2017-05-17 12:52:21 |
Bogdan Dobrelya |
bug |
|
|
added bug |
2017-05-17 12:53:38 |
Bogdan Dobrelya |
tripleo: milestone |
|
pike-2 |
|
2017-05-17 12:53:43 |
Bogdan Dobrelya |
tripleo: status |
New |
In Progress |
|
2017-05-17 12:53:45 |
Bogdan Dobrelya |
tripleo: assignee |
|
Bogdan Dobrelya (bogdando) |
|
2017-05-17 12:53:48 |
Bogdan Dobrelya |
tripleo: importance |
Undecided |
Wishlist |
|
2017-05-17 12:53:58 |
Bogdan Dobrelya |
tags |
|
quickstart |
|
2017-05-17 12:58:29 |
Bogdan Dobrelya |
bug |
|
|
added subscriber James Slagle |
2017-05-19 08:03:13 |
Bogdan Dobrelya |
description |
Note, this is not related to OVB/CI devmode RFEs like https://bugs.launchpad.net/tripleo/+bug/1657191
The use case is:
* given a set of VM instances on generic openstack cloud provider, deployed either as a heat stack or the like (for example, like Traas does for RDO cloud) and given access creds/url to that cloud;
* given a dynamic openstack inventory generated by openstack.py from ansible repo;
* given a custom ansible play comprising quickstart/extras roles, with custom vars included.
* to deploy a multinode (like tripleo-ci multinode scenarios do for nodepool), based on the deployed-server quickstart-extras role;
* and having applied the following shortcuts for development environments/hacking/testing WIP patches:
* No images/containers/packages to be built, no VMs provisioned with Ironic, and no repos or images to be injected (the deployed-server extras role covers that case, it seems)
Caveats:
* No tripleo-ci/quickstart shell scripts used, only ansible-playbook commands wanted.
* Deployment split into the steps: an undercloud may be deployed as a separate, then deployment continued to finish multinode overcloud as well. |
Note, this is not related to OVB/CI devmode RFEs like https://bugs.launchpad.net/tripleo/+bug/1657191
The use case is:
* given a set of VM instances on generic openstack cloud provider, deployed either as a heat stack or the like (for example, like Traas does for RDO cloud) and given access creds/url to that cloud;
* given a dynamic openstack inventory generated by openstack.py from ansible repo;
* given a custom ansible play comprising quickstart/extras roles, with custom vars included.
* to deploy a multinode (like tripleo-ci multinode scenarios do for nodepool), based on the deployed-server quickstart-extras role;
* and having applied the following shortcuts for development environments/hacking/testing WIP patches:
* No images/containers/packages to be built, no VMs provisioned with Ironic, and no repos or images to be injected (the deployed-server extras role covers that case, it seems)
* Dependencies should not be cloned & built as artifacts with zuul_cloner et al. Instead, custom undercloud/overcloud install scripts are expected to contain all the steps needed to get those installed from dev forks into the dev env being deployed, see an example https://github.com/bogdando/tripleo-quickstart-extras/blob/dev/roles/undercloud-deploy/templates/undercloud-deploy-dev.sh.j2#L53
Caveats:
* No tripleo-ci/quickstart shell scripts used, only ansible-playbook commands wanted.
* Deployment split into the steps: an undercloud may be deployed as a separate, then deployment continued to finish multinode overcloud as well. |
|
2017-05-22 15:16:42 |
Bogdan Dobrelya |
summary |
[RFE] openstack provider for multinode development environments |
[RFE][deployed-server] openstack provider for multinode development environments |
|
2017-05-22 15:17:55 |
Bogdan Dobrelya |
description |
Note, this is not related to OVB/CI devmode RFEs like https://bugs.launchpad.net/tripleo/+bug/1657191
The use case is:
* given a set of VM instances on generic openstack cloud provider, deployed either as a heat stack or the like (for example, like Traas does for RDO cloud) and given access creds/url to that cloud;
* given a dynamic openstack inventory generated by openstack.py from ansible repo;
* given a custom ansible play comprising quickstart/extras roles, with custom vars included.
