OpenShift Infrastructure安裝和使用指導

1 Installation

1.1 Preparation

Note: This article will take pre-production ofSpain as an environment to show the steps.

1.1.1 Infrastructure plan

According to the OpenShift specification, see https://docs.openshift.com/container-platform/3.7/install_config/install/prerequisites.html#install-config-install-prerequisites,each VM should be created with minimum flavor required by OpenShift. Ifcustomer wants to expand business scale, they could reference the same link toadd new node or scale up the node flavor.

The infrastructure will incorporate thefollowing VMs:

Load balancer: 1 VM, acts as NTP server, and HAProxyfor distributing traffic to masters.

Master: 3 VMs

Node: 2 VMs

Infra node: 1 VM, router POD, metric POD,logging POD will be created here.

DNS: 1 VM.

1.1.2 Domain name plan

LB:(NTP, LB)

    -lb.openshift.example.com 192.168.0.131

master:

    -master01.openshift.example.com 192.168.0.164

    -master02.openshift.example.com 192.168.0.159

    -master03.openshift.example.com 192.168.0.88

node:

    -node01.openshift.example.com 192.168.0.180

    -node02.openshift.example.com 192.168.0.135

infra:

    -infra-node01.openshift.example.com 192.168.0.195

1.1.3 Provision VMs

Based on the infrastructure plan, create eachVM with 8 vCPU, 16GB memory, 40GB system disk and 20GB data disk, and with OSof RHEL 7.3.

Note:currently we provision VMs using RHEL image, but in the future we will improveby creating another new image which contains some packages and configurationsbeforehand.

Meanwhile, bind LB VM and Infra node VM withelastic IP (EIP) respectively.

During the installation, all of the nodes needto access Internet to download Docker images, therefore you need to configureSNAT for the subnet. Steps:

Switch to NAT Gateway on VPC console, andcreate one by:

Here, you need to choose VPC and subnet namewhich are same with your infrastructure’s network definition.

After the NAT Gateway is created, you need to add SNAT rules, like:

This will allow all the VMs in subnet-docker tovisit Internet using SNAT. In this step, you have to apply an EIP in advance.Therefore, considering LB and Infra node bund with EIP, you need 3 EIPs intotal during the installation.

1.2 Configurations before installing

1.2.1 Change host name

On each VM, change host name by command ntmui.

If you provision the VMs with the exact name of domain name in plan, you dont have to change host name. If not, you have to change it.

1.2.2 Disable firewalld

systemctl disable firewalld

systemctl stop firewalld

Finally check if all the firewall ruleshave been removed:

iptables -S

Normally, it should be:

-P INPUT ACCEPT

-P FORWARD ACCEPT

-P OUTPUT ACCEPT

1.2.3 Allow root to access

On each VM, edit /etc/ssh/sshd_config to allow root to login using password, mainly options about:

PasswordAuthenticationyes

PermitRootLoginyes

 

Then restart sshd by:

systemctlrestart sshd

Check if you can login by root account with ssh.

1.2.4 Configure NTP

On LB VM, install NTP by:

yum install ntp

Edit /etc/ntp.conf and add thefollowing configuration to allow the other client in the same subnet tosynchronize time:

restrict 192.168.0.0 mask 255.255.255.0 nomodifynotrap

Enable NTP service and start it:

systemctl enable ntpd

systemctl restart ntpd

Check if NTP server synchronizes time withother servers in Internet: ntpstat

 

On the rest VMs, install NTP with the samecommand:

yum install ntp

But edit /etc/ntp.conf and set the upperlevel NTP server to be LB VM’s internal address:

# Use public servers from the pool.ntp.org project.

# Please consider joining the pool(http://www.pool.ntp.org/join.html).

server 192.168.0.131

And enable the NTP service and start it:

systemctl enable ntpd

systemctl restartntpd

1.2.5 Configure DNS

Create a VM with OS of Windows Server 2012 R2 DC.

Install DNS Server in the VM, with Add Roles and Features Wizard:

After DNS service installed, open the Properties for openshift-dns, and configure the Forwarders to be “100.125.1.11”, which is the default and internal DNS server when you create a new subnet.

