Skip to content

Commit

Permalink
BGP DTs
Browse files Browse the repository at this point in the history
Configuration #1 (kernel-routing-nbdb-driver):
- BGP
- Ceph and Networker nodes
- DVR
- expose-tenant-networks disabled
- default sync-interval
- kernel routing
- NBDB driver
- FRR: interface-level BGP peers
- FRR: different ASN per rack
- Tempest (neutron+octavia) + Tobiko (scenario+sanity+faults)

Configuration #2 (ovn-routing-nbdb-driver):
- BGP
- Networker nodes
- DVR
- expose-tenant-networks enabled
- sync-interval configured with high value
- OVN routing
- NBDB driver
- FRR: IP-level BGP peers
- FRR: different ASN per rack
- Tempest (neutron+octavia) + Tobiko (scenario+sanity+faults)

Configuration #3 (ovn-routing-nbdb-driver-nondvr):
- BGP
- Networker nodes
- DVR disabled
- expose-tenant-networks enabled
- default sync-interval
- OVN routing
- NBDB driver
- FRR: IP-level BGP peers
- FRR: different ASN per rack
- Tempest (neutron+octavia) + Tobiko (scenario+sanity+faults)

Configuration #4 (kernel-routing-sbdb-driver-ipv6):
- BGP
- Networker nodes
- IPv6 single-stack controlplane
- DVR
- expose-tenant-networks disabled
- default sync-interval
- kernel routing
- SBDB driver
- FRR: IP-level BGP peers
- FRR: different ASN per rack
- Tempest (neutron+octavia) + Tobiko (scenario+sanity+faults)
  • Loading branch information
eduolivares committed Jan 26, 2024
1 parent 7354503 commit 8da5a06
Show file tree
Hide file tree
Showing 4 changed files with 224 additions and 0 deletions.
57 changes: 57 additions & 0 deletions examples/dt/bgp/kernel-routing-nbdb-driver/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,57 @@
# RHOSO Deployed Topology - BGP using kernel routing and OVN NB DB driver

## General information
| Revision | Change | Date |
|--------: | :-------------------- | :--------------: |
| v0.1 | Initial publication | 2024-01-12 |

## Node topology
| Node role | bm/vm | amount |
| --------------------------------------------- | ----- | ------ |
| Openshift master/worker combo-node cluster | vm | 3 |
| Compute nodes | vm | 3 |
| Networker nodes | vm | 3 |
| Ceph nodes | vm | 3 |
| Leaf routers | vm | 6 |
| Spine routers | vm | 2 |
| External routers | vm | 1 |
| External VMs | vm | 1 |

## Services, enabled features and configurations
| Service | configuration | Lock-in coverage? |
| ------------------------------------------- | ---------------------------------------------------------------------------------------- | ------------------ |
| Neutron | ML2/OVN, DVR | Must have |
| Octavia | | Must have |
| OVN-BGP-Agent | kernel-routing, nbdb-driver expose-tenant-networks disabled, sync-interval default value | Must have |
| FRR | different ASN per rack, interface-level BGP peers | Must have |

## Considerations/Constraints
1. Virtual networks should be created to connect the VMs between them.
2. All the VMs that are neither Openstack nor Openshift nodes need to be
properly configured in order to support the BGP protocol.
3. The spine/leaf topology separates the overcloud nodes into different L2
network segments, called racks. Each rack includes one compute, one
networker, one ceph and two leafs.
4. A separate provisioning network is necessary to install Openstack on those
nodes.
5. Once Openstack is installed on them, controlplane and dataplane connectivity
between them and with external machines (extvm) is achieved using the BGP
protocol.
6. Tests are executed from the extvm machine, in order to verify external
connectivity.

## Testing tree
| Test framework | Stage to run | Special configuration | Test case to report |
| ------------------------ | ------------ | --------------------------------------------------------------------- | :----------------------: |
| Tempest/neutron+octavia | stage 6 | skips due to usupported features (multicast, FIP port-forwarding, etc)| scenario |
| Tobiko/Faults | stage 7 | | sanity, scenario, faults |

