#metabrainz

OK. I've recreated my ubuntu 18.04 - best make a snapshot.

So I can't use the docker than comes with ubuntu? I have to install my own version?

I give up. Thanks for your help anyway.

I'm back, doing the release now. updating beta/prod containers atm

nelgin: why can't you use the docker that comes with ubuntu?

Because the instructions say 1.21.1 or higher and Ubuntu comes with 1.17.1

oh wait, I did the update but not the upgrade...sorry, I work almost exclusively on rpm systems.

ok, did an update and upgrade and the ubuntu docker-compose is still 1.17.1

huh, I wonder why it needs 1.21

That's what the instructions say I was given.

I guess you can get a later version from https://docs.docker.com/install/linux/docker-ce...

That's what I was told, fail at the first hurdle.

hmm, it explains why 1.21 is needed there https://github.com/metabrainz/musicbrainz-docke...

but I can't find that in the current version of the readme

yvanzo: ^ is 1.21 still needed? if we can just fix the commands to use the correct names instead that'd be a lot less annoying for people using ubuntu 18.04 I think

I can try 1.17.1 and see.

if you run into an issue with the names maybe we can find a workaround

not sure that's a good reason to require people to upgrade at least

It's at the "setting up" stage

Unescaped left brace in regex is deprecated, passed through in regex; marked by <-- HERE in m/(?<!\\)\${ <-- HERE ([^}]+)}/ at /usr/sbin/pam_getenv line 78.

Got a bunch of those after Setting up postgresql

invoke-rc.d: could not determine current runlevel

invoke-rc.d: policy-rc.d denied execution of start.

Some runlevel stuff

does the whole process abort after those message, or does it keep going? (wondering if they're non-fatal warnings that can be ignored)

Keep going

It's building the perl modules now

ok, maybe they're harmless then. I've seen the rc.d ones before and the regex one looks non-fatal

(I haven't used musicbrainz-docker before, but I've seen them elsewhere)

arning about asset size..

yarn install v1.19.1

[1/4] Resolving packages...

[2/4] Fetching packages...

info fsevents@1.2.7: The platform "linux" is incompatible with this module.

info "fsevents@1.2.7"; is an optional dependency and failed compatibility check. Excluding it from installation.

info fsevents@1.2.4: The platform "linux" is incompatible with this module.

info "fsevents@1.2.4"; is an optional dependency and failed compatibility check. Excluding it from installation.

A bit further up I got these

Creating the database, guess I'll check in tomorrow lol

heh, okay

those you can ignore too

alright, finished the mbs release now

databas eimported. working on creating indexes

sbvkrishna: Hey guys, i got my GitHub account back! It was a random flag done by the GitHub spam control system. Anyways my PR is still up https://github.com/bookbrainz/bookbrainz-site/p.... Could anyone check into it?

Also can i too contribute to the Google Code-In as a mentor ?

I was a code-in participant for 2018 and 2017 and have done some task with Musicbrainz and bookbrainz.

iliekcomputers: hi

Ok, I'm at InitDB.pl succeeded. Let's see what's next.

Docker is up. Let's see if I can build the search indexes. I'm running on a vm with 2 CPUs and 4GB of memory but I'm using a fixed rather than dynamic disk...so maybe take a day or two :)

musicbrainz@musicbrainz:~/musicbrainz-docker$ sudo docker-compose exec indexer python -m sir reindex

ERROR: No container found for indexer_1

FAIL!

Also get "Style sheet could not be loaded http://localhost:5000/build/common-696a1a7.css";

When I go to look at the page in the browser.

Loading failed for the <script> with source “http://localhost:5000/static/build/runtime-e9fd0a9.js”. 192.168.168.159:5000:1:1

Loading failed for the <script> with source “http://localhost:5000/static/build/common-chunks-d38d211.js”. 192.168.168.159:5000:1:1

Loading failed for the <script> with source “http://localhost:5000/static/build/jed-data-fb8a297.js”. 192.168.168.159:5000:1:1

Loading failed for the <script> with source “http://localhost:5000/static/build/common-8c1f1ae.js”. 192.168.168.159:5000:1:1

Yeah so...that doesn't work either. If someone wants to make suggestions, I'll check in later.

