#metabrainz

moooin!

lucifer: zas: what is the status of aretha / willaims?

mo’’in’

morning

jasondk: cool, thanks!

ruaok: from what i see, barman has been moved but everything else is on williams.

zas: can we work on this today, please>

lucifer: what is the status with the dumps and importing data to spark from the 15th?

sentry's clear and the dumps also imported fine yesterday. so i think it was a one time connectivity issue.

let's generate the parquet dumps for testing?

yes, let's. I'll get that moving.

we're still running old cron right?

yes

k

we should do one full dump and 2 inc dumps. so that we can test all 3 scenarios. full dump import, first inc dump import and subsequent inc dump imports.

ok

ruaok: also, how do continous aggregates work? do they refresh counts for current year or only older than 1 year?

let me adjust the dummy dump code.

👍

the cont aggs are currently set to refresh only 1 year automatically. then the cron job wakes up and processes 2 or 3 years in one go, until all time is done.

this keeps the disk space usage low.

makes sense, due to some reason counts on my profile have been broken for ~3 days now. i was hoping the cron job would fix it but it didn't.

ok, I'll look into that in a bit once I get the spark stuff going

like that?

thanks.

yes looks good.

ruaok: moooin, yes, I'm transferring ftp data atm, we can start to move services, machine looks stable

zas: cool and you can take care of the ssh keys? do we need to install new ones or reuse the existing ones?

ssh keys for ? I guess we can copy them over

for LB and other servivces to copy things to FTP.

would be ideal if they could be copied over.

lucifer: dumps started.

lucifer: do you know if the listen counts are broken for only you or for all people?

ruaok: they are broken for at least me. i checked a few other profiles and it looked okay for them. for me the counts were 0 (now ~100) so i noticed.

have you deleted your listen count key in redis and then gone back to your profile page to force a recount?

redis should be the first place to look at this.

no i haven't done that yet. because i thought it might be useful to know why this happened in the first place.

i'll try forcing the recount.

its going to be hard to figure out why -- we should really replace the redis stuff with the solution that I suggested -- keeping the counts in TS and be done with it.

(without relying on cont agg)

agreed. +1

ok, to get back into the swing of things and since i like pain, I'll add the listen_mbid_mapping table to the dumps.

I suppose that ought to be a new dump file...

yeah, new file makes sense.

forcing recount fixed the issue.

ok, keep your eye on it and see if it gets borked again.

👍

I'm doing the same.

hello, buenos dias

back to it this week

hiya!

guten morgen!

lucifer: my listen counts are incrementing.

hi!

https://www.adventurebikeracing.com/threepeaksb... that looks fun

> Distance of the Race in 2021 is around 2450km which makes it possible to finish it in around a week.

ruaok: yes, the listen counts increment for me as well. the issue is my listens were around 8000 last week, then those reset to 0 due to some reason. from then they kept on incrementing as i listened to ~100 today.

yeah, sounds like your redis key was lost. for whatever reason.

possibly.

alastairp: those rules. ouch.

winner rolled in to arc de triomf 2 days ago

after 6 days of riding

sheeesh. thats freaking insane.

~400km a day? over mountains? insane.

I guess some people like doing that to themselves...

did they get a PCR test for the borders?? 🤭

[musicbrainz-android] 14amCap1712 merged pull request #78 (03master…onboarding): Onboarding Tweaks https://github.com/metabrainz/musicbrainz-andro...

question for react experts: as a matter of style do you always have exactly 1 component per file, or if you have small components that are only used as a helper in a larger one will you include them in the same file?

im trying to submit this listen, but im getting an error 500, any ideas?

500? hmmm, lemme look/

i think its because the json has escape characters

I doubt it -- then you'd get a 400 error.

did you retry submitting it?

it may be a transient error.

no i didnt retry, but i have tried sending the submission a couple times

```Error: unhandled exception: ERROR 500 Data requested not available [HttpRequestError]```

thats the error

please submit it again, so I can look for the error in the logs. tell me right after you submit.

done

the user id is tandy1000

if that helps

alastairp: I'm not a purist, so if the helper component is small and not used anywhere else I think it's acceptable to keep it in the same file as the main component.

If there are more helper components I usually create a separate helpers file for them.

oh hai monkey !

Hai !

tandy[m]: I see the error. let me look closer.

feeling back to 100%, monkey ?

Yep, all good now

great.

lucifer: what do you make of tandy[m] 's error? https://sentry.metabrainz.org/metabrainz/listen...

the listen pasted is looks ok, but we can't parse it.

I'm going to deploy PR #1539 to test.LB unless anyone is is actively using it.

Some BrainzPlayer improvements

alastairp: same as monkey for MBS

@ruaok the json is generated by some unmarshallling code, so i might have missed something in my custom types

the JSON looks ok to me, but somehow we get an array parsed from the JSON, not a dict.

hmm i used this guide to write the types for which a submissin payload is generated

ruaok: I see the JSON starts with a curly bracket, and I think maybe the code expects an array instead

Would that be possible?

Not sure why that would result in a 500, but…

shouldn't be, no. That would make our docs wrong. :)

Ignore me :)

thanks for looking. :)

I can't see the whole JSON, was trying to see if there could be some issue with the formatting

Can you see the whole JSON somewhere ruaok ? I only see a truncated version

this https://pastebin.pl/view/4817772d ?

Thanks

It's anoying that we can't see the whole data in the sentry issue

sorry had fallen asleep. 😅

i have seen that error before tandy[m]. my hunch is on an encoding issue. let me look deeper.

lucifer: Hi! I have another one for your plate :) There are some uncommitted changes in docker-server-configs on kiss that I don't want to mess up

monkey: hi! just took a look, you can safely discard that. those are probably from ruaok updating cron earlier today for testing.

OK, thanks :)

tandy[m]: your entire json is wrapped inside extra `''`. so its not a correct json payload. i am not sure of the source of those but place would be to look into your listen submission code.

*first place

Hi lucifer ! Thanks for reviewing the PR. I am available in case any discussions are to be made.

I even got some good feedback from the community regarding the design, which I intend to incorporate soon.

akshaaatt[m]: hi! sounds nice. i am trying to test your picard addition PR again and it does not seem to work this time. can you try again and see if it works for you.

ruaok: the alert you set up seems to working nicely ;)

Sure!

oh good, I hadn't had a chance to test it yet.

did you stop the metric writer to test it? or did something else happen?

its hung up itself. i just came to know of it from the alerts.

ok, that begs the question: why does it stop?