## Stages
All stages must be executed in the order listed below. Everything is required unless otherwise indicated.
1. [Pre-deployment: create virtual networks, virtual routers and nodes](bgp-pre-deplyment.md)
2. [Install the OpenStack K8S operators and their dependencies](../../common/)
3. [Configure networking and deploy the Openstack control plane](bgp-control-plane.md)
4. [Configure and deploy the initial data plane to prepare for Ceph installation](bgp-pre-data-plane)
5. [Update the control plane and finish deploying the data plane after Ceph has been installed](bgp-post-data-plane)
6. [Execute non destructive testing](bgp-tempest.md)
7. [Execute HA testing](bgp-tobiko.md)
57 changes: 57 additions & 0 deletions examples/dt/bgp/kernel-routing-sbdb-driver-ipv6/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,57 @@
# RHOSO Deployed Topology - BGP using kernel routing and OVN SB DB driver and IPv6 single-stack control plane

## General information
| Revision | Change | Date |
|--------: | :-------------------- | :--------------: |
| v0.1 | Initial publication | 2024-01-12 |

## Node topology
| Node role | bm/vm | amount |
| --------------------------------------------- | ----- | ------ |
| Openshift master/worker combo-node cluster | vm | 3 |
| Compute nodes | vm | 3 |
| Networker nodes | vm | 3 |
| Leaf routers | vm | 6 |
| Spine routers | vm | 2 |
| External routers | vm | 1 |
| External VMs | vm | 1 |

## Services, enabled features and configurations
| Service | configuration | Lock-in coverage? |
| ------------------------------------------- | ---------------------------------------------------------------------------------------- | ------------------ |
| Neutron | ML2/OVN, DVR | Must have |
| Octavia | | Must have |
| OVN-BGP-Agent | kernel-routing, sbdb-driver expose-tenant-networks disabled, sync-interval default value | Must have |
| FRR | different ASN per rack, IP-level BGP peers | Must have |

## Considerations/Constraints
1. Virtual networks should be created to connect the VMs between them.
2. All the VMs that are neither Openstack nor Openshift nodes need to be
properly configured in order to support the BGP protocol.
3. The spine/leaf topology separates the overcloud nodes into different L2
network segments, called racks. Each rack includes one compute, one
networker, one ceph and two leafs.
4. A separate provisioning network is necessary to install Openstack on those
nodes.
5. All the controlplane interfaces should be configured with IPv6 single-stack
5. Once Openstack is installed on the nodes, controlplane and dataplane
connectivity between them and with external machines (extvm) is achieved
using the BGP protocol.
6. Tests are executed from the extvm machine, in order to verify external
connectivity.

## Testing tree
| Test framework | Stage to run | Special configuration | Test case to report |
| ------------------------ | ------------ | --------------------------------------------------------------------- | :----------------------: |
| Tempest/neutron+octavia | stage 6 | skips due to usupported features (multicast, FIP port-forwarding, etc)| scenario |
| Tobiko/Faults | stage 7 | | sanity, scenario, faults |

## Stages
All stages must be executed in the order listed below. Everything is required unless otherwise indicated.
1. [Pre-deployment: create virtual networks, virtual routers and nodes](bgp-pre-deplyment.md)
2. [Install the OpenStack K8S operators and their dependencies](../../common/)
3. [Configure networking and deploy the Openstack control plane](bgp-control-plane-ipv6.md)
4. [Configure and deploy the initial data plane to prepare for Ceph installation](bgp-pre-data-plane)
5. [Update the control plane and finish deploying the data plane after Ceph has been installed](bgp-post-data-plane)
6. [Execute non destructive testing](bgp-tempest.md)
7. [Execute HA testing](bgp-tobiko.md)
55 changes: 55 additions & 0 deletions examples/dt/bgp/ovn-routing-nbdb-driver-nondvr/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,55 @@
# RHOSO Deployed Topology - BGP using OVN routing and OVN NB DB driver

## General information
| Revision | Change | Date |
|--------: | :-------------------- | :--------------: |
| v0.1 | Initial publication | 2024-01-12 |

## Node topology
| Node role | bm/vm | amount |
| --------------------------------------------- | ----- | ------ |
| Openshift master/worker combo-node cluster | vm | 3 |
| Compute nodes | vm | 3 |
| Networker nodes | vm | 3 |
| Leaf routers | vm | 6 |
| Spine routers | vm | 2 |
| External routers | vm | 1 |
| External VMs | vm | 1 |

## Services, enabled features and configurations
| Service | configuration | Lock-in coverage? |
| ------------------------------------------- | ---------------------------------------------------------------------------------------------- | ------------------ |
| Neutron | ML2/OVN, DVR disabled | Must have |
| Octavia | | Must have |
| OVN-BGP-Agent | ovn-routing, nbdb-driver expose-tenant-networks enabled, sync-interval default value | Must have |
| FRR | different ASN per rack, IP-level BGP peers | Must have |

