#musicbrainz-devel

two my $sizes

I guess I can just remove the second 'my'

I'll read it as 'let urlSize = maybe mempty ('-' : ) size' and it's less scary

:P

oh, I'll name it urlsize then. slightly less confusing.

or just sprintf to form the url

then you get: sprintf "http://ueauea/mbid-%s%s.%s", $mbid, defined($size) ? .., $suffix

which also requires less munging

I agree that's prettier.

(but I pushed the urlsize variant)

ocharles: can you check https://github.com/metabrainz/CAA-indexer/pull/... ? and if you're happy with it we can merge that and redeploy.

warp: i'm happy with that

ruaok:

[2013-07-18T10:16:17+00:00] INFO: Processing service[postgresql] action enable (postgresql::default line 20)

[2013-07-18T10:16:17+00:00] INFO: Processing service[postgresql] action start (postgresql::default line 20)

[2013-07-18T10:16:17+00:00] INFO: Processing service[postgresql] action reload (postgresql::default line 20)

[2013-07-18T10:16:18+00:00] INFO: service[postgresql] reloaded

which is indeed wrong

it does appear that postgresql doesn't start, it only reloads

hmm no, pg is running fine

odd

mine was running too

hrm, I get no route to host trying to get to hobbes.

warp: I did get a nagios that hobbes was unreachable recently

evidently, that was not one to ignore

should I call in a reboot?

I didn't realize we already moved all the important bits off it, because I haven't seen anyone complaining here :)

jesus kept complaining.

Isn't that always true? :p

ruaok: yes, I cannot ssh into the machine from lenny and such.

srsly

no, jesus is complaining because i turned off test.mb.org

when i upgraded to pg 9.1

i wondered if anyone would notice :)

I didn't intend to entirely nuke hobbes :)

ocharles: hm, CAA-indexer doesn't have an origin, can I rename the github remote?

yes please

that confused me

ok, reboot ordered, but apparently they are fighting some other fires too.

ping me in 10 if its not back.

ruaok: ok, there's no hurry. I don't need it for anything.

k

ocharles: caa-indexer updated and restarted

warp: goodo

try and index something with a png

yay, my caa-indexer VM came up and has a rabbitmq with all the right queues and exchanges

ocharles: yay! I see .png in http://archive.org/download/mbid-442b3814-b2c1-...

woohoo. and how does it look on mb.org?

looks good afaics

the "original" link on the third image of https://beta.musicbrainz.org/release/442b3814-b... is to a .png

http://archive.org/details/mbid-442b3814-b2c1-4... thumbnails \o/

\o/

firefox seems to be happy it is indeed a png

that album looks awful, too

:P

alright, for reindexing already submitted .png/.gif images. can I just make a list, and munge that into a SELECT pgq.insert_event() statements in a .sql file, and source that on totoro?

just run 'SELECT pgq.insert_event(..) FROM release JOIN whatever WHERE foo' and run that on totoro

that call will run in a few ms

and then the indexer will do its thing

ah, ok

oh it's only 43 releases.

yeah, I discouraged people from going nuts because of the thumbnails and the fact the images don't work at all on the main servers right now :P

done

https://beta.musicbrainz.org/release/9a17b124-5... also looking good.

ocharles: shall I merge the trigger changes into master now? (considering that has already been deployed to production).

warp: yea, master or beta

oh yes, it has to beta, it is branched off of beta.

gah.

I think the AC in my office is making me ill.

went to office, starting feeling bad. went home, feeling better.

i'm not feeling too perky today

at least you have AC though, the lack of that is why i feel off

well, when I've sat in the office without AC, I was fine.

so, no I have no AC. :(

which may drive me to rent an office here, outside of MTG.

you don't think you will adapt to the ac then?

I have no problems with AC. the office in cali had great AC.

I suspect that there is mold in the AC system here.

and that everytime I turn it on, I get blasted with mold spores.

i know

i meant adapting to that bit

not something I want to adapt to. mold can be really bad for you.

and in the last working week, we've lost quite a bit of working time to being mysteriously ill.

and I'm quite done with that, honestly.

anyways, I'm back home where I have a working vagrant install. I'll keep working on it.

understandably

ocharles: any idea on this one?

Are you trying to run a bash script?

I've seen this when a bash script fails to exit with 0

it's an extremely useless stack trace, and I can't figure out why we get one like that

while running vagaran provision

*vagrant

yea, but presumably you have a bash script in one of the recipes you're running

Not really.

package "openjdk-7-jre"

and it still fails.

and even when I remove chef.add_recipe "search-server" from Vagrantfile it still fails.

paste your Vagrantfile

the most minimal one you've got it to fail on

ok, let me work on whittling that down.

I think maybe you're using something that depends on an attribute that's not set

I've undone all of my changes now, I think.

this is the most minimal one that fails:

if I remove "chef.add_recipe "virtualmachine"" then I get [default] Warning: Chef solo run list is empty. This may not be what you want.

right

nikki has changed the topic to: agenda: questions on reviews, MBS-3841 (nikki), sortnames (nikki)

you have a cookbooks/virtualmachine folder?

yes with recipes/default.rb

did you write that yourself? i can't see it on github

no

is this more of the submodules hell?

what is git-status on vagrant-images

and what is git rev-parse HEAD

# Your branch is ahead of 'origin/master' by 28 commits.

# modified: musicbrainz-server/Vagrantfile

2834095

ok, have you modified the cookbooks project at all?

and I don't know what 2834095 is, so I guess you've commited

yes, I added the search-server subdir

ok

yes, its a local commit on my forked repo

go into the cookbooks directory, and git rev-parse HEAD there

ok

when did you make this commit?

ef94b90f94b

i've ripped vagrantimages apart quite a lot when i was getting ready to review

so I think you may have branched off something that no longer exists

yes, I've gotten that feeling.

something has been quite odd.

I'd start over again, but then I would have to wait to build the VM and import the data again.

reluctant to do that, but if I have to...

no, you don't need to do that

we just need to get you on the right commits

guide me obi-wan!

first, lets do vagrant-images

you have a single commit, or multiple commits?

(of your own work)