* to deploy a multinode (like tripleo-ci multinode scenarios do for nodepool), based on the deployed-server quickstart-extras role;
* and having applied the following shortcuts for development environments/hacking/testing WIP patches:
* No images/containers/packages to be built, no VMs provisioned with Ironic, and no repos or images to be injected (the deployed-server extras role covers that case, it seems)
* Dependencies should not be cloned & built as artifacts with zuul_cloner et al. Instead, custom undercloud/overcloud install scripts are expected to contain all the steps needed to get those installed from dev forks into the dev env being deployed, see an example https://github.com/bogdando/tripleo-quickstart-extras/blob/dev/roles/undercloud-deploy/templates/undercloud-deploy-dev.sh.j2#L53
Caveats:
* No tripleo-ci/quickstart shell scripts used, only ansible-playbook commands wanted.
* Deployment split into the steps: an undercloud may be deployed as a separate, then deployment continued to finish multinode overcloud as well. |
Note, this is not related to OVB/CI devmode RFEs like https://bugs.launchpad.net/tripleo/+bug/1657191
The use case is:
* given a set of VM instances on generic openstack cloud provider, deployed either as a heat stack or the like (for example, like Traas does for RDO cloud) and given access creds/url to that cloud;
* given a dynamic openstack inventory generated by openstack.py from ansible repo;
* given a custom ansible play comprising quickstart/extras roles, with custom vars included,
* to DEPLOY a multinode (like tripleo-ci multinode scenarios do for nodepool, but here VM nodes pre-provisioned by hosting openstack cloud), based on the deployed-server quickstart-extras role;
* and having applied the following shortcuts for development environments/hacking/testing WIP patches:
* No images/containers/packages to be built, no VMs provisioned with Ironic, and no repos or images to be injected (the deployed-server extras role covers that case, it seems)
* Dependencies should not be cloned & built as artifacts with zuul_cloner et al. Instead, custom undercloud/overcloud install scripts are expected to contain all the steps needed to get those installed from dev forks into the dev env being deployed, see an example https://github.com/bogdando/tripleo-quickstart-extras/blob/dev/roles/undercloud-deploy/templates/undercloud-deploy-dev.sh.j2#L53
Caveats:
* No tripleo-ci/quickstart shell scripts used, only ansible-playbook commands wanted.
* Deployment split into the steps: an undercloud may be deployed as a separate, then deployment continued to finish multinode overcloud as well. |
|
2017-05-23 12:17:06 |
Bogdan Dobrelya |
description |
Note, this is not related to OVB/CI devmode RFEs like https://bugs.launchpad.net/tripleo/+bug/1657191
The use case is:
* given a set of VM instances on generic openstack cloud provider, deployed either as a heat stack or the like (for example, like Traas does for RDO cloud) and given access creds/url to that cloud;
* given a dynamic openstack inventory generated by openstack.py from ansible repo;
* given a custom ansible play comprising quickstart/extras roles, with custom vars included,
* to DEPLOY a multinode (like tripleo-ci multinode scenarios do for nodepool, but here VM nodes pre-provisioned by hosting openstack cloud), based on the deployed-server quickstart-extras role;
* and having applied the following shortcuts for development environments/hacking/testing WIP patches:
* No images/containers/packages to be built, no VMs provisioned with Ironic, and no repos or images to be injected (the deployed-server extras role covers that case, it seems)
* Dependencies should not be cloned & built as artifacts with zuul_cloner et al. Instead, custom undercloud/overcloud install scripts are expected to contain all the steps needed to get those installed from dev forks into the dev env being deployed, see an example https://github.com/bogdando/tripleo-quickstart-extras/blob/dev/roles/undercloud-deploy/templates/undercloud-deploy-dev.sh.j2#L53
Caveats:
* No tripleo-ci/quickstart shell scripts used, only ansible-playbook commands wanted.
* Deployment split into the steps: an undercloud may be deployed as a separate, then deployment continued to finish multinode overcloud as well. |
Note, this is not related to OVB/CI devmode RFEs like https://bugs.launchpad.net/tripleo/+bug/1657191
This extends the https://blueprints.launchpad.net/tripleo/+spec/split-stack-software-configuration (Split Stack Phase 2): "deploy OpenStack on already deployed baremetal nodes. We have this functionality today with the deployed-server templates, but we need to harden this work and make it more consumable outside of CI." as the following below.