Meanwhile add the DNS records in the DNS server, like:

Finally modify the DNS server address for the subnet:

Here, 192.168.0.168 is the internal address of the new DNS server, and subnet-0 is the subnet in which to deploy OpenShift infrastructure.

1.2.6 Install depended packages

On each VM:

yum -y install wget git net-tools bind-utilsiptables-services bridge-utils bash-completion kexec-tools sos psacct

yum -y update

yum -y install atomic-openshift-utils

yum -y install atomic-openshift-excluder atomic-openshift-docker-excluder

atomic-openshift-excluder unexclude

1.2.7 Install and configure Docker

On the nodes except for LB VM, install Docker by:

yum install docker-1.12.6

Check if it's installed correctly:

docker version

In Chapter 1.1.3, we introduced the VM being provisioned with one 20GB data disk. This disk is used to be Docker's backend storage. According to OpenShift's proposal, in production environment, it should be at least 100GB, and 300GB as suggested.

Edit /etc/sysconfig/docker-storage-setup,and set it to be:

CONTAINER_THINPOOL=docker-pool

DEVS=/dev/xvdb

VG=docker-vg

Note: /dev/xvdb is the data volume, and you can get your path by“fdisk -l”.

Configure it using command:

docker-storage-setup

Verify your configuration. You should have a dm.thinpooldev value in the /etc/sysconfig/docker-storage file and a docker-pool logical volume:

DOCKER_STORAGE_OPTIONS="--storage-driverdevicemapper --storage-opt dm.fs=xfs --storage-opt dm.thinpooldev=/dev/mapper/docker--vg-docker--pool--storage-opt dm.use_deferred_removal=true --storage-optdm.use_deferred_deletion=true "

 

Check if docker is running:

systemctl status docker

If it's not running, enable it and start it:

systemctl enable docker

systemctl start docker

If Docker is already running, re-initialize Docker:

systemctl stop docker

rm -rf /var/lib/docker/*

systemctl restart docker

1.2.8 Enable ssh trusty

During the installation, such as executing the command on the LB VM with root, the root user needs to access the other VMs without password.

On the LB VM, generate an ssh key, without any password:

ssh-keygen

Then distribute the ssh key to all the other VMs:

for host in lb.openshift.example.com \

           master01.openshift.example.com \

            master02.openshift.example.com \

            master03.openshift.example.com\

            node01.openshift.example.com\

            node02.openshift.example.com\

            infra-node01.openshift.example.com\

do ssh-copy-id $host; \

done

Check if root on LB VM could access other VM without password:

sshmaster01.openshift.example.com

1.2.9 Deploy NFS on LB VM

Starting with OpenShift Container Platform3.7, the OpenShift Ansible broker (OAB) is enabled by default. The OAB deploys its own etcd instance separate from the etcd used by the rest of the OpenShift Container Platform cluster. The OAB’s etcd instance requires separate storage using persistent volumes (PVs) to function. If no PV is available, etcd will wait until the PV can be satisfied. The OAB application will enter a CrashLoop state until its etcd instance is available.

Therefore you could deploy NFS service on LB VM and provide the PV by NFS path.

First check if NFS service related RPM package has been installed:

rpm -qa | grep nfs

If not installed, install NFS by:

yum install nfs-utils

Create a directory used for exporting to outside, such as /nfsexports.

Edit /etc/exports, and add thefollowing content:

/nfsexports *(rw,insecure,sync,all_squash)

Enable and start NFS service:

systemctl enable nfs-server

systemctl enable rpcbind

systemctl restart rpcbind

systemctl restart nfs-server

Check if the path is exported correctly. Execute the mount command on other VM, such as mater VM:

mount -t nfs lb.openshift.example.com:/nfsexports /tmp/test

Finally unmounts it after checks withsuccess:

umount /tmp/test

1.3 Install OpenShift Infrastructure

Reference Chapter 2.1 for ansible inventory file, and copy the content to /etc/ansible/hosts.

