Age | Commit message (Collapse) | Author | Files | Lines |
|
|
|
In the previous release[1], the services were stopped before the
pacemaker services, so that they get a chance to send last message to
the database/rabbitmq queue:
Let's do the upgrade in the same order.
[1] https://github.com/openstack/tripleo-heat-templates/blob/stable/newton/extraconfig/tasks/major_upgrade_controller_pacemaker_2.sh#L13-L71
Change-Id: I1c4045e8b9167396c9dfa4da99973102f1af1218
|
|
The default is to deploy v2.0 endpoints, but this is not the recommended
approach. we should instead be using versionless endpoints
Change-Id: Icbfae1c2ff2b7312646fd8e817dd8209220a0d96
Related-Bug: #1667679
|
|
When fixing LP#1643487 we added ?bind_address to all DB URIs.
Since this clashes with Cellsv2 due to the URIs becoming host
dependent, we need a new approach to pass bind_address to pymysql
that leaves the DB URIs host-independent.
In change Iff8bd2d9ee85f7bb1445aa2e1b3cfbff1f397b18 we first create a
/etc/my.cnf.d/tripleo.cnf file with a [tripleo] section with the correct
bind-address option.
In this change we make sure that the DB URIs will point to the added
file and to the specific section containing the necessary bind-address
option. We do introduce a new MySQLClient profile which will hold all
this more client-specific configuration so that this change can fit
better in the composable roles work. Also, in the future it might
contain the necessary configuration for SSL for example.
Note that in case the /etc/my.cnf.d/tripleo.cnf file does not exist
(because it is created via the mysqlclient profile), things keep on
working as usual and the bind-address option simply won't be set, which
has no impact on hosts where there are no VIPs.
Co-Authored-By: Damien Ciabrini <dciabrin@redhat.com>
Change-Id: Ieac33efe38f32e949fd89545eb1cd8e0fe114a12
Related-Bug: #1643487
Closes-Bug: #1663181
Closes-Bug: #1664524
Depends-On: Iff8bd2d9ee85f7bb1445aa2e1b3cfbff1f397b18
|
|
These are handle by puppet as usual (puppet run comes after the
ansible steps) so remove them from these remaining upgrade_tasks
Change-Id: Ic341f31251622ccb11a5f7818b2edf7a82391560
|
|
We don't need all the steps currently enabled for either batched
or concurrent updates, so decrease them. In future we can perhaps
introspect the task tags during plan creation and set these
dynamically.
Change-Id: I0358886a332dfbecd03bc4a67086b08d25756c22
Partially-Implements: blueprint overcloud-upgrades-per-service
|
|
These are only used for TLS-everywhere, and fills up the kerberos
principals that will need to be created for the certs used by the
overcloud. With this, the metadata hook will format these principals
correctly and will further pass them on to the nova metadata service.
Where they can be used if there's a plugin enabled.
bp tls-via-certmonger
bp novajoin
Change-Id: I873094bb69200052febda629fda698a7a782c031
|
|
|
|
Currently we start all OpenStack services in step6, but puppet
already does this, and sometimes services require configuration
to account for the new version after the yum update before they
will start.
So instead of reimplementing that configuration management in
ansible, just defer starting the services until puppet has run
which will happen right after the ansible upgrade steps complete.
Note there are some DB sync operations etc that we may also be able
to remove and let puppet do those steps, but I've left those in
for now, as we know there are some actions during that phase
e.g nova cells setup, which aren't yet handled by puppet.
Change-Id: Idc8e253167a4bc74b086830cfabf28d4aab97d28
|
|
Change-Id: I1b2c0025e363d0387ddc0514decd3bc2dc80f9ae
Closes-Bug: 1650680
|
|
|
|
When a service connects to the database VIP from the node hosting this
VIP, the resulting TCP socket has a src address which is by default
bound to the VIP as well. If the VIP is failed over to another node
while the socket's Send-Q is not empty, TCP keepalive won't engage and
the service will become unavailable for a very long time (by default
more than 10m).
To prevent failover issues, DB connections should have the src address
of their TCP socket bound to the IP of the network interface used for
MySQL traffic. This is achieved by passing a new option to the
database connection URIs. This option is available starting from
PyMySQL 0.7.9-2.
We use a new intermediate variable in hiera to hold the IP to be used
as a source address for all DB connections. All services adapt their
database URI accordingly.
Moreover, a new YAML validation check is added to guarantee that new
services will construct their database URI appropriately.