The use case is:
* given a set of VM instances on generic openstack cloud provider, deployed either as a heat stack or the like (for example, like Traas does for RDO cloud) and given access creds/url to that cloud;
* given a dynamic openstack inventory generated by openstack.py from ansible repo;
* given a custom ansible play comprising quickstart/extras roles, with custom vars included,
* to DEPLOY a multinode (like tripleo-ci multinode scenarios do for nodepool, but here VM nodes pre-provisioned by hosting openstack cloud), based on the deployed-server quickstart-extras role;
* and having applied the following shortcuts for development environments/hacking/testing WIP patches:
* No images/containers/packages to be built, no VMs provisioned with Ironic, and no repos or images to be injected (the deployed-server extras role covers that case, it seems)
* Dependencies should not be cloned & built as artifacts with zuul_cloner et al. Instead, custom undercloud/overcloud install scripts are expected to contain all the steps needed to get those installed from dev forks into the dev env being deployed, see an example https://github.com/bogdando/tripleo-quickstart-extras/blob/dev/roles/undercloud-deploy/templates/undercloud-deploy-dev.sh.j2#L53
Caveats:
* No tripleo-ci/quickstart shell scripts used, only ansible-playbook commands wanted.
* Deployment split into the steps: an undercloud may be deployed as a separate, then deployment continued to finish multinode overcloud as well. |
|
2017-05-23 12:17:17 |
Bogdan Dobrelya |
summary |
[RFE][deployed-server] openstack provider for multinode development environments |
[RFE][deployed-server][split-stack] openstack provider for multinode development environments |
|
2017-05-25 07:44:12 |
Bogdan Dobrelya |
description |
Note, this is not related to OVB/CI devmode RFEs like https://bugs.launchpad.net/tripleo/+bug/1657191
This extends the https://blueprints.launchpad.net/tripleo/+spec/split-stack-software-configuration (Split Stack Phase 2): "deploy OpenStack on already deployed baremetal nodes. We have this functionality today with the deployed-server templates, but we need to harden this work and make it more consumable outside of CI." as the following below.
The use case is:
* given a set of VM instances on generic openstack cloud provider, deployed either as a heat stack or the like (for example, like Traas does for RDO cloud) and given access creds/url to that cloud;
* given a dynamic openstack inventory generated by openstack.py from ansible repo;
* given a custom ansible play comprising quickstart/extras roles, with custom vars included,
* to DEPLOY a multinode (like tripleo-ci multinode scenarios do for nodepool, but here VM nodes pre-provisioned by hosting openstack cloud), based on the deployed-server quickstart-extras role;
* and having applied the following shortcuts for development environments/hacking/testing WIP patches:
* No images/containers/packages to be built, no VMs provisioned with Ironic, and no repos or images to be injected (the deployed-server extras role covers that case, it seems)
* Dependencies should not be cloned & built as artifacts with zuul_cloner et al. Instead, custom undercloud/overcloud install scripts are expected to contain all the steps needed to get those installed from dev forks into the dev env being deployed, see an example https://github.com/bogdando/tripleo-quickstart-extras/blob/dev/roles/undercloud-deploy/templates/undercloud-deploy-dev.sh.j2#L53
Caveats:
* No tripleo-ci/quickstart shell scripts used, only ansible-playbook commands wanted.
* Deployment split into the steps: an undercloud may be deployed as a separate, then deployment continued to finish multinode overcloud as well. |
Note, this is not related to OVB/CI devmode RFEs like https://bugs.launchpad.net/tripleo/+bug/1657191
This extends the https://blueprints.launchpad.net/tripleo/+spec/split-stack-software-configuration (Split Stack Phase 2): "deploy OpenStack on already deployed baremetal nodes. We have this functionality today with the deployed-server templates, but we need to harden this work and make it more consumable outside of CI." as the following below.