Install command:

ansible-playbook -i /etc/ansible/hosts /usr/share/ansible/openshift-ansible/playbooks/byo/config.yml

 

Uninstall command:

ansible-playbook -i /etc/ansible/hosts /usr/share/ansible/openshift-ansible/playbooks/adhoc/uninstall.yml

1.4 Verification

1.4.1 Login portal and provision PODs

oc get node

It should show the following output:

NAME                                 STATUS                     AGE       VERSION

infra-node01.openshift.example.com   Ready                      38m       v1.7.6+a08f5eeb62

master01.openshift.example.com       Ready,SchedulingDisabled   38m      v1.7.6+a08f5eeb62

master02.openshift.example.com       Ready,SchedulingDisabled   38m      v1.7.6+a08f5eeb62

master03.openshift.example.com       Ready,SchedulingDisabled   38m      v1.7.6+a08f5eeb62

node01.openshift.example.com         Ready                      38m      v1.7.6+a08f5eeb62

node02.openshift.example.com         Ready                      38m       v1.7.6+a08f5eeb62

Generate a new user account on each master:

htpasswd -b /etc/origin/master/htpasswd admin admin

Restart service on each master by:

systemctl restart atomic-openshift-master-api atomic-openshift-master-controllers

Add following domain name mappings in the hosts file of laptop:

195.235.204.174   lb.openshift.example.com       

195.235.204.175   *.cloudapps.openshift.example.com   #### should be the public IP of router infra node

Finally login console with URL: https://lb.openshift.example.com:8443/console, using credentials admin/admin.

Afterlogin with success, you need to create a project for yourself and then you can provision POD by selecting application images in the catalog.

1.4.2 Create persistent volume

Edit a PV file in any one of the master, withfollowing content:

apiVersion:v1

kind:PersistentVolume

metadata:

  name: pv0001

spec:

  capacity:

    storage: 5Gi

  accessModes:

  - ReadWriteOnce

  nfs:

    path: /share-7eaade1d

    server:sfs-nas1.eu-preprod.telefonicaopencloud.com

  persistentVolumeReclaimPolicy: Recycle

Create PV with command:

occreate -f ./nfs-pv.yaml

You can check PV created successfully or not:

oc get pv       

On the console, you can provision storage, andattach it to some pod.

Note: openshift willschedule by find a PV with capacity more than required.

1.4.3 Add nodes to the cluster

Create new VMs, and install the necessary packages, and configure them properly by referencing the above steps.

Edit /etc/ansible/hosts:

    [OSEv3:children]

    masters

    nodes

    etcd

    lb

    nfs

    ...

    ...

    [new_nodes]

    node03.openshift.example.com openshift_node_labels="{'region':'primary','zone':'default'}"

    node04.openshift.example.comopenshift_node_labels="{'region':'primary', 'zone':'default'}"

Add the nodes to cluster:

ansible-playbook-i /etc/ansible/hosts /usr/share/ansible/openshift-ansible/playbooks/byo/openshift-node/scaleup.yml

Finally, move any hosts you had defined in the[new_<host_type>] section into their appropriate section (but leave the[new_<host_type>] section definition itself in place) so that subsequent runs using this inventory file are aware of the nodes but do not handle them as new nodes. For example, when adding new nodes:

    [OSEv3:children]

    masters

    nodes

    etcd

    lb

    nfs

    ...

    ...

    [nodes]

    master01.openshift.example.com

    master02.openshift.example.com

    master03.openshift.example.com

    infra-node01.openshift.example.comopenshift_node_labels="{'region':'infra', 'zone':'default'}"

    infra-node02.openshift.example.comopenshift_node_labels="{'region':'infra', 'zone':'default'}"

    node01.openshift.example.comopenshift_node_labels="{'region':'primary', 'zone':'default'}"

    node02.openshift.example.comopenshift_node_labels="{'region':'primary', 'zone':'default'}"

    node03.openshift.example.comopenshift_node_labels="{'region':'primary', 'zone':'default'}"

    node04.openshift.example.comopenshift_node_labels="{'region':'primary', 'zone':'default'}"

    [new_nodes]

1.4.4 Dynamic persistent volume provisioning

In order to use storage class to provisionpersistent volume, we need first to configure OpenStack provider:

Reference: https://docs.openshift.com/container-platform/3.7/install_config/configuring_openstack.html

Run the following command to delete all thenodes:

oc delete node master.openshift.example.com

On master node, edit /etc/origin/master/master-config.yaml, and add cloud-provider andcloud-config in apiServerArguments and controllerArguments sections:

kubernetesMasterConfig:

  ...

 apiServerArguments:

   cloud-provider:

      -"openstack"

    cloud-config:

      -"/etc/cloud.conf"

 controllerArguments:

   cloud-provider:

      -"openstack"

   cloud-config:

      -"/etc/cloud.conf"

On all the node nodes, edit /etc/origin/node/node-config.yamland add cloud-provider and cloud-config in kubeletArguments section:

nodeName:

 OpenShift-node01/02

kubeletArguments:

 cloud-provider:

    -"openstack"

  cloud-config:

-"/etc/cloud.conf"

Meanwhile on all the nodes, edit /etc/cloud.conf:

[Global]

auth-url =https://iam.eu-preprod.telefonicaopencloud.com:443/v3

username = haobinbin

password = "******"

domain-id = e80f19ea6bf4442381f01995ad5c7b9a

tenant-id = 64105d651c3948cbb7eb25e589821d49

region = eu-preprod-1

 

[LoadBalancer]

subnet-id = fb99d7d7-976d-4631-9935-4f7e12ae86be

 

[BlockStorage]

bs-version = v2

On all the master nodes, restart serviceby:

systemctl restart atomic-openshift-master-apiatomic-openshift-master-controllers

On all the node nodes, restart service by:

systemctl restart atomic-openshift-node

Notes: because certificate of pre-productionis self-signed, we need to import the ca certificate into all the nodes (seethe following commands). This operation should be performed in advance,otherwise the node service will not be able to restart.

update-ca-trust enable

cp ca.crt /etc/pki/ca-trust/source/anchors/

update-ca-trust extract

openssl verify ca.crt

rapidSSL-ca.crt: OK  #### screen output expected

 

Notes: after all the nodes areup, maybe the master is schedule enabled state. We need to change them to bedisabled.

oadm manage-node master01.openshift.example.com--schedulable=false

 

Edit storage class yaml file cinder-storageclass.yaml:

kind: StorageClass

apiVersion: storage.k8s.io/v1

metadata:

  name: gold

provisioner: kubernetes.io/cinder

parameters:

  type:SATA 

  availability:eu-preprod-1a

  fsType: ext4

Create storage class by:

oc create -f ./cinder-storageclass.yaml

Check if it is created successfully:

oc get storageclass

Configure the above storage class as thedefault one:

oc patch storageclass gold -p '{"metadata":{"annotations": {"storageclass.kubernetes.io/is-default-class":"true"}}}'

 

Edit PVC yaml file pvc-gold.yaml:

apiVersion: v1

kind: PersistentVolumeClaim

metadata:

 name:cinder-vol-binbin01

spec:

 accessModes:

  -ReadWriteOnce

 resources:

   requests:

     storage:10Gi

 storageClassName: gold

 

Create PVC by:

oc create -f pvc-gold.yaml

Check if PVC is created successfully:

oc get pvc

Login OpenCloud console and check if volume iscreated:

1.4.5 (Optional) PVC provision by SFS storage

Dynamic PVC creation is only supported inKVM scenario instead of Xen. If there’s no KVM POD, end user could choose NFSstorage as PV. We take SFS service as an example to explain.

Create a new yaml file with namesfs-pv01.yaml on OpenShift Master node, and edit the content:

apiVersion: v1

kind: PersistentVolume

metadata:

  name:sfs-pv0001

spec:

  capacity:

    storage:10Gi

  accessModes:

  -ReadWriteOnce

  nfs:

    path:/share-f7d41805

    server:sfs-nas1.eu-preprod.telefonicaopencloud.com

 persistentVolumeReclaimPolicy: Recycle

Notes: the path /share-f7d41805is SFS path applied in OpenCloud, and storage is its size.