Change-Id: Ic69de63acbfb992314ea30a3a9b17c0b5341c035
Closes-Bug: #1643487
|
|
Heat now supports release name aliases, so we can replace
the inconsistent mix of date related versions with one consistent
version that aligns with the supported version of heat for this
t-h-t branch.
This should also help new users who sometimes copy/paste old templates
and discover intrinsic functions in the t-h-t docs don't work because
their template version is too old.
Change-Id: Ib415e7290fea27447460baa280291492df197e54
|
|
Keystone UUID tokens require the token_flush job to delete expired
tokens to prevent the disk from filling. When the job runs, it should
be allowed to log to the disk so that the job can be traced if required.
Change-Id: I62e36e0968902564b97093a45df15e963ad08242
Closes-Bug: #1648174
|
|
|
|
This shows how we could wire in the upgrade steps using Ansible
as was previously proposed e.g in https://review.openstack.org/#/c/321416/
but it's more closely integrated with the new composable services
architecture.
It's also very similar to the approach taken by SpinalStack where
ansible snippets per-service were combined then run in a series of
steps using Ansible tags.
This patch just enables upgrade of keystone - we'll add support for
other patches in subsequent patches.
Partially-Implements: blueprint overcloud-upgrades-per-service
Change-Id: I39f5426cb9da0b40bec4a7a3a4a353f69319bdf9
|
|
This changes how we get the network-based FQDNs for the specific
services, from using the custom fact, to the new hiera entries.
Change-Id: Iae668a5d89fb7bee091db4a761aa6c91d369b276
|
|
This patch optimizes how we deploy hiera by using a new
heat hook specifically designed to help compose hiera
within heat templates. As part of this change:
- we update all the 'hiera' software configurations to set the group to hiera
instead of os-apply-config.
- The new format uses JSON instead of YAML. The hook actually writes
out the hiera JSON directly so no conversion takes place. Arrays,
Strings, Booleans all stay in their native formats. As such we can avoid
having to do many of the awkward string and list conversions in t-h-t to
support the previous YAML formatting.
- The new hook prefers JSON over YAML so upgrading users will have the
new files prefered. (we will post a cleanup routine for the old files
soon but this isn't a new behavior, JSON is now simply prefered.)
- A lot of services required edits to account for default settings that
worked in YAML that no longer work correctly in the native JSON
format. In almost all these cases I think the resulting codes looks
cleaner and is more explicit with regards to what is getting
configured in hiera on the actual nodes.
Depends-On: I6a383b1ad4ec29458569763bd3f56fd3f2bd726b
Closes-bug: #1596373
Change-Id: Ibe7e2044e200e2c947223286fdf4fd5bcf98c2e1
|
|
Provision the Keystone Fernet Token provider
by installing 2 keys with dynamic content
generated by python-tripleoclient.
Note that this only sets up the necessary keys to use fernet as a token
provider, however, this does not intend to set it up as the default
provider; This will be discussed and will come as part of another
commit.
Co-Authored-By: Juan Antonio Osorio Robles <jaosorior@redhat.com>
Depends-On: Ic070d160b519b8637997dbde165dbf15275e0dfe
Change-Id: Iaa5499614417000c1b9ba42a776a50cb22c1bb30
|
|
In order to eventually enable fernet tokens for keystone, we need to be
specify the token provider. This change codifies the current default
used by TripleO of uuid tokens and fernet token setup disabled.
Change-Id: I7c03ed7b6495d0b9a57986458d020b3e3bf7224a
Closes-Bug: #1641763
|
|
Closes-Bug: rhbz#1392428
Closes-Bug: #1640834
Change-Id: I2a1a869493ccb4c8d5b9aea26b8ef947750d2cfe
|
|
|
|
|
|
Change-Id: Ib25849565c617f32357ef545957f58454b2a53f8
|
|
This adds an environment file that can be used to enable TLS in
the internal endpoints via certmonger if used. This will include
a nested stack that will create the hash that will be used to
create the certmonger certificates.
When setting up a service over apache via puppet, we used to disable
explicitly ssl (which sets modd_ssl-related fields for that vhost).
We now make this depend on the EnableInternalTLS flag. This has only
been done for keystone, but more services will be added as the
puppet code lands
bp tls-via-certmonger
Depends-On: I303f6cf47859284785c0cdc65284a7eb89a4e039
Change-Id: I12e794f2d4076be9505dabfe456c1ca6cfbd359c
|
|
This patch movs the various db::mysql hiera settings into a
'mysql' specific service_config_settings section for each
service so that these will only get applied on the MySQL service
node. This follows a similar puppet-tripleo change where we
create the actual databases for all services locally on
the MySQL service node to avoid permission issues.