The user story is "Infrastructure scripts are doing many different things that I am as a developer don't need/want to do"
The use case is:
* given a set of VM instances on generic openstack cloud provider, deployed either as a heat stack or the like (for example, like Traas does for RDO cloud) and given access creds/url to that cloud;
* given a dynamic openstack inventory generated by openstack.py from ansible repo;
* given a custom ansible play comprising quickstart/extras roles, with custom vars included,
* to DEPLOY a multinode (like tripleo-ci multinode scenarios do for nodepool, but here VM nodes pre-provisioned by hosting openstack cloud), based on the deployed-server quickstart-extras role;
* and having applied the following shortcuts for development environments/hacking/testing WIP patches:
* No images/containers/packages to be built, no VMs provisioned with Ironic, and no repos or images to be injected (the deployed-server extras role covers that case, it seems)
* Dependencies should not be cloned & built as artifacts with zuul_cloner et al. Instead, custom undercloud/overcloud install scripts are expected to contain all the steps needed to get those installed from dev forks into the dev env being deployed, see an example https://github.com/bogdando/tripleo-quickstart-extras/blob/dev/roles/undercloud-deploy/templates/undercloud-deploy-dev.sh.j2#L53
Caveats:
* No tripleo-ci/quickstart shell scripts used, only ansible-playbook commands wanted.
* Deployment split into the steps: an undercloud may be deployed as a separate, then deployment continued to finish multinode overcloud as well. |
|
2017-05-25 07:44:40 |
Bogdan Dobrelya |
tags |
quickstart |
deployment-time quickstart |
|
2017-06-08 20:37:26 |
Emilien Macchi |
tripleo: milestone |
pike-2 |
pike-3 |
|
2017-06-26 14:00:59 |
Bogdan Dobrelya |
description |
Note, this is not related to OVB/CI devmode RFEs like https://bugs.launchpad.net/tripleo/+bug/1657191
This extends the https://blueprints.launchpad.net/tripleo/+spec/split-stack-software-configuration (Split Stack Phase 2): "deploy OpenStack on already deployed baremetal nodes. We have this functionality today with the deployed-server templates, but we need to harden this work and make it more consumable outside of CI." as the following below.
The user story is "Infrastructure scripts are doing many different things that I am as a developer don't need/want to do"
The use case is:
* given a set of VM instances on generic openstack cloud provider, deployed either as a heat stack or the like (for example, like Traas does for RDO cloud) and given access creds/url to that cloud;
* given a dynamic openstack inventory generated by openstack.py from ansible repo;
* given a custom ansible play comprising quickstart/extras roles, with custom vars included,
* to DEPLOY a multinode (like tripleo-ci multinode scenarios do for nodepool, but here VM nodes pre-provisioned by hosting openstack cloud), based on the deployed-server quickstart-extras role;
* and having applied the following shortcuts for development environments/hacking/testing WIP patches:
* No images/containers/packages to be built, no VMs provisioned with Ironic, and no repos or images to be injected (the deployed-server extras role covers that case, it seems)
* Dependencies should not be cloned & built as artifacts with zuul_cloner et al. Instead, custom undercloud/overcloud install scripts are expected to contain all the steps needed to get those installed from dev forks into the dev env being deployed, see an example https://github.com/bogdando/tripleo-quickstart-extras/blob/dev/roles/undercloud-deploy/templates/undercloud-deploy-dev.sh.j2#L53
Caveats:
* No tripleo-ci/quickstart shell scripts used, only ansible-playbook commands wanted.
* Deployment split into the steps: an undercloud may be deployed as a separate, then deployment continued to finish multinode overcloud as well. |
Note, this is not related to OVB/CI devmode RFEs like https://bugs.launchpad.net/tripleo/+bug/1657191
This extends the https://blueprints.launchpad.net/tripleo/+spec/split-stack-software-configuration (Split Stack Phase 2): "deploy OpenStack on already deployed baremetal nodes. We have this functionality today with the deployed-server templates, but we need to harden this work and make it more consumable outside of CI." as the following below.
