Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/giannisalinetti/os-inventory

Inventory generator for OpenShift advanced installation
https://github.com/giannisalinetti/os-inventory

Last synced: 2 months ago
JSON representation

Inventory generator for OpenShift advanced installation

Awesome Lists containing this project

README 

        
# os-inventory: an OpenShift installation inventory generator



## DISCLAIMER: This project is deprecated. 

Red Hat OpenShift 4 drop the Ansible deployment method and introduces a totally 

new approach in the installation of OpenShift clusters based on Terraform and a 

combination of tecniques which leverage on Red Hat CoreOS **Ignition Configs** 

on boostrap nodes and control plane nodes.

Ansible roles are now only related to worker nodes based on RHEL.



## Why?



Red Hat OpenShift Container Platform is installed using Ansible. The community project

https://github.com/openshift/openshift-ansible provides related playbooks and roles.

To deploy OpenShift a lot of parameters can be set using inventory variables.

To speed up the deployment process, a tool that creates inventories from a template can

be really helpful. Written in Go, si single binary tool can generate custom inventories

starting from a simple yaml file with a minimun set of parameters.

Default values are provided to generate a basic inventory from scratch.

Despite being a very simple tool, os-inventory implements a full working CLI that can be

extended with new features in the future, yet maintaining the KISS phylosophy.

A basic sanity check on data, especially strings, is done to avoid errors on install time.



## Build and Install



To build and install os-inventory:



```

$ make

$ sudo make install

```



This will run tests and compile the binary install it under */usr/local/bin* on your system.



## Usage



To get the generic help and overall description simply run the command without arguments:



```

$ os-inventory

```



To print a specific command help:



```

$ os-inventory  --help

```



To dump all the default values in YAML format to stdout:



```

$ os-inventory defaults

```



To generate an inventory and dump it to stdout:



```

$ os-inventory generate

```



To generate the inventory loading configurations from a YAML file:



```

$ os-inventory generate -f examples/example.yml

```



The generated inventory is redirected by defayult to stdout. To create a custom file

use the options *-o* or *--output*:



```

$ os-inventory generate -f example.yml -o /tmp/myinventory

```



## The YAML config



The YAML configuration file can include a different number of parameters:



```

---

deployment_type: "enterprise"

multi_master: true

haproxy_enabled: true

masters_list:

    - m1.example.com

    - m2.example.com

    - m3.example.com

etcd_list:

    - m1.example.com

    - m2.example.com

    - m3.example.com

lb_list:

    - lb.example.com

nodes_map:

    m1.example.com: ""

    m2.example.com: ""

    m3.example.com: ""

    n1.example.com: "openshift_node_labes=\"{'region': 'infra'}\""

    n2.example.com: "openshift_node_labes=\"{'region': 'infra'}\""

    n3.example.com: "openshift_node_labes=\"{'region': 'infra'}\""

    n4.example.com: "openshift_node_labes=\"{'region': 'primary'}\""

    n5.example.com: "openshift_node_labes=\"{'region': 'primary'}\""

```



A more minimalistic config file can be produces, maybe with just one custom parameter:



```

---

deployment_type: "origin"

```



It is possible to generate CNS ready configurations. This is a small example of Registry 

using Container Native Storage, a very common scenario:



```

---

deployment_type: "enterprise"

multi_master: true

haproxy_enabled: true

nfs_enabled: false

glusterfs_registry_enabled: true

registry_cns: true

registry_native_nfs: false

masters_list:

    - m1.example.com

    - m2.example.com

    - m3.example.com

etcd_list:

    - m1.example.com

    - m2.example.com

    - m3.example.com

lb_list:

    - lb.example.com

glusterfs_registry_map:

    n1.example.com: "glusterfs_devices='[ \"/dev/vdb\", \"/dev/vdc\", \"/dev/vdd\" ]'"

    n2.example.com: "glusterfs_devices='[ \"/dev/vdb\", \"/dev/vdc\", \"/dev/vdd\" ]'"

    n3.example.com: "glusterfs_devices='[ \"/dev/vdb\", \"/dev/vdc\", \"/dev/vdd\" ]'"

nodes_map:

    m1.example.com: ""

    m2.example.com: ""

    m3.example.com: ""

    n1.example.com: "openshift_node_labes=\"{'region': 'infra'}\""

    n2.example.com: "openshift_node_labes=\"{'region': 'infra'}\""

    n3.example.com: "openshift_node_labes=\"{'region': 'infra'}\""

    n4.example.com: "openshift_node_labes=\"{'region': 'primary'}\""

    n5.example.com: "openshift_node_labes=\"{'region': 'primary'}\""

```



## Defaults



Currently, the configured defaults are the following:



```

---

deployment_type: origin     # Choose between "origin" or "enterprise" deployment

ssh_user: root              # The user that connects to the nodes 

nfs_enabled: true           # Use native NFS for storage

glusterfs_enabled: false    # Enable Glusterfs based CNS or CRS

glusterfs_registry_enabled: false   # Use Glusterfs for registry storage

registry_native_nfs: true   # Use NFS for registry storage

registry_cns: false         # Enable Container Native Storage for registy

haproxy_enabled: false      # Enable HAProxy loadbalancer in HA deployments

install_version: v3.9       # Installation version

skip_checks: false          # Skip storage, memory, package checks

multi_master: false         # Multi master HA deployment

cluster_method: native      # Cluster method 

cluster_hostname: osapi.example.com # Private cluster endpoint name for API

cluster_public_hostname: osapi.example.com  # Public cluster endpoint name for API

containerized_deploy: false # Enable containerized deployment

containerized_ovs: false    # Containerize OpenvSwitch

containerized_node: false   # Containerize atomic-openshift-node service

containerized_master: false # Containerize master services (api and controllers)

containerized_etcd: false   # Containerize etcd service

system_images_registry: registry.access.redhat.com  # Default system images registry

openshift_use_crio: false   # Enable CRI-O deployment

openshift_crio_use_rpm: false   # Use rpm basesd installation for CRI-O

multi_infra: false          # Deploy multiple infra nodes

use_xip: false              # Use XIP for wildcard DNS (useful in training labs)

infra_ipv4: ""              # Infra IPv4 address for XIP based wildcard

ext_dns_wildcard: osapps.example.com    # Wildcard address (the common scenario)

sdn_plugin: ovs-multitenant # Cluster SDN plugin

disable_servicecatalog: true    # Disable the Service Catalog (from v3.7)

infra_replicas: 3           # Number of replicas for infra nodes

metrics_enabled: true       # Enable metrics subsystem

deploy_hosa: true           # Deploy Hawkular OpenShift Agent (HOSA)

metrics_native_nfs: true    # Use cluster NFS for metrics storage

prometheus_enabled: false   # Enable Prometheus deployment

prometheus_native_nfs: true # Use cluster NFS for Prometheus

logging_enabled: true       # Enable logging subsystem

logging_native_nfs: true    # Use Cluster NFS for logging

masters_list: []            # List of the master nodes

etcd_list: []               # List of the etcd nodes

lb_list: []                 # List of the load balancers

nfs_list: []                # List of cluster NFS nodes

glusterfs_map: {}           # Key/Value map of GlusterFS nodes

glusterfs_registry_map: {}  # Key/Value map of GlusterFS nodes for registry

nodes_map: {}               # Key/Value map of nodes.

```



## TODO



- Create new entries in the template for Container Native Storage (CNS)

- Create more sanity checks on string and int values

- Implement a good code testing