Create PV in this Master node:

oc create -f sfs-pv01.yaml

Checkif PV is created: oc get pv

Initially its status is Available.

Thenend user could create persistent volume claim on the OpenShift portal, andfinally OpenShift will schedule the request on this PV, and this PV status ischanged to Bound.

1.4.6 (Optional) Use PVC for docker registry

Reference: https://docs.openshift.com/container-platform/3.7/install_config/registry/deploy_registry_existing_clusters.html#install-config-deploy-registry-existing-clusters

 

As said in the guide, a deploymentconfigs,service, and pod named with docker-registry will be created during theOpenShift infrastructure deployment. On default, the docker registry will usetemporary directory for storage of docker images. After the POD is re-created,all the images will be lost.

In order to keep the images permanently,we’d better replace the deploymentconfigs.

 

First we need to create PV and PVC. TakeSFS as an example, and create a SFS path in advance.

Edit sfs-pv-docker-registry.yaml:

apiVersion: v1

kind: PersistentVolume

metadata:

  name:sfs-pv-docker-registry

spec:

  capacity:

    storage:10Gi

  accessModes:

  -ReadWriteMany

  nfs:

    path:/share-152b6f67

    server:sfs-nas1.eu-preprod.telefonicaopencloud.com

 persistentVolumeReclaimPolicy: Recycle

Create PV by: oc create -fsfs-pv-docker-registry.yaml

Edit sfs-pvc-docker-registry.yaml:

apiVersion: v1

kind: PersistentVolumeClaim

metadata:

  name:sfs-pvc-docker-registry

spec:

  accessModes:

    -ReadWriteMany

  resources:

    requests:

      storage:10Gi

Create PVC by: oc create -fsfs-pvc-docker-registry.yaml

 

Get the current volumes indeploymentconfigs of docker-registry:

oc volumes dc/docker-registry

deploymentconfigs/docker-registry

  empty directory as registry-storage

    mounted at /registry

  secret/registry-certificates asregistry-certificates

    mounted at /etc/secrets

Replace the current volume ofregistry-storage by:

oc volume dc/docker-registry --add--name=registry-storage -t pvc --claim-name=sfs-pvc-docker-registry--mount-path=/registry --overwrite

Finally the POD docker-registry will bere-created, and the deploymentconfigs of docker-registry will also be updated:

deploymentconfigs/docker-registry

 secret/registry-certificates as registry-certificates

    mountedat /etc/secrets

 pvc/sfs-pvc-docker-registry (allocated 10GiB) as registry-storage

   mounted at /registry

 

In order for test, we can access the PODand check the path:

ocrsh docker-registry-4-zrxpb

1.4.7 (Optional) docker registry

oc login -u haobinbin -p ******

oc whoami -t

docker login -u haobinbin -p yFX6tpyff8kkldVGWcuTLYNkDiMbr1_7-3Q-vaf-sIcdocker-registry-default.cloudapps.openshift.example.com:443

docker pull docker.io/nginx:latest

docker tag docker.io/nginx:latestdocker-registry-default.cloudapps.openshift.example.com:443/opencloud2/nginx:latest

docker pushdocker-registry-default.cloudapps.openshift.example.com:443/opencloud2/nginx:latest

ocget is

1.4.8 (Optional) Manage users

1.4.8.1 AllowAllPasswordIdentityProvider

 identityProviders:

  - challenge:true

    login: true

    mappingMethod:claim

    name:my_allow_provider

    provider:

     apiVersion: v1

      kind:AllowAllPasswordIdentityProvider

 

Configure the above part in all the master’sconfiguration file, and restart service on each master by:

systemctl restart atomic-openshift-master-apiatomic-openshift-master-controllers

Any user and password could login theOpenShift portal, even if the user account does not exists.

1.4.8.2 DenyAllPasswordIdentityProvider

  identityProviders:

  - name:my_deny_provider

    challenge:true

    login: true

   mappingMethod: claim

    provider:

     apiVersion: v1

      kind:DenyAllPasswordIdentityProvider

Configure the above part and restartservice by:

systemctl restart atomic-openshift-master-apiatomic-openshift-master-controllers

Notes:this identity provider does not make sense in production environment.