The user story is "Infrastructure scripts are doing many different things that I am as a developer don't need/want to do"
The use case is:
* given a set of VM instances on generic openstack cloud provider, deployed either as a heat stack or the like (for example, like Traas does for RDO cloud) and given access creds/url to that cloud;
* given a dynamic openstack inventory generated by openstack.py from ansible repo;
* given a custom ansible play comprising quickstart/extras roles, with custom vars included,
* to DEPLOY a multinode (like tripleo-ci multinode scenarios do for nodepool, but here VM nodes pre-provisioned by hosting openstack cloud), based on the deployed-server quickstart-extras role;
* and having applied the following shortcuts for development environments/hacking/testing WIP patches:
* No images/containers/packages to be built, no VMs provisioned with Ironic, and no repos or images to be injected (the deployed-server extras role covers that case, it seems)
* Dependencies should not be cloned & built as artifacts with zuul_cloner et al. Instead, custom undercloud/overcloud install scripts are expected to contain all the steps needed to get those installed from dev forks into the dev env being deployed, see an example https://github.com/bogdando/tripleo-quickstart-extras/blob/dev/roles/undercloud-deploy/templates/undercloud-deploy-dev.sh.j2#L53
Caveats:
* No tripleo-ci/quickstart shell scripts used, only ansible-playbook commands wanted.
* Deployment split into the steps: an undercloud may be deployed as a separate, then deployment continued to finish multinode overcloud as well.
* Ansible-playbook commands may overlap for long running undercloud deploy and overcloud-prep tasks, bringing a drastic shortcut to deploy time. Example https://github.com/bogdando/oooq-warp/commit/d0b068f66b8ff92c5ba34541f24f235ab7dd0c97 |
|
2017-06-26 14:01:29 |
Bogdan Dobrelya |
description |
Note, this is not related to OVB/CI devmode RFEs like https://bugs.launchpad.net/tripleo/+bug/1657191
This extends the https://blueprints.launchpad.net/tripleo/+spec/split-stack-software-configuration (Split Stack Phase 2): "deploy OpenStack on already deployed baremetal nodes. We have this functionality today with the deployed-server templates, but we need to harden this work and make it more consumable outside of CI." as the following below.
The user story is "Infrastructure scripts are doing many different things that I am as a developer don't need/want to do"
The use case is:
* given a set of VM instances on generic openstack cloud provider, deployed either as a heat stack or the like (for example, like Traas does for RDO cloud) and given access creds/url to that cloud;
* given a dynamic openstack inventory generated by openstack.py from ansible repo;
* given a custom ansible play comprising quickstart/extras roles, with custom vars included,
* to DEPLOY a multinode (like tripleo-ci multinode scenarios do for nodepool, but here VM nodes pre-provisioned by hosting openstack cloud), based on the deployed-server quickstart-extras role;
* and having applied the following shortcuts for development environments/hacking/testing WIP patches:
* No images/containers/packages to be built, no VMs provisioned with Ironic, and no repos or images to be injected (the deployed-server extras role covers that case, it seems)
* Dependencies should not be cloned & built as artifacts with zuul_cloner et al. Instead, custom undercloud/overcloud install scripts are expected to contain all the steps needed to get those installed from dev forks into the dev env being deployed, see an example https://github.com/bogdando/tripleo-quickstart-extras/blob/dev/roles/undercloud-deploy/templates/undercloud-deploy-dev.sh.j2#L53
Caveats:
* No tripleo-ci/quickstart shell scripts used, only ansible-playbook commands wanted.
* Deployment split into the steps: an undercloud may be deployed as a separate, then deployment continued to finish multinode overcloud as well.
* Ansible-playbook commands may overlap for long running undercloud deploy and overcloud-prep tasks, bringing a drastic shortcut to deploy time. Example https://github.com/bogdando/oooq-warp/commit/d0b068f66b8ff92c5ba34541f24f235ab7dd0c97 |
Note, this is not related to OVB/CI devmode RFEs like https://bugs.launchpad.net/tripleo/+bug/1657191
This extends the https://blueprints.launchpad.net/tripleo/+spec/split-stack-software-configuration (Split Stack Phase 2): "deploy OpenStack on already deployed baremetal nodes. We have this functionality today with the deployed-server templates, but we need to harden this work and make it more consumable outside of CI." as the following below.
