#metabrainz

hi

not much really. early last week I fell sick, and was mostly useless for the rest of the week.

ok again mostly since yesterday. I looked a bit what we need to do to get a picard 2 release out. zas let's chat about that separately

also yesterday started looking into the harmony tidal issue kellnerd mentioned. I have it mostly working, with some cases being a bit annoying with their changed API.

that's all . aerozol , you want to finish the round?

Heyo

Things heating up with work and life this week, so just got the usual stuff done. My misc tasks have built up so it still ends up being a bit (comms, discussions, tickets, style…)

I checked in on CB spam for the first time in ages and might have found an early contender for review of the year

Nothing else that stands out! Take us away reosarevok

Thanks!

And thanks all for attending

zas: are there news about "Hetzner mainboard repl" then?

But since they'll have a lot to do, they'll just alert us 4 hours before, and it starts in Feb

more details on https://docs.hetzner.com/robot/dedicated-server...

For us, it means we have to handle that on 4 servers: hendrix, jimmy, yamaoka, randy

That is rather critical ones.

We can order new servers (but we aren't certain about delays to get them)

given that these are our two DB servers, can we migrate to a new server without downtime?

given that we have two rack-slots right now, that might work ok.

Yes, if we can get servers in time, and they don't start by ours... ;)

well, I suspect that if we opt to get a new server, that they would hold off on fixing ours.

> separate email with a scheduled appointment time within a 4-hour window

This reads more like they will send an appointment ahead of time. The 4hour window refers to the appointment, not the notice

I'd hope so

Ah yes, right

bitmap: if we opt for new servers, downtime is very short right?

Doesn't change that we need a solid plan, but at least it won't be sprung on us

so, downtime or a slower migration. which do we prefer?

IIRC, we were able to perform a switchover in under 30 mins

hmmm, they may exchange mainboards faster than that

they'll have lots of practice!

(the switchover is very fast actually, stopping/restoring traffic to the DB is the slow part)

So, new servers or we accept a downtime?

Whatever is easiest on you / the tea?

in any case, 30 mins downtime min

Team*

if it's under an hour downtime seems acceptable and less work for us

New servers will cost more time, and we gain nothing (not like they will be faster, they will be the same)

I tend to agree

ok by me

We can perhaps switch between hendrix/jimmy anyway if they don't do both at same time/day

It seems unlikely.

We ordered these two servers with the same config at the same time.

holycow23[m]: Any ideas on what could be the reason behind this

Ok, solved then. BTW, it starts in Feb, but they say it can take a year long....

cant hurt to ask them for what we need and if they can't accomodate, we'll fall back

It would be great to be able to plan the day with them.

Well, they say -> https://docs.hetzner.com/robot/dedicated-server...

Yeah. I mean, how problematic is the issue? Might they fail entirely before they get to it possibly a year from now? :D

holycow23: we'll get to you after our meeting :) not long now

We don't know much about the exact issues

Anyway, we can move on.

Ok, if you're happy with that, let's finish with a

</BANG>>

... oops :)

* \</BANG>

Thanks again everyone!

Thanks reo!

My autocorrect wanted to call you "feo"

Oh, is anyone coming to Brussels for fosdem?

monkey: little kids and autocorrects never lie

reosarevok[m]: with any luck, lucifer will attend

zas: after reading their whole docs, it seems more reliable to order a new server

He told me :)

Wondering if anyone else is going to be in town for frites

yvanzo, bitmap: we should schedule a meeting :)

(not in Brussels, unless you want to)

A "call de Brusellas"?

reosarevok[m]: are you both available tomorrow?

I have a call at 15 UTC, I expect to be done by 16

After that, yes

zas: The main reason is that the replacement will happen at a random time on a random day, with a short time to get ready for it.

bitmap, reosarevok: 4pm UTC as usual works for me. Are you available too lucifer?

zas: something needs a kicking again: https://ci.metabrainz.org/job/musicbrainz-docke...

Guess it was just restarting the jenkins container

But it keeps breaking and it didn't use to, so that's something to investigate I guess?

Releasing prod

Done

bitmap: can you check the blog post?

(no huge huge rush, still doing docker)

"this release includes the usual small bugs" s/bugs/bug fixes/ maybe 😛

LGTM though

I mean....

But yeah, that's what I wanted to say. Sorry for being too honest!

yvanzo: yes.

[musicbrainz-docker] release 03v-2025-01-27.0 has been published by 14reosarevok: https://github.com/metabrainz/musicbrainz-docke...

[listenbrainz-server] 14amCap1712 merged pull request #3147 (03master…refactor-incremental): Refactor incremental stats framework https://github.com/metabrainz/listenbrainz-serv...

[listenbrainz-server] release 03v-2025-01-27.0 has been published by 14github-actions[bot]: https://github.com/metabrainz/listenbrainz-serv...

Did reosarevok and monkey both release LB and MB at the exact same time?? (comparing LB repo update email time + MB blog post time)

Or maybe lucifer or mayhem. This year is gonna be in sync!

1729