csscorp / openstack-automation Goto Github PK

View Code? Open in Web Editor NEW

74.0 74.0 44.0 763 KB

Openstack deployment using saltstack

License: GNU General Public License v2.0

Python 99.46% Scheme 0.54%

openstack-automation's People

Contributors

Stargazers

Watchers

openstack-automation's Issues

horizon/init.sls looks for conf-enabled when it should be sites-enabled

or is it different in debian and ubuntu ?

(I have a fix I can pull request)

I want to talk about CSSCorps/openstack-automation repo at OSCON

I'm working on my slides for OSCON http://www.oscon.com/oscon2014/ and would love to promote this repo and the developers working on it. Would that be ok with CSSCorp? Have a logo or other info I can add.

Thanks, Yazz

Add a single node nova.conf state file.

Something like nova.controller_compute_combined so nova can be removed in controller pillar and nova.compute_kvm can be replaced by nova.controller_compute_combined in compute pillar.

Are these states actually being used?

Are these states actually being used? I was able to get this working, but there were several sls errors that I had to fix.

glance_sync step fails on debian deployment.

After some tweaking for debian (will do a pull request shortly) I get a working highstate. Only exception is the following:

          ID: glance_sync
    Function: cmd.run
        Name: glance-manage db_sync
      Result: False
     Comment: Command "glance-manage db_sync" run
     Changes:   
              ----------
              pid:
                  2958
              retcode:
                  1
              stderr:
                  2014-06-19 14:47:12.420 2959 INFO glance.db.sqlalchemy.migration [-] Upgrading database to version latest
                  2014-06-19 14:47:12.434 2959 INFO migrate.versioning.api [-] 13 -> 14... 
                  2014-06-19 14:47:12.437 2959 CRITICAL glance [-] `images`
                  2014-06-19 14:47:12.437 2959 TRACE glance Traceback (most recent call last):
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/bin/glance-manage", line 10, in <module>
                  2014-06-19 14:47:12.437 2959 TRACE glance     sys.exit(main())
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/glance/cmd/manage.py", line 127, in main
                  2014-06-19 14:47:12.437 2959 TRACE glance     CONF.command.func()
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/glance/cmd/manage.py", line 77, in do_db_sync
                  2014-06-19 14:47:12.437 2959 TRACE glance     CONF.command.current_version)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/glance/db/sqlalchemy/migration.py", line 122, in db_sync
                  2014-06-19 14:47:12.437 2959 TRACE glance     upgrade(version=version)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/glance/db/sqlalchemy/migration.py", line 61, in upgrade
                  2014-06-19 14:47:12.437 2959 TRACE glance     return versioning_api.upgrade(sql_connection, repo_path, version)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/migrate/versioning/api.py", line 186, in upgrade
                  2014-06-19 14:47:12.437 2959 TRACE glance     return _migrate(url, repository, version, upgrade=True, err=err, **opts)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "<string>", line 2, in _migrate
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/migrate/versioning/util/__init__.py", line 159, in with_en
gine
                  2014-06-19 14:47:12.437 2959 TRACE glance     return f(*a, **kw)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/migrate/versioning/api.py", line 366, in _migrate
                  2014-06-19 14:47:12.437 2959 TRACE glance     schema.runchange(ver, change, changeset.step)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/migrate/versioning/schema.py", line 91, in runchange
                  2014-06-19 14:47:12.437 2959 TRACE glance     change.run(self.engine, step)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/migrate/versioning/script/py.py", line 145, in run
                  2014-06-19 14:47:12.437 2959 TRACE glance     script_func(engine)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/glance/db/sqlalchemy/migrate_repo/versions/014_add_image_t
ags_table.py", line 64, in upgrade
                  2014-06-19 14:47:12.437 2959 TRACE glance     tables = [define_image_tags_table(meta)]
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/glance/db/sqlalchemy/migrate_repo/versions/014_add_image_t
ags_table.py", line 23, in define_image_tags_table
                  2014-06-19 14:47:12.437 2959 TRACE glance     schema.Table('images', meta, autoload=True)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/sqlalchemy/schema.py", line 318, in __new__
                  2014-06-19 14:47:12.437 2959 TRACE glance     table._init(name, metadata, *args, **kw)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/sqlalchemy/schema.py", line 381, in _init
                  2014-06-19 14:47:12.437 2959 TRACE glance     self._autoload(metadata, autoload_with, include_columns)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/sqlalchemy/schema.py", line 408, in _autoload
                  2014-06-19 14:47:12.437 2959 TRACE glance     self, include_columns, exclude_columns
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/sqlalchemy/engine/base.py", line 2426, in run_callable
                  2014-06-19 14:47:12.437 2959 TRACE glance     return conn.run_callable(callable_, *args, **kwargs)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/sqlalchemy/engine/base.py", line 1969, in run_callable
                  2014-06-19 14:47:12.437 2959 TRACE glance     return callable_(self, *args, **kwargs)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/sqlalchemy/engine/default.py", line 260, in reflecttable
                  2014-06-19 14:47:12.437 2959 TRACE glance     return insp.reflecttable(table, include_columns, exclude_columns)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/sqlalchemy/engine/reflection.py", line 350, in reflecttabl
e
                  2014-06-19 14:47:12.437 2959 TRACE glance     tbl_opts = self.get_table_options(table_name, schema, **table.kwargs)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/sqlalchemy/engine/reflection.py", line 178, in get_table_options
                  2014-06-19 14:47:12.437 2959 TRACE glance     **kw)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "<string>", line 1, in <lambda>
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/sqlalchemy/engine/reflection.py", line 47, in cache
                  2014-06-19 14:47:12.437 2959 TRACE glance     ret = fn(self, con, *args, **kw)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/sqlalchemy/dialects/mysql/base.py", line 2057, in get_table_options
                  2014-06-19 14:47:12.437 2959 TRACE glance     parsed_state = self._parsed_state_or_create(connection, table_name, schema, **kw)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/sqlalchemy/dialects/mysql/base.py", line 2156, in _parsed_state_or_create
                  2014-06-19 14:47:12.437 2959 TRACE glance     info_cache=kw.get('info_cache', None)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "<string>", line 1, in <lambda>
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/sqlalchemy/engine/reflection.py", line 47, in cache
                  2014-06-19 14:47:12.437 2959 TRACE glance     ret = fn(self, con, *args, **kw)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/sqlalchemy/dialects/mysql/base.py", line 2181, in _setup_parser
                  2014-06-19 14:47:12.437 2959 TRACE glance     full_name=full_name)
                  2014-06-19 14:47:12.437 2959 TRACE glance   File "/usr/lib/python2.7/dist-packages/sqlalchemy/dialects/mysql/base.py", line 2268, in _show_create_table
                  2014-06-19 14:47:12.437 2959 TRACE glance     raise exc.NoSuchTableError(full_name)
                  2014-06-19 14:47:12.437 2959 TRACE glance NoSuchTableError: `images`
                  2014-06-19 14:47:12.437 2959 TRACE glance
              stdout:

