#metabrainz

I have nothing to report for last week, was afk most of the time.

That's it I guess.

Thanks, back to reosarevok (IRC) then!

Hi!

Thanks everyone for your reviews

SolrCloud? (and very much not SOLR Cloud)

whats the diff?

It has to be upgraded

The cluster is now 100% resources almost all the times

So we need to migrate to increase the overall performance, and upgrade SOLR in the process

So is that why the search fails relatively often?

The idea is to deploy a set of VMs in pararel

reosarevok: yes

oh!

yvanzo & I can do most of the work on that I think, we'll need atj for Ansible support (if we go this way for it)

There are some resources available for deploying through Ansible already, just need to adapt.

The current cost of SOLRCloud is rather low (<200€/month), we can prolly drop our own load balancers since Hetzner has now a solution for that

Of course, more nodes will be needed, we'll evaluate that during testing phase

The other topic was about the need to upgrade docker on all machines (and rather soon)

Ok

zas: Servers reboot!

We need to organize that this week if possible

While at it, we'll upgrade consul too

so some downtime this week?

and reboot all servers

that suits me atleast fine that it's this week

ApeKattQuest: yes, likely, but it should be short

noproblem at all

Ok, say when and I'll be available

zas: ping lucifer (IRC) and I tomorrow during the day when we're both about and we can do the LB servers.

ok, what about tomorrow meeting time?

let me know when you'd like to reboot floyd and I can publish a blog post for the downtime

er, jimmy

sure works

I'm not around that time tomorrow.

mayhem[m]: sooner then?

but if lucifer (IRC) is willing to do it, then by all means.

bitmap: how long would the downtime be?

sooner works for me

Not sure if worth a blog if short?

Maybe just a tweet?

reosarevok: not sure, however long it takes to upgrade and reboot the server

upgrade is quick, reboot usually few mins at worse

What is still causing a downtime? PG?

mayhem[m], zas: after 12:00PM UTC anytime works for me.

PG

should be quick. but just in case

we can start during the day to upgrade servers that aren't critical (or are redundant) as usual

makes sense

k

ok, let's say 17 PM UTC?

bitmap: ^^?

worksforme

bitmap, can you also do the relevant changes to enable logical replication?

for sure

thanks!

I suggest we discuss details outside the meeting

reosarevok: back to you

So, that sounds like a wrap?

Ok, thanks!

So, thanks again everyone for attending

</BANG>

lucifer (IRC): Just a "heads up" that the topic in matrix doesn't seem to be getting updated when it changes in IRC. Will that be an issue regarding the meeting agenda?

Hmm, that's something we should try to bridge better :)

Thanks IRCers, but you’re missing all the fluff of using Matrix!

rdswift[m]: hmm, right. i think it works when you set if from matrix but not vice versa

hmm

is that why the topic got changed back and forth accidentally

Also, I'm not us normal users can change it in matrix.

* I'm not sure that us normal

I doubt that it might due to it, because changing the description of the room in Matrix requires privileges.

*be

what, are there OP in matrix?

what does matrix say for topic now

Might need to maintain the meeting agenda elsewhere?

akshaaatt: lucifer: I get the feeling the app YIM momentum has waned - are you okay with me sharing the news that we reran it akshaaatt? If you are planning another push for the app YIM it would be a good excuse to plug it, but it has been a while now!

hmm..

ApeKattQuest: #metabrainz MetaBrainz Community and Development channel | MusicBrainz non-development: #musicbrainz | BookBrainz: #bookbrainz | Channel is logged; see https://musicbrainz.org/doc/IRC for details | Agenda: Reviews, IRC alternatives (mayhem, lucifer), Summit 2024 (ruaok)

Looks like something has changed because I now have ability to edit it.

Apparently I can change it too from Matrix 😄

yvanzo[m], rdswift[m]: yes i changed the room topic permissions

but the topic isn't actually changing. only a lot of people it "so and such has changed topic to...."

I haven’t got the chance yet to debug the issues we were facing last time, aerozol. pranavkonidena[m was looking into it but it might take a few more days to release so we should release it separately

You can go ahead with the re run yim news

zas: bitmap: just checking for the meeting notes, downtime confirmed for 17 PM UTC tomorrow? Or not 100% sure yet

That’s a feint.

i thgouht it was lb

aerozol: I thought we are only rebooting redundant servers tomorrow, but not 100% sure

Okay thanks akshaaatt. If it’s a ‘few more days’ we can wait, if ‘a few more days’ = no idea when then I will share the news today ;)

aerozol (IRC): Time not confirmed yet, but not before 17:00 UTC

Thanks, I won’t confirm a time

We can say that there will be a downtime at some point after 17:00 UTC.