Change-Id: Ic0692b1f7aa8409699630ef3924c4be98ca6ffb2
Closes-bug: #1620595
Depends-On: I05cc0afa9373429a3197c194c3e8f784ae96de5f
Depends-On: I5e1ef2dc6de6f67d7c509e299855baec371f614d
|
|
In upstream puppet-keystone, the boostrap process should use an admin
password not the admin token for the bootstrapping of keystone. The
admin password option is being added to the upstream class so we will
need to provide it to properly have keystone bootstrapped.
Change-Id: Icab4b0cb70d6caf2f2792c4fe730f060b807fbc1
Depends-On: I7a706d93b43ec025bdb4b29667f64ff2f7dd52a0
Related-Bug: #1621959
|
|
Change-Id: I291bfb1e5736864ea504cd82eea1d4001fcdd931
|
|
Currently the servername is incorrectly set for the services running
over apache. It currently takes the default value which is just the
regular FQDN, when the services actually might be running on
different IPs that require alternative FQDNs.
This fixes that by filling that value from a fact in hiera that's
dependant on the service's network.
Closes-Bug: #1625677
Change-Id: Ib7ea5fd2d18a376eaa2f5a3fa5687cb9b719a8e2
|
|
This implements support for installing fluentd agents as a composable
service on the overcloud.
Depends-On: I2e1abe4d8c8359e56ff626255ee50c9cacca1940
Implements: tripleo-opstools-centralized-logging
Change-Id: I23b0e23881b742158fcfb6b8c145a3211d45086e
|
|
|
|
The keystone public_endpoint value should be deduced from the calling
request and not hardcoded, or it makes network isolation impossible.
Change-Id: Ide6a65aa9393cb84591b0015ec5966cc01ffbcf8
Closes-Bug: 1381961
|
|
Configure Keystone credentials by installing 2 keys with dynamic content
generated by python-tripleoclient.
Note: this is a first iteration of managing Keystone credentials. It has
a few limitations:
- keys are not exported to external storage.
- keys are not rotated automatically.
Change-Id: I45cf8821eadf528dfcdc8d74e6e0484597b0d2c0
|
|
- adds possibility to install sensu-client on all nodes
- each composable service has it's own subscription
Co-Authored-By: Emilien Macchi <emilien@redhat.com>
Co-Authored-By: Michele Baldessari <michele@redhat.com>
Implements: blueprint tripleo-opstools-availability-monitoring
Change-Id: I6a215763fd0f0015285b3573305d18d0f56c7770
|
|
This patch moves the settings for Nova, Neutron, and Horizon
out of controller.yaml.
Also fixes the NovaPassword settings in nova-base.yaml
so they don't use get_input.
Also, creates a new apache.yaml base service to contain shared
apache settings for several services which use Apache for WSGI.
Co-Authored-By: Giulio Fidente <gfidente@redhat.com>
Change-Id: I35d909bd5abc23976b5732a2b9af31cf1448838e
Related-bug: #1604414
|
|
This patch moves keystone, mongodb, and manila
bind host settings out of controller.yaml and into
composable services.
Change-Id: I1874dc47fffa30606107999da702442badde35c9
|
|
This patch adds a new DefaultPasswords parameter to
composable services. This is needed to help provide
access to top level password resources that overcloud.yaml
currently manages (passwords for Rabbit, Mysql, etc.).
Moving the RandomString resources into composable services
would cause them to regenerate within the stack. With this
approach we can leave them where they are while we deprecate
the top level mechanism and move the code that uses the
passwords into the composable services.
Change-Id: I4f21603c58a169a093962594e860933306879e3f
|
|
This will be needed to pick the network where the service has
to bind to from within the service template.
Change-Id: I52652e1ad8c7b360efd2c7af199e35932aaaea8c
|
|
Currently we have a hard-coded set of per-service parameters, which
will cause problems for custom roles and full composability.
As a first step towards making this more configurable, remove the
hard-coded per-service parameters from overcloud.yaml, and adjust
the EndpointMap generation to instead accept two mappings, the
ServiceNetMap and a mapping of networks to IPs (effectively this
just moves the map lookup inside the endpoint map instead of
inside overcloud.yaml)
Change-Id: Ib522e89c36eed2115a6586dd5a6770907d9b33db
Partially-Implements: blueprint custom-roles
|
|
These interfaces have all been deprecated by keystone, and we don't
offer any parameter interface to select PKI token format anyway,
so remove these to align with keystone reccomendations.
