here is good for code questions, deltafarce
deltafarce
Ah, okay. I've spammed the other channel, lol
brb
'm new to Postgres, and have been working with the MusicBrainz dump in a Flask webapp, so my question is: like the MusicBrainz site, is there a good programmatic way to hop from a sub-area (e.g. Beverly Hills) to its country (United States)?
I'm using the Postgres dumps in a Flask webapp, with SQLAlchemy.
From what I could dig, 'musicbrainz.l_area_area' is where relationships between the various areas is stored.
This looks like a many to many relation table, but I'm not sure how I can get the code to pick that up.
ruaok
I don't personally know much about areas. reosarevok, yvanzo?
it might also be hard to do with sqlalchemy. ORMs tend to not be good at such complicated things.
deltafarce
Hmm, I'm aware. All I want to do is substitute the city/municipality area codes I'm getting with their appropriate country ones.
Because it's annoying to have a mix of countries, cities, and districts when linking artists to their areas.
iliekcomputers
ruaok:
spark performance is very degraded
we haven't been able to ascertain why
ruaok
spark or the VM? we know those VMs suck ass.
iliekcomputers
i thought it might be the VM.
ruaok
I think it has to do with the fact that our instances are idle, so they get "paged out". and paging them back in take a while.
iliekcomputers
but i have no way to confirm that
this has been going on for a couple of hours though.
ruaok
which is exactly why I want to have a dedicated cluster that we control.
its probably been going on for ages. and complaining to hetztner does nothing since they do not guarantee the performance of these VMs.
Nyanko-sensei joined the channel
Nyanko-sensei has quit
iliekcomputers
I remember that the thing we're trying to do used to take like 40 min when I did it last time (probably months ago). and it is so slow now that it's hard to even estimate how long it'd take.
we thought it might be some config changes that had been made recently, but reverting them didn't help.
Nyanko-sensei joined the channel
ruaok
what if we spin up 4 machines at azure, which has guaranteed performance? we'd need to change the tunnel to lemmy IPs, but this gives us a solution for TODAY.
iliekcomputers
makes sense to me. If the problem still exists, then we know that it's spark and not the VM
ruaok
ok, let me get on that.
iliekcomputers
thank you!
ruaok
once I get VMs running, do you want to do the docker install from there?
or should I do that?
c1e0 joined the channel
iliekcomputers
i think you doing it might end up being faster
or if there's docs that I can follow, I'll do it.
sarthak_jain joined the channel
ruaok
ok, let me get started and see where things go. gonna take a while though.