Currently the issue is at a no idea stage because pranavkonidena[m tried to debug it but couldn’t reproduce it

But there will be some tweet about it when confirmed I guess.

Got it thanks akshaaatt

:)

apologies if I wasn't paying attention, I can prepare the PG changes for tomorrow anyway

bitmap: Is there anything special to do/check regarding cache changes for today’s MBS release?

let me just double check that no DBDefs changes are needed...

our production DBDefs files do not use the new REDIS_*_DATABASE settings, but I guess that's not really an issue. (I'm not even sure it was worth adding these.)

yvanzo[m]: nothing special needed, then, but I will keep an eye on things after it's deployed

bitmap: I don’t see any change about REDIS_*DATABASE in the code.

in lib/DBDefs/Default.pm, I introduced three new settings

[design-system] 14Aerozol merged pull request #123 (03master…ListenBrainz-Local-logos): ListenBrainz Local logos https://github.com/metabrainz/design-system/pul...

bitmap: When? Not in `git diff beta..production`

we are talking about https://github.com/metabrainz/musicbrainz-serve... right? it hasn't been merged yet (nor approved)

Oops

Sorry, will review it again.

(Every time I’m thinking it’s already merged.)

all good

a problem with REDIS_CACHE_DATABASE and REDIS_STORE_DATABASE is that they are not used at all if you already have CACHE_MANAGER_OPTIONS, DATASTORE_REDIS_ARGS, etc. configured without them.

p.s. happy Waitangi Day everyone! Today Aotearoa pays respect to the treaty between Māori and Europeans, the founding document of ‘modern’ NZ

And to you aerozol !

Thanks monkey!

Reminds me, I had a q for you, that I accidentally pinged reo for instead

Thanks for all the notes on the LB recap document! I started writing it in current tense, which was fun, but about halfway through it had tied me in knots a little bit. If you think it is too confusing I can rewrite it to all be past tense. Otherwise I’ll have another pass over it to make sure it’s consistently current tense

^ monkey

Honestly I think it's fine in current tense

start in past tense, then segue into present and end in future !!!!

(joking)

That sounded, sounds and will sound like a great idea ApeKattQuest

[bookbrainz-site] 14dependabot[bot] opened pull request #1059 (03master…dependabot/npm_and_yarn/babel/preset-typescript-7.23.3): chore(deps-dev): bump @babel/preset-typescript from 7.16.0 to 7.23.3 https://github.com/metabrainz/bookbrainz-site/p...

💪

[bookbrainz-site] 14dependabot[bot] opened pull request #1060 (03master…dependabot/npm_and_yarn/elastic/elasticsearch-8.12.0): chore(deps): bump @elastic/elasticsearch from 5.6.22 to 8.12.0 https://github.com/metabrainz/bookbrainz-site/p...

[bookbrainz-site] 14dependabot[bot] closed pull request #1054 (03master…dependabot/npm_and_yarn/elastic/elasticsearch-8.11.0): chore(deps): bump @elastic/elasticsearch from 5.6.22 to 8.11.0 https://github.com/metabrainz/bookbrainz-site/p...

[troi-recommendation-playground] 14mayhem opened pull request #127 (03main…merge-content-resolver): Merge content resolver https://github.com/metabrainz/troi-recommendati...

lucifer: a big PR for ya -- looks scary, but most is copy-pasta from another reviewed repo, so not that much to review.

lucifer: thoughts on what to do with: https://github.com/nmslib/nmslib ?

the maintainer appears to be AWOL. I suppose we can clone and release on pypi, but that is meh.

mayhem: do you know of an alternative library that might work?

sadly, no. I was tickled pink to find this one.

but there are a lot of people in our position. this lib seems popular, so I hope that someone will pick it up.

i'll see if i can find any workaround than pypi.

maybe a gitsubmodule?

gah, I hate those.

yeah true but its better than releasing on pypi imo.

neither are great. :(

mayhem: pypi github issues suggest using github submodules