Went back to the install manual http://docs.openstack.org/havana/install-guide/install/apt/content/glance-install.html dropped the database and re-created from the instructions there. And the db_sync worked.

Would it be because the mysql handling is different between ubuntu and debian ? Any ideas why this would fail ?

Use mariadb instead of mysql

featurerequest

(if we start using theses states in production I might contribute this if no one has since I posted this ticket)

config file owner changes to root during deployment

config files pushed from master are owned by root. so services fail to start as they do not have permission.

Provide documentation to use with gitfs

Hopefully one can use gitfs http://salt.readthedocs.org/en/v2014.1.4/topics/tutorials/gitfs.html?highlight=gitfs to use these pillars and states, it would be nice to document how to set it up (if possible).

Activate ssl version of dashboard by default

using dashboard-ssl.conf and dashboard-ssl-redirect.conf (port 80 redirects to port 443) by default would be nice

Changes to be ammended to this repo

Hi All,
The repo has been lacking activity from my side for quiet some time now. The following changes will be made to this project.

All branches except the master shall be removed
The changes and commits to all other branches will be manually updated to the master
Pull requests to all branches will be consolidated and merged manually to master
'yaml' will be default format for sls files. This is done as maintaining sls files across format is causing mismatches and errors
Support for cinder sls will be added
Repo will be modified in a way to enable users to use the repo directly as salt gitfs backend
Support will be added to pillar file to make the project distro agnostic

The changes mentioned above shall first be made in https://github.com/Akilesh1597/openstack-automation , tested and then merged here.

Thank you and please do leave a comment for your suggestions

Deployment stuck on bridge creating

Here almost default config file: http://paste.ubuntu.com/7125446/
Here is venus ps axf: http://paste.ubuntu.com/7125460/
Here is mercury ps axf: http://paste.ubuntu.com/7125464/

I tried to add one more tunnel, and it freezes too:

$ sudo ovs-vsctl add-br br-int1

Both venus and mercury are vm's with fresh Ubuntu-12.04.4

Jinja variable 'dict object' has no attribute 'install'; line 2

Hello,

Really nice work. Thank you.
I am trying to test the automatically deployment of openstack using salt but I receive an error:
On one of the minion:
root@hawk:~# salt-call state.highstate -l debug
the message is:
local:

Data failed to compile:

Traceback (most recent call last):

File "/usr/lib/pymodules/python2.7/salt/state.py", line 2349, in call_highstate
top = self.get_top()
File "/usr/lib/pymodules/python2.7/salt/state.py", line 1926, in get_top
tops = self.get_tops()
File "/usr/lib/pymodules/python2.7/salt/state.py", line 1809, in get_tops
env=env
File "/usr/lib/pymodules/python2.7/salt/template.py", line 69, in compile_template
ret = render(input_data, env, sls, **render_kwargs)
File "/usr/lib/pymodules/python2.7/salt/renderers/jinja.py", line 42, in render
tmp_data.get('data', 'Unknown render error in jinja renderer')
SaltRenderError: Jinja variable 'dict object' has no attribute 'install'; line 2

havana:
{% for cluster_component in pillar['install'] %} <======================
{% for server in pillar[cluster_component] %}
{{ server }}:
{% for sls in pillar['install'][cluster_component] %}
- {{ sls }}
{% endfor %}
[...]