## Considerations/Constraints
1. Virtual networks should be created to connect the VMs between them.
2. All the VMs that are neither Openstack nor Openshift nodes need to be
properly configured in order to support the BGP protocol.
3. The spine/leaf topology separates the overcloud nodes into different L2
network segments, called racks. Each rack includes one compute, one
networker and two leafs.
4. A separate provisioning network is necessary to install Openstack on those
nodes.
5. Once Openstack is installed on them, controlplane and dataplane connectivity
between them and with external machines (extvm) is achieved using the BGP
protocol.
6. Tests are executed from the extvm machine, in order to verify external
connectivity.

## Testing tree
| Test framework | Stage to run | Special configuration | Test case to report |
| ------------------------ | ------------ | --------------------------------------------------------------------- | :----------------------: |
| Tempest/neutron+octavia | stage 5 | skips due to usupported features (multicast, FIP port-forwarding, etc)| scenario |
| Tobiko/Faults | stage 6 | | sanity, scenario, faults |

## Stages
All stages must be executed in the order listed below. Everything is required unless otherwise indicated.
1. [Pre-deployment: create virtual networks, virtual routers and nodes](bgp-pre-deplyment.md)
2. [Install the OpenStack K8S operators and their dependencies](../../common/)
3. [Configure networking and deploy the Openstack control plane](bgp-control-plane.md)
4. [Configure and deploy the data plane](bgp-data-plane)
5. [Execute non destructive testing](bgp-tempest.md)
6. [Execute HA testing](bgp-tobiko.md)
55 changes: 55 additions & 0 deletions examples/dt/bgp/ovn-routing-nbdb-driver/README.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,55 @@
# RHOSO Deployed Topology - BGP using OVN routing and OVN NB DB driver

## General information
| Revision | Change | Date |
|--------: | :-------------------- | :--------------: |
| v0.1 | Initial publication | 2024-01-12 |

## Node topology
| Node role | bm/vm | amount |
| --------------------------------------------- | ----- | ------ |
| Openshift master/worker combo-node cluster | vm | 3 |
| Compute nodes | vm | 3 |
| Networker nodes | vm | 3 |
| Leaf routers | vm | 6 |
| Spine routers | vm | 2 |
| External routers | vm | 1 |
| External VMs | vm | 1 |

## Services, enabled features and configurations
| Service | configuration | Lock-in coverage? |
| ------------------------------------------- | ---------------------------------------------------------------------------------------------- | ------------------ |
| Neutron | ML2/OVN, DVR | Must have |
| Octavia | | Must have |
| OVN-BGP-Agent | ovn-routing, nbdb-driver expose-tenant-networks enabled, sync-interval set to a high value | Must have |
| FRR | different ASN per rack, IP-level BGP peers | Must have |

## Considerations/Constraints
1. Virtual networks should be created to connect the VMs between them.
2. All the VMs that are neither Openstack nor Openshift nodes need to be
properly configured in order to support the BGP protocol.
3. The spine/leaf topology separates the overcloud nodes into different L2
network segments, called racks. Each rack includes one compute, one
networker and two leafs.
4. A separate provisioning network is necessary to install Openstack on those
nodes.
5. Once Openstack is installed on them, controlplane and dataplane connectivity
between them and with external machines (extvm) is achieved using the BGP
protocol.
6. Tests are executed from the extvm machine, in order to verify external
connectivity.

## Testing tree
| Test framework | Stage to run | Special configuration | Test case to report |
| ------------------------ | ------------ | --------------------------------------------------------------------- | :----------------------: |
| Tempest/neutron+octavia | stage 5 | skips due to usupported features (multicast, FIP port-forwarding, etc)| scenario |
| Tobiko/Faults | stage 6 | | sanity, scenario, faults |

## Stages
All stages must be executed in the order listed below. Everything is required unless otherwise indicated.
1. [Pre-deployment: create virtual networks, virtual routers and nodes](bgp-pre-deplyment.md)
2. [Install the OpenStack K8S operators and their dependencies](../../common/)
3. [Configure networking and deploy the Openstack control plane](bgp-control-plane.md)
4. [Configure and deploy the data plane](bgp-data-plane)
5. [Execute non destructive testing](bgp-tempest.md)
6. [Execute HA testing](bgp-tobiko.md)

0 comments on commit 8da5a06

Please sign in to comment.