Nobody?

Looking at the log if I run docker-compose up without the -d

Creating musicbrainzdocker_indexer_1 ... done

Attaching to musicbrainzdocker_search_1, musicbrainzdocker_db_1, musicbrainzdocker_redis_1, musicbrainzdocker_musicbrainz_1, musicbrainzdocker_indexer_1

indexer_1 | stdin: is not a tty

musicbrainzdocker_indexer_1 exited with code 1

If I had stuck with the ova and the crawling slow replication from 2018, it'd be done by now :(

Oh ffs.

[picard:github-actions@7251f64 - build #293] CI canceled! (https://travis-ci.org/phw/picard/builds/617065082)

[picard:github-actions@de63080 - build #291] CI errored! (https://travis-ci.org/phw/picard/builds/617064360)

[picard:github-actions@6c87ea3 - build #295] CI passed! (https://travis-ci.org/phw/picard/builds/617068176)

yvanzo: turns out we're just inconsistent: https://github.com/metabrainz/musicbrainz-serve...

[picard:github-actions@ccb68bc - build #306] CI errored! (https://travis-ci.org/phw/picard/builds/617071457)

[picard:github-actions@00d7fe6 - build #307] CI passed! (https://travis-ci.org/phw/picard/builds/617074684)

[picard:debug-build@b88f205 - build #5369] CI canceled! (https://travis-ci.org/metabrainz/picard/builds/...)

[picard:debug-build@6ef7ccb - build #5370] CI canceled! (https://travis-ci.org/metabrainz/picard/builds/...)

mo’’in’

reosarevok: use 'any' then :)

bitmap: thanks for the release!

Or we could change our code to be consistent like we do with everything else? :/

bitmap: 1.21 is needed for https://github.com/metabrainz/musicbrainz-docke...

As an alternative, you can either use COMPOSE_PROJECT_NAME=musicbrainz-docker (and rebuild every image so far) or replace musicbrainz-docker_ with musicbrainzdocker_ in this file.

To install a more recent version of Docker: https://docs.docker.com/install/linux/docker-ce...

Moin !

Then to install a more recent version of Docker Compose: https://docs.docker.com/compose/install/

reosarevok: it is already consistent, with 'any' :)

I mean, that to me is like saying "it's already consistent if we turn the rule off"

reosarevok: this rule is here to spot common mistakes, 'any' still detects such mistakes.

But why would we choose to have our code be less standardised if we can make it more standardised?

No, turning the rule off wouldn't detect such error.

It feels super weird to me to choose to have one thing that can be done in two ways done in different ways throughout the code

This is way too complicated for someone who justs wants to get it up and running. I've been doing unix/linux admin for over 30 years and I'm still having problems following along.

nelgin: I agree, sorry, it is currently a bit of a mess, due to the ongoing search server migration, plus the schema change.

So it'll really fail if I try and build/run it with docker-compose 1.17.1 even tho that's the latest that comes with Ubuntu?

Yes, not because of features, because it has a bug that is fixed in 1.21.0-rc1 only.

Note, it used to work with older versions of docker-compose before it broke stuff.

It still is probably less pain than handling dependencies, pg extensions, and all by hand.

Oh well...

At least I have the database, I'll just write some perl scripts to extract what I need rather than having the fun api.

or mariadb anyway, who the hell uses postgres any more ;)

Because then we'd have to write PHP

(I think that's the answer you'll get from all our devs, not just me :p )

And I mean, there's really not that much of a difference between mysql and postgres for just using the data anyway. I suspect it'd be as complicated in mysql/mariadb, it's just a complicated DB with a lot of things going on

moin!

yvanzo: can't you explicitly set the project name to be without the hyphen though?

bitmap: yes but then you would have to set COMPOSE_PROJECT_NAME for recent versions...

I mean the readme would have to include -p or whatever in all the commands so it'd work in either version

...which is probably more pain than installing latest version from dedicated apt repos.

I don't see how making people to upgrade is less painful tbh, since they're copy/pasting the compose commands from the readme anyway

or we could provide scripts for that

Ultimately, there should be a global script indeed.

But if we change project name, that would break current users setup on git pull.