On the salt-master I configured:
file_roots:
base:
- /srv/salt/
havana:
- /srv/havana/file
pillar_roots:
base:
- /srv/pillar
havana:
- /srv/havana/pillar

The project I put in /srv/havana

Thanks a lot,
Gabriel

No License on this Project

Hi there,

you're stating nowhere what's the license on this project. Not even
"this is licensed under the same term's as (OpenStack||SaltStack)"
(which would be Apache Public License in both cases).

Regards, Florian

PS: I took a quick look on your neutron-module but I guess I can't
use it until the license is clear.

openvswitch gre tunnel does not work with hostname

file '/etc/neutron/plugins/openvswitch/ovs_neutron_plugin.ini' has a key named 'local_ip' giving the host name of the machine here renders the setup useless.

Work around is to manually find the host name of the machine where neutron-ovs-plugin is installed and edit the corresponding config files cached in the master in the location
'file_root/config/<cluster_name>/<machine_name>/etc/neutron/plugins/openvswitch/ovs_neutron_plugin.ini' and set the local_ip key to the machines ip address.

neutron/ml2.sls fails when run twice

The following step in neutron/ml2.sls fails when run twice

----------
          ID: intergrationg_bridge
    Function: cmd.run
        Name: ovs-vsctl add-br br-int
      Result: False
     Comment: Command "ovs-vsctl add-br br-int" run
     Changes:   
              ----------
              pid:
                  16413
              retcode:
                  1
              stderr:
                  ovs-vsctl: cannot create a bridge named br-int because a bridge named br-int already exists
              stdout:

It should check if the bridge is already configured (and configured according to the spec) and not do anything if it is.

Running everything on one physical server

I would like to run everything one one server with just a eth0 interface.

I run into three issues:

nova.conf in nova.compute_kvm conflicts with the one in nova.init
wich pillar should i use for net when i only have a eth0
the way i tried, the net it goes down and i have to log in without net and remove the bridges before i can reach the physical server again.

Any tips? :)

Support for icehouse installation

Do you have any plans to support an icehouse (next version of havana) installation ?

license change request from GPL to apache 2

Would it be possible to change the license for this project?

It is important to use Apache 2 license in order to get our code contributed to the OpenStack and SaltStack community. They will not accept code with GPL license.

Can you please revise the license from GPL to Apache2? Here are the steps to do this: http://www.apache.org/dev/apply-license.html#new

Reference: #43

Release versions of this project

First off, am glad there is some activity on the project, looking forward to re-using this for our next openstack deployment.

It would be nice to have a release cycle on openstack-automation, some versionning would enable users to see some sort of changelog and keep track of what is planned and what works for which version.

implement cinder deployment

I see there are empty cinder.sls any plans on implementing this ?

I might give it a shot soon.

Comment: State ini.options_present found in sls mysql is unavailable

I believe s/ini/ini_manage/

Has ini_manage been integrated in salt ? http://docs.saltstack.com/en/latest/ref/states/all/salt.states.ini_manage.html If so, maybe it could be removed from _modules to avoid code duplication ?

neutron module passes profile setting onto command so fails on many

salt-call --local neutron.delete_subnet profile=virl flat1
[INFO ] Executing command '/sbin/zfs help' in directory '/home/virl'
[INFO ] Executing command '/sbin/zfs help' in directory '/home/virl'
[INFO ] Starting new HTTP connection (1): 127.0.0.1
[ERROR ] calling with args ('flat1',)
[ERROR ] calling with kwargs {'profile': 'virl'}
Passed invalid arguments: delete_subnet() got an unexpected keyword argument 'profile'

Latest OpenStack release version

Hi
Is this support installation of newer than icehouse?
I saw icehouse as the example version in README

have example pillars in the repo so pillars aren't tracked by git

Openstack_Icehouse/pillar/openstack_cluster.sls and Openstack_Icehouse/pillar/openstack_cluster_inverse.sls could be moved to .example files and then .gitignore could include these files. In the README you would instruct to copy both these files and then edit them.

Why JSON?

Why did you decide to write salt states in JSON / Jinja format?

JSON doesn't allow for comments and such. Its really not a great choice for configuration files that a human has to edit.

Merge efforts with salt-openstack

As discussed in cloudbase/salt-openstack#4 (comment) the https://github.com/cloudbase/salt-openstack/ project looks very similar to what is being done here and it would be a shame to duplicate the work.

It would be great to see the two projects share states and code (if possible).

csscorp / openstack-automation Goto Github PK

openstack-automation's People

Contributors

Stargazers

Watchers

Forkers

openstack-automation's Issues

featurerequest

Data failed to compile:

Recommend Projects

Recommend Topics

Recommend Org