Puppet provisioning fails about mozillians HOT 13 CLOSED

I did a fresh vagrant up yesterday, got a bunch of errors and everything worked. Either way, this is a valid issue, but did you try starting the server after the errors?

from mozillians.

I tried running the manage.py commands listed at http://mozillians.readthedocs.org/en/latest/installation.html, but creating a superuser failed. I'm unsure if that's a step I was supposed to enact.

from mozillians.

However, the server appears to be running ok, so thanks for the suggestion!

from mozillians.

I may have spoken to soon. I can't log in or register an account without getting a "DoesNotExist at /browserid/verify/ UserProfile matching query does not exist." traceback :/

from mozillians.

I would like to point out that I see this:

godot:mozillians jdm$ find . -name slapd.conf
./puppet/files/directory/devslapd/slapd.conf
godot:mozillians jdm$ find . -name setup.sh
./puppet/files/directory/devslapd/setup.sh

The puppet rules specify that the files should appear at /home/vagrant/mozillians/directory/devslapd, so I assume that this is supposed to be part of the magic ssh setup that lets me edit files in my git clone and see the changes be reflected in the server that's running in the vm. What's going wrong here?

from mozillians.

That's interesting. By using vagrant ssh and mkdir -p mozillians/directory/devslapd I was able to fix the problem and provisioning continued.

from mozillians.

I now hit this error instead:

[default] err: /Stage[main]/Playdoh_site/Exec[sql_migrate]/returns: change from notrun to 0 failed: /usr/bin/python2.6 ./vendor/src/schematic/schematic migrations/ returned 2 instead of one of [0] at /tmp/vagrant-puppet/manifests/classes/playdoh.pp:30

Running the command inside the vm shows:

(mozillians)lucid32 :: ~/mozillians » python2.6 ./vendor/src/schematic/schematic migrations/
Usage: schematic /path/to/migrations/dir

schematic: error: Expected a directory: /path/to/migrations/dir

Given that there is no migrations/ directory, this does not surprise me. Faking a mozillians/migrations directory yields:

[default] err: /Stage[main]/Playdoh_site/Exec[south_migrate]/returns: change from notrun to 0 failed: /usr/bin/python2.6 manage.py migrate returned 1 instead of one of [0] at /tmp/vagrant-puppet/manifests/classes/playdoh.pp:37

which turns out to be a traceback ending in

django.db.utils.DatabaseError: (1146, "Table 'mozillians.south_migrationhistory' doesn't exist")

Does any of this help?

from mozillians.

I think that broken puppet scripts is sortof a known issue right now. Were in the midst of removing LDAP all together and sometimes that means keeping code around that we aren't going to be using.

I know that if you run:

$ vagrant up && vagrant ssh
$ mysql -uroot
> drop database mozillians;
> create database mozillians;
$ cd mozillians
$ dj syncdb
$ dj migrate

everything works out. These things should be done by puppet but nobody has had the time to fix it.

from mozillians.

That did take me to a seemingly-running server, but trying to log in with browserid fails after attempting 5 connections to browserid.org.

from mozillians.

Actually it fails with Max retries exceeded for url: /verify?timeout=5

from mozillians.

That means elastic search is failing you. Try:

$ vagrant restart && vagrant ssh
$ cd mozillians/
$ rs (or dj runserver 0.0.0.0:8001 or ./manage.py runserver 0.0.0.0:8001)

Vagrant tries hard but sometimes it seems to get messed up with the interfaces especially when you move between different networks (on you host computer)

from mozillians.

I made it work by not attempting to create a superuser in the syncdb step. Hooray!

from mozillians.

vagrant up runs with zero issues now thanks to 3e81801.

from mozillians.