1.4.8.3 KeystonePasswordIdentityProvider

 identityProviders:

  - name:my_keystone_provider

    challenge:true

    login: true

   mappingMethod: claim

    provider:

      apiVersion:v1

      kind:KeystonePasswordIdentityProvider

     domainName: zhangchi

      url:https://iam.eu-preprod.telefonicaopencloud.com:443/v3

      ca:/root/ca.crt

Configure the above part and restartservice on each master by:

systemctl restart atomic-openshift-master-apiatomic-openshift-master-controllers

Notes: the OpenShift will call OpenCloudIAM API to validate username and password under domain zhangchi.

1.4.8.4 HTPasswdPasswordIdentityProvider

 identityProviders:

  - challenge:true

    login: true

    mappingMethod: claim

    name:htpasswd_auth

    provider:

     apiVersion: v1

      file:/etc/origin/master/htpasswd

      kind:HTPasswdPasswordIdentityProvider

Configure the above part.

Meanwhile on each master, using thefollowing command to generate account and password:

htpasswd -b /etc/origin/master/htpasswd admin admin

Restart master service on each master by:

systemctl restart atomic-openshift-master-apiatomic-openshift-master-controllers

Notes: if htpassword cannot be run, try to install it by yum installhttpd-tools.

1.4.8.5 LDAPPasswordIdentityProvider

Reference: https://docs.openshift.com/container-platform/3.7/install_config/configuring_authentication.html#LookupMappingMethod

2 Appendix: ansible host files

2.1 multi_master_in_HA_without_persistent_storage

[OSEv3:children]

masters

nodes

etcd

lb

nfs

 

 

[OSEv3:vars]

ansible_ssh_user=root

openshift_deployment_type=openshift-enterprise

openshift_master_identity_providers=[{'name':'htpasswd_auth','login':'true', 'challenge':'true', 'kind':'HTPasswdPasswordIdentityProvider','filename':'/etc/origin/master/htpasswd'}]

openshift_master_default_subdomain=cloudapps.openshift.example.com

 

openshift_logging_install_logging=true

openshift_metrics_install_metrics=true

 

openshift_disable_check=disk_availability,memory_availability

openshift_clock_enabled=true

 

#### multiple master, need to add thefollowing parameters

openshift_master_cluster_method=native

openshift_master_cluster_hostname=lb.openshift.example.com

openshift_master_cluster_public_hostname=lb.openshift.example.com

 

openshift_template_service_broker_namespaces=['openshift','myproject']

 

openshift_hosted_etcd_storage_kind=nfs

openshift_hosted_etcd_storage_nfs_options="*(rw,root_squash,sync,no_wdelay)"

openshift_hosted_etcd_storage_nfs_directory=/nfsexports

openshift_hosted_etcd_storage_volume_name=

openshift_hosted_etcd_storage_access_modes=["ReadWriteOnce"]

openshift_hosted_etcd_storage_volume_size=20G

openshift_hosted_etcd_storage_labels={'storage':'etcd'}

 

ansible_service_broker_local_registry_whitelist=['.*-apb$']

 

[masters]

master01.openshift.example.com

master02.openshift.example.com

master03.openshift.example.com

 

#### multiple master, need to add thefollowing etcd and lb hosts

[etcd]

master01.openshift.example.com

master02.openshift.example.com

master03.openshift.example.com

 

[lb]

lb.openshift.example.com

 

[nodes]

master01.openshift.example.com

master02.openshift.example.com

master03.openshift.example.com

infra-node01.openshift.example.comopenshift_node_labels="{'region':'infra', 'zone':'default'}"

node01.openshift.example.comopenshift_node_labels="{'region':'primary', 'zone':'default'}"

node02.openshift.example.comopenshift_node_labels="{'region':'primary', 'zone':'default'}"

 

[nfs]

lb.openshift.example.com

2.2 allinone_with_persistent_storage

[OSEv3:children]

masters

nodes

etcd

nfs

 