The user story is "Infrastructure scripts are doing many different things that I am as a developer don't need/want to do"
The use case is:
* given a set of VM instances on generic openstack cloud provider, deployed either as a heat stack or the like (for example, like Traas does for RDO cloud) and given access creds/url to that cloud;
* given a dynamic openstack inventory generated by openstack.py from ansible repo;
* given a custom ansible play comprising quickstart/extras roles, with custom vars included,
* to DEPLOY a multinode (like tripleo-ci multinode scenarios do for nodepool, but here VM nodes pre-provisioned by hosting openstack cloud), based on the deployed-server quickstart-extras role;
* and having applied the following shortcuts for development environments/hacking/testing WIP patches:
* No images/containers/packages to be built, no VMs provisioned with Ironic, and no repos or images to be injected (the deployed-server extras role covers that case, it seems)
* Dependencies should not be cloned & built as artifacts with zuul_cloner et al. Instead, custom undercloud/overcloud install scripts are expected to contain all the steps needed to get those installed from dev forks into the dev env being deployed, see an example https://github.com/bogdando/tripleo-quickstart-extras/blob/dev/roles/undercloud-deploy/templates/undercloud-deploy-dev.sh.j2#L53
Caveats:
* No tripleo-ci/quickstart shell scripts used, only ansible-playbook commands wanted.
* Deployment split into the steps: an undercloud may be deployed as a separate, then deployment continued to finish multinode overcloud as well.
* Ansible-playbook commands may overlap for long running undercloud deploy and overcloud-prep and repo-setup tasks, bringing a drastic shortcut to deploy time. Example https://github.com/bogdando/oooq-warp/commit/d0b068f66b8ff92c5ba34541f24f235ab7dd0c97 |
|
2017-06-26 14:02:10 |
Bogdan Dobrelya |
summary |
[RFE][deployed-server][split-stack] openstack provider for multinode development environments |
[RFE][deployed-server][split-stack] openstack provider for multinode/pre-provisioned development environments |
|
2017-07-30 04:58:10 |
Emilien Macchi |
tripleo: milestone |
pike-3 |
pike-rc1 |
|
2017-07-31 15:32:12 |
Emilien Macchi |
tripleo: milestone |
pike-rc1 |
queens-1 |
|
2017-08-01 08:53:12 |
Bogdan Dobrelya |
description |
Note, this is not related to OVB/CI devmode RFEs like https://bugs.launchpad.net/tripleo/+bug/1657191
This extends the https://blueprints.launchpad.net/tripleo/+spec/split-stack-software-configuration (Split Stack Phase 2): "deploy OpenStack on already deployed baremetal nodes. We have this functionality today with the deployed-server templates, but we need to harden this work and make it more consumable outside of CI." as the following below.
The user story is "Infrastructure scripts are doing many different things that I am as a developer don't need/want to do"
The use case is:
* given a set of VM instances on generic openstack cloud provider, deployed either as a heat stack or the like (for example, like Traas does for RDO cloud) and given access creds/url to that cloud;
* given a dynamic openstack inventory generated by openstack.py from ansible repo;
* given a custom ansible play comprising quickstart/extras roles, with custom vars included,
* to DEPLOY a multinode (like tripleo-ci multinode scenarios do for nodepool, but here VM nodes pre-provisioned by hosting openstack cloud), based on the deployed-server quickstart-extras role;
* and having applied the following shortcuts for development environments/hacking/testing WIP patches:
* No images/containers/packages to be built, no VMs provisioned with Ironic, and no repos or images to be injected (the deployed-server extras role covers that case, it seems)
* Dependencies should not be cloned & built as artifacts with zuul_cloner et al. Instead, custom undercloud/overcloud install scripts are expected to contain all the steps needed to get those installed from dev forks into the dev env being deployed, see an example https://github.com/bogdando/tripleo-quickstart-extras/blob/dev/roles/undercloud-deploy/templates/undercloud-deploy-dev.sh.j2#L53
Caveats:
* No tripleo-ci/quickstart shell scripts used, only ansible-playbook commands wanted.
* Deployment split into the steps: an undercloud may be deployed as a separate, then deployment continued to finish multinode overcloud as well.