The keystone.conf.sample says these values may be silently ignored or
removed, so it seems reasonable to do the same here (parameter_defaults
should be ignored from old stacks).
Change-Id: Ic88d584863a98ed49fc335825fbfba7a52b0f14e
Depends-On: I8232262b928c91dcde7bea2f23fa2a7c2660719e
|
|
This is the THT part that brings us the next generation architecture
as described in the following spec:
https://review.openstack.org/#/c/299628/
Blueprint:
https://blueprints.launchpad.net/tripleo/+spec/ha-lightweight-architecture
So far we tested deployment + tripleo.sh --overcloud-pingtest and
failover + tripleo.sh --overcloud-pingtest
Note that many of the Pacemaker template files become redundant with
this change, but to simplify the process of getting this change landed,
those templates will not be removed until a future commit.
Depends-On: I5e7585c08675d8a4bd071523b94210d325d79b59
Change-Id: I00bccb2563c006f80baed623b64f1e17af20dd4e
Implements: blueprint ha-lightweight-architecture
Co-Author: cmsj@tenshu.net
|
|
The http_proxy_to_wsgi middleware was recently added to keystone as
default in the pipeline [1]. So this takes it into use instead of the
non-standard option we were using before, which will be deprecated.
We already enable this middleware for nova, cinder and heat.
[1] Iad628a863e55cbf20c89ef23ebc7527ba8e1a835
Depends-On: I0fec98a6e1d9c8be4d8b8df382b78ba2815790f9
Change-Id: I8c1b84adc828a2b8c9ea11c4e2b8349427b1b206
|
|
Migrate puppet/hieradata/*.yaml parameters to puppet/services/*.yaml
except for some services that are not composable yet.
Co-Authored-By: Juan Antonio Osorio Robles <jaosorior@redhat.com>
Change-Id: I7e5f8b18ee9aa63a1dffc6facaf88315b07d5fd7
|
|
Split out the firewall rules in puppet/hieradata/controller.yaml
into the composable services
Depends-On: Id370362ab57347b75b1ab25afda877885b047263
Change-Id: Icaecab100d3f278035fbbb3facb9bf6c62c76c03
|
|
This patch adds a new service_name section to each composable
service. We now have an explicit unit test check to ensure that
service_name exists in tools/yaml-validate.py.
This patch also wires service_names into hieradata on each
of the roles so that tools can access the deployed services locally
during deployment and upgrades.
Change-Id: I60861c5aa760534db3e314bba16a13b90ea72f0c
|
|
The current default doesn't get expanded to its real value correctly,
the new one should work better, it's according to hiera docs:
https://docs.puppet.com/hiera/3.2/variables.html
Change-Id: I8d1df131d2f1eee4ae0f725358d33a3ecfa3175a
|
|
Now that keystone is run with wsgi apache the KeystoneWorkers
param has no efffect [1]. This repurposes KeystoneWorkers to now
set wsgi::apache::workers and adds the
keystone::wsgi::apache::threads hardcoding to 1 (can overrride
via extraconfig data)
[1] https://github.com/openstack/puppet-keystone/blob/51c68fb127c28b9748b352871783c3ec6ef5c83b/manifests/init.pp#L396
Closes-Bug: 1598092
Change-Id: I503d1f0d3f4c56abfaf4609ea8290e78668013b6
|
|
Change-Id: I934aff4960588a3957b7a7bcc90385ae48d34d1e
|
|
By passing the MysqlVirtualIP via the EndpointMap we won't need it
to be provided as a parameter to the services.
This follows what is already happening for the glance registry
service with I9186e56cd4746a60e65dc5ac12e6595ac56505f0.
Change-Id: Iad2ab389bf64d0fc8b06eb0e7d29b5370ff27dff
Co-Authored-By: Juan Antonio Osorio Robles <jaosorior@redhat.com>
|
|
Adds new puppet and puppet pacemaker specific services for
Keystone.
The puppet manifests for keystone now live in puppet-tripleo.
Hiera settings are driven by the nested stack heat templates
and used to control puppet-keystone and puppet-tripleo
directly.
The Pacemaker template extends the default keystone service and
swaps in the pacemaker specific puppet-tripleo profile instead.
Change-Id: I8b30438a27e9d5ec4e7d335e0bd1a931a20b03a2
Depends-On: I2faf5a78db802549053ec41678bf83bf28108189
|