[OSEv3:vars]

ansible_ssh_user=root

debug_level=4

openshift_deployment_type=openshift-enterprise

openshift_master_identity_providers=[{'name':'htpasswd_auth','login':'true', 'challenge':'true', 'kind':'HTPasswdPasswordIdentityProvider','filename':'/etc/origin/master/htpasswd'}]

openshift_master_default_subdomain=cloudapps.openshift.example.com

 

####install cluster logging, with persistent volume

openshift_logging_install_logging=true

openshift_logging_storage_kind=nfs

openshift_logging_storage_access_modes=['ReadWriteOnce']

openshift_logging_storage_host=sfs-nas1.eu-preprod.telefonicaopencloud.com

openshift_logging_storage_nfs_directory=/share-84a7b897

openshift_logging_storage_volume_name=

openshift_logging_storage_volume_size=20Gi

 

####install cluster metrics, with persistent volume

openshift_metrics_install_metrics=true

openshift_metrics_storage_kind=nfs

openshift_metrics_storage_access_modes=['ReadWriteOnce']

openshift_metrics_storage_host=sfs-nas1.eu-preprod.telefonicaopencloud.com

openshift_metrics_storage_nfs_directory=/share-7d9c5e07

openshift_metrics_storage_volume_name=

openshift_metrics_storage_volume_size=20Gi

 

openshift_disable_check=disk_availability,memory_availability

openshift_clock_enabled=true

 

openshift_template_service_broker_namespaces=['openshift','myproject']

 

openshift_hosted_etcd_storage_kind=nfs

openshift_hosted_etcd_storage_nfs_options="*(rw,root_squash,sync,no_wdelay)"

openshift_hosted_etcd_storage_nfs_directory=/nfsexports

openshift_hosted_etcd_storage_volume_name=

openshift_hosted_etcd_storage_access_modes=["ReadWriteOnce"]

openshift_hosted_etcd_storage_volume_size=20G

openshift_hosted_etcd_storage_labels={'storage':'etcd'}

 

ansible_service_broker_local_registry_whitelist=['.*-apb$']

 

[masters]

master.openshift.example.comopenshift_schedulable=true

 

[nodes]

master.openshift.example.comopenshift_node_labels="{'region':'infra'}"openshift_node_kubelet_args="{'image-gc-high-threshold':['90'],'image-gc-low-threshold': ['80']}" openshift_schedulable=true

 

[etcd]

master.openshift.example.com

 

[nfs]

master.openshift.example.com

2.3 allinone_without_persistent_storage

[OSEv3:children]

masters

nodes

etcd

nfs

 

[OSEv3:vars]

ansible_ssh_user=root

debug_level=4

openshift_deployment_type=openshift-enterprise

openshift_master_identity_providers=[{'name':'htpasswd_auth','login':'true', 'challenge':'true', 'kind':'HTPasswdPasswordIdentityProvider','filename':'/etc/origin/master/htpasswd'}]

openshift_master_default_subdomain=cloudapps.openshift.example.com

 

####install cluster logging, without persistent volume

openshift_logging_install_logging=true

 

openshift_disable_check=disk_availability,memory_availability

openshift_clock_enabled=true

 

openshift_template_service_broker_namespaces=['openshift','myproject']

 

openshift_hosted_etcd_storage_kind=nfs

openshift_hosted_etcd_storage_nfs_options="*(rw,root_squash,sync,no_wdelay)"

openshift_hosted_etcd_storage_nfs_directory=/nfsexports

openshift_hosted_etcd_storage_volume_name=

openshift_hosted_etcd_storage_access_modes=["ReadWriteOnce"]

openshift_hosted_etcd_storage_volume_size=20G

openshift_hosted_etcd_storage_labels={'storage':'etcd'}

 

ansible_service_broker_local_registry_whitelist=['.*-apb$']

 

[masters]

master.openshift.example.comopenshift_schedulable=true

 

[nodes]

master.openshift.example.comopenshift_node_labels="{'region':'infra'}"openshift_node_kubelet_args="{'image-gc-high-threshold':['90'],'image-gc-low-threshold': ['80']}" openshift_schedulable=true

 

[etcd]

master.openshift.example.com

 

[nfs]

master.openshift.example.com