This enables us to set several TLS requirements for the users created
by the host_access resource.
Change-Id: If550f184f85f8fdbc197fc9f930d4446de67090a
Adding additional options for os_workers.rb:
os_workers_small
os_workers_large
os_workers will provide response times close to what
$::processorcount had (with two sockets machines).
os_workers_small will be the existing calculation.
os_workers_large can be used where services are distributed across
multiple machines and high worker count is not a concern.
Closes-bug: #1650424
Change-Id: I5dce760044d49ef9e3c88736c73cc267fd039ae7
when ceilometer api is deployed in httpd, however
the process name is like '/usr/sbin/httpd -DFOREGROUND',
it is ambiguous and make monitoring more difficult.
the aodh, gnocchi module are also like that.
so we should set an appropriate name for wsgi process.
Change-Id: I5e8c9be062a88e9ed6442e6cbce6573fba385030
Related-Bug: #1626550
If an array with a single host is passed into the hosts param, the
function was erroring because the host was getting converted to a string
when it was checking for bracketing for ipv6. This change checks the
type and fixes the case where a single host is being passed so it
continues to function.
Change-Id: I91b9959a6f71b4e6885e55a568116cc28cf16ddd
Closes-Bug: #1650042
Previously if you wanted to use the openstack::db::mysql for to create a
database and use the same user for multiple databases, the catalog would
fail due to a duplicate mysql_user definition. This change adds the
ability to disable the user creation as well as the grant if the user
does not need it.
Change-Id: Id04a622cc900254fe60bc257a9e42d16c676bf40
Related-Bug: 1649341
This sets the WSGIChunkedRequest setting in the vhost, which is
something that services like glance need when running over httpd.
Change-Id: I1c8816e9e5a9a38a3a86f3c0c5016df18f09ea62
This change adds a os_transport_url function that can be used to
generate correct URIs for the transport_url setting used by
oslo.messaging.
Change-Id: If83c0f0e61a08061334536399a42767a305966b7
I want to be able to use the internal API endpoints from Puppet. This is done
by setting $OS_INTERFACE to internal for use with the openstack CLI. This
change will do this by also reading OS_INTERFACE from the environment or
/root/openrc.
Change-Id: Id826836ccfe8f61780a6201ef2851068d75391ab
This change updates the tests to use rspec-puppet-facts when doing
different OS testing. Additionally as part of this change, there are
improvements to the openstacklib::policycrd testing which uncovered
issues with the verify_contents catalog test. The verify_contents calls
have been replaced with heredocs to better test when multiple services
are excluded.
Change-Id: I86bae2b16026e15b6e4445f3749419b8802bc94d
This patch changes the default worker count from ::processorcount to the
new ::os_workers fact. ::os_workers is based on the number of processors
(currently cpu/4) but is capped at a maximum of 8 worker processors.
This is a much more reasonable default in general and prevents excessive
resource consumption on systems with a large number of CPUs.
Change-Id: I458791aa8027cffeeec49698b302cb96ae5af2e2
Releasenote translation publishing is being prepared. 'locale_dirs'
needs to be defined in conf.py to generate translated version of the
release notes.
Note that this repository might not get translated release notes - or
no translations at all - but we add the entry here nevertheless to
prepare for it.
Change-Id: I4e38d2da134afa1e2bbbe3b747deec2b3adb6413
This fixes the credentials classes to properly accept region as a
property to manage. This allows OS_REGION_NAME to be set properly if
the child class passes it in.
Change-Id: I3345dac8bebd68f93290c1d45aa9a4d80bf3fb34
Partial-Bug: #1517220
This change updates the module to use the rspec-puppet-facts as defined
in the puppet-openstack_spec_helper.
Change-Id: Ida010fef458cbf86346928360971d87efba6b5d1
This will allow to run the service_validation on demand when required
from a notification, for example.
Change-Id: I18ea45f39cd7f4930de4916d658653d22d8530b3
This change adds a new fact that can be used as the default values for
any worker configuration. When used this fact will reduce the number of
default workers configured for services. Previously we were using
$::processorcount which could lead to excessive memory consumption or
connection limits being execeeded when services were deployed on
baremetal with large number of processors. The os_workers fact uses the
larger of two following values, '2' or '# processors / 4' but will not
exceed '8'.
Change-Id: Ia588817fde9203a134bc9eaad9a5bd5453d7639a
We should only ignore openstack from the root directory because
otherwise new files in the openstack/ subfolders under lib are not
tracked.
Change-Id: Ife454df6ae8d45826608bea59c3df2eb04e43016
Closes-Bug: #1541896
Allow to give an array of IP addresses to normalize_ip_for_url function.
Each IP in the list will be normalized like it would be for a string.
Change-Id: I8d361ce9cfcfe6a3f8592b2b7991971a3c748c75
Co-Authored-By: Athlan-Guyot sofer <sathlang@redhat.com>
The default exec timeout in puppet is 300 seconds, this can be a
very long while on services that do not return immediately.
With default values, a command that never returns we're looking at
10*300 seconds (or 50 minutes).
Let's default to something more reasonable, 60 seconds, and also
make it configurable.
Change-Id: I53bd4935fbbb3a86ac1ad522a0f3d8f017151dd0
Removing puppet-lint warnings
in favor of upgrading to latest gem
2016-09-13 21:12:33.947117 | manifests/db/postgresql.pp:37:WARNING: line has more than 140 characters
Change-Id: If40b8e5874c791e0a94bd634c09768c77f0c2d93
openstack_config contains a method 'self.instances' that is required for
purging unmanaged resources that references a static namevar method that
is not defined in openstack_config and may not be provided by base
classes. The 'namevar' method doesn't require any information from the
derived class so a default implementation in the base provider type
suffices.
Closes-Bug: #1612009
Change-Id: If27152863b9df5a28bf45e23ca989019666a958c
For instance this would enable this:
neutron_l3_agent_config {
'DEFAULT/router_id': value => $router_name, transform_to => 'uuid';
}
The neutron_l3_agent_config would only have to implement this:
def to_uuid(name)
# code to get the uuid
end
def from_uuid(uuid)
# code to return the name
end
Change-Id: I3b7c17590b27cd3a22c5458342d049969ade2281
Co-Authored-By: Drew Fisher <drew.fisher@oracle.com>
We recently patched the openstackclient code to increase the number
of retries when running commands[1], but I think what we really wanted
was to increase the amount of time to wait on each retry.
The original approach did help somewhat, but it is doing so by just
racing more times. Instead, it would be better to win the race more
often.
This is mostly a CI or developer environment issue where we are
deploying with the minimal amount of computing resources to get
a functioning cloud.
[1] https://git.openstack.org/cgit/openstack/puppet-openstacklib/commit/?id=489099632446b1903986549e5feb80b509f9b77d
Change-Id: I0707165a9e8054c1b3e0ad299604356cecc9f99f
When validating a service, it's useful to have logs when exec output
fails.
This patch adds this feature.
Change-Id: I43c9eff8516ea562bf9bde8d5651ddceeb496163