* Ansible-playbook commands may overlap for long running undercloud deploy and overcloud-prep and repo-setup tasks, bringing a drastic shortcut to deploy time. Example https://github.com/bogdando/oooq-warp/commit/d0b068f66b8ff92c5ba34541f24f235ab7dd0c97 |
Note, this is not related to OVB/CI devmode RFEs like https://bugs.launchpad.net/tripleo/+bug/1657191
This extends the https://blueprints.launchpad.net/tripleo/+spec/split-stack-software-configuration (Split Stack Phase 2): "deploy OpenStack on already deployed baremetal nodes. We have this functionality today with the deployed-server templates, but we need to harden this work and make it more consumable outside of CI." as the following below.
There is also a nice write-up explaining the use case https://blog-slagle.rhcloud.com/?p=299
The user story is "Infrastructure scripts are doing many different things that I am as a developer don't need/want to do"
The use case is:
* given a set of VM instances on generic openstack cloud provider, deployed either as a heat stack or the like (for example, like Traas does for RDO cloud) and given access creds/url to that cloud;
* given a dynamic openstack inventory generated by openstack.py from ansible repo;
* given a custom ansible play comprising quickstart/extras roles, with custom vars included,
* to DEPLOY a multinode (like tripleo-ci multinode scenarios do for nodepool, but here VM nodes pre-provisioned by hosting openstack cloud), based on the deployed-server quickstart-extras role;
* and having applied the following shortcuts for development environments/hacking/testing WIP patches:
* No images/containers/packages to be built, no VMs provisioned with Ironic, and no repos or images to be injected (the deployed-server extras role covers that case, it seems)
* Dependencies should not be cloned & built as artifacts with zuul_cloner et al. Instead, custom undercloud/overcloud install scripts are expected to contain all the steps needed to get those installed from dev forks into the dev env being deployed, see an example https://github.com/bogdando/tripleo-quickstart-extras/blob/dev/roles/undercloud-deploy/templates/undercloud-deploy-dev.sh.j2#L53
Caveats:
* No tripleo-ci/quickstart shell scripts used, only ansible-playbook commands wanted.
* Deployment split into the steps: an undercloud may be deployed as a separate, then deployment continued to finish multinode overcloud as well.
* Ansible-playbook commands may overlap for long running undercloud deploy and overcloud-prep and repo-setup tasks, bringing a drastic shortcut to deploy time. Example https://github.com/bogdando/oooq-warp/commit/d0b068f66b8ff92c5ba34541f24f235ab7dd0c97 |
|
2017-08-01 08:54:02 |
Bogdan Dobrelya |
description |
Note, this is not related to OVB/CI devmode RFEs like https://bugs.launchpad.net/tripleo/+bug/1657191
This extends the https://blueprints.launchpad.net/tripleo/+spec/split-stack-software-configuration (Split Stack Phase 2): "deploy OpenStack on already deployed baremetal nodes. We have this functionality today with the deployed-server templates, but we need to harden this work and make it more consumable outside of CI." as the following below.
There is also a nice write-up explaining the use case https://blog-slagle.rhcloud.com/?p=299
The user story is "Infrastructure scripts are doing many different things that I am as a developer don't need/want to do"
The use case is:
* given a set of VM instances on generic openstack cloud provider, deployed either as a heat stack or the like (for example, like Traas does for RDO cloud) and given access creds/url to that cloud;
* given a dynamic openstack inventory generated by openstack.py from ansible repo;
* given a custom ansible play comprising quickstart/extras roles, with custom vars included,
* to DEPLOY a multinode (like tripleo-ci multinode scenarios do for nodepool, but here VM nodes pre-provisioned by hosting openstack cloud), based on the deployed-server quickstart-extras role;
* and having applied the following shortcuts for development environments/hacking/testing WIP patches:
* No images/containers/packages to be built, no VMs provisioned with Ironic, and no repos or images to be injected (the deployed-server extras role covers that case, it seems)
* Dependencies should not be cloned & built as artifacts with zuul_cloner et al. Instead, custom undercloud/overcloud install scripts are expected to contain all the steps needed to get those installed from dev forks into the dev env being deployed, see an example https://github.com/bogdando/tripleo-quickstart-extras/blob/dev/roles/undercloud-deploy/templates/undercloud-deploy-dev.sh.j2#L53
Caveats:
* No tripleo-ci/quickstart shell scripts used, only ansible-playbook commands wanted.
* Deployment split into the steps: an undercloud may be deployed as a separate, then deployment continued to finish multinode overcloud as well.
* Ansible-playbook commands may overlap for long running undercloud deploy and overcloud-prep and repo-setup tasks, bringing a drastic shortcut to deploy time. Example https://github.com/bogdando/oooq-warp/commit/d0b068f66b8ff92c5ba34541f24f235ab7dd0c97 |
Note, this is not related to OVB/CI devmode RFEs like https://bugs.launchpad.net/tripleo/+bug/1657191
This extends the https://blueprints.launchpad.net/tripleo/+spec/split-stack-software-configuration (Split Stack Phase 2): "deploy OpenStack on already deployed baremetal nodes. We have this functionality today with the deployed-server templates, but we need to harden this work and make it more consumable outside of CI." as the following below.
There is also a nice write-up explaining the use case https://blog-slagle.rhcloud.com/?p=299
The user story is "Infrastructure scripts are doing many different things that I am as a developer don't need/want to do. Instead I want to deploy overcloud openstack on pre-provisioned VMs, using ansible playbooks only (composed from quickstart roles)"
The use case is:
* given a set of VM instances on generic openstack cloud provider, deployed either as a heat stack or the like (for example, like Traas does for RDO cloud) and given access creds/url to that cloud;
* given a dynamic openstack inventory generated by openstack.py from ansible repo;
* given a custom ansible play comprising quickstart/extras roles, with custom vars included,
* to DEPLOY a multinode (like tripleo-ci multinode scenarios do for nodepool, but here VM nodes pre-provisioned by hosting openstack cloud), based on the deployed-server quickstart-extras role;
* and having applied the following shortcuts for development environments/hacking/testing WIP patches:
* No images/containers/packages to be built, no VMs provisioned with Ironic, and no repos or images to be injected (the deployed-server extras role covers that case, it seems)
* Dependencies should not be cloned & built as artifacts with zuul_cloner et al. Instead, custom undercloud/overcloud install scripts are expected to contain all the steps needed to get those installed from dev forks into the dev env being deployed, see an example https://github.com/bogdando/tripleo-quickstart-extras/blob/dev/roles/undercloud-deploy/templates/undercloud-deploy-dev.sh.j2#L53
Caveats:
* No tripleo-ci/quickstart shell scripts used, only ansible-playbook commands wanted.
* Deployment split into the steps: an undercloud may be deployed as a separate, then deployment continued to finish multinode overcloud as well.
* Ansible-playbook commands may overlap for long running undercloud deploy and overcloud-prep and repo-setup tasks, bringing a drastic shortcut to deploy time. Example https://github.com/bogdando/oooq-warp/commit/d0b068f66b8ff92c5ba34541f24f235ab7dd0c97 |
|
2017-10-23 08:57:41 |
Bogdan Dobrelya |
tags |
deployment-time quickstart |
deployment-time quickstart ux |
|
2017-10-23 08:58:42 |
Bogdan Dobrelya |
tripleo: status |
In Progress |
Fix Committed |
|
2017-10-23 09:00:12 |
Bogdan Dobrelya |
nominated for series |
|
tripleo/pike |
|
2017-10-23 09:00:12 |
Bogdan Dobrelya |
bug task added |
|
tripleo/pike |
|
2017-10-23 09:00:12 |
Bogdan Dobrelya |
nominated for series |
|
tripleo/queens |
|
2017-10-23 09:00:12 |
Bogdan Dobrelya |
bug task added |
|
tripleo/queens |
|
2017-10-23 09:00:21 |
Bogdan Dobrelya |
tripleo/pike: status |
New |
Fix Committed |
|
2017-10-23 09:00:24 |
Bogdan Dobrelya |
tripleo/pike: importance |
Undecided |
Wishlist |
|
2017-10-23 09:00:26 |
Bogdan Dobrelya |
tripleo/pike: assignee |
|
Bogdan Dobrelya (bogdando) |
|
2017-10-23 16:09:28 |
Emilien Macchi |
tripleo/queens: status |
Fix Committed |
Fix Released |
|
2017-10-23 16:09:30 |
Emilien Macchi |
tripleo/pike: status |
Fix Committed |
Fix Released |
|