Dataset Viewer
text
stringlengths 26
1.5k
|
|---|
SAS® Viya® Platform:
Deployment Guide
2023.08**This document might apply to additional versions of the software. Open this document in SAS Help Center and click
on the version in the banner to see all available versions.
SAS® Documentation
September 5, 2023
|
The correct bibliographic citation for this manual is as follows: SAS Institute Inc. 2023. SAS® Viya® Platform: Deployment Guide . Cary, NC:
SAS Institute Inc.
SAS® Viya® Platform: Deployment Guide
Copyright © 2023, SAS Institute Inc., Cary, NC, USA
All Rights Reserved. Produced in the United States of America.
For a hard copy book: No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any
means, electronic, mechanical, photocopying, or otherwise, without the prior written permission of the publisher, SAS Institute Inc.
For a web download or e-book: Your use of this publication shall be governed by the terms established by the vendor at the time you
acquire this publication.
The scanning, uploading, and distribution of this book via the Internet or any other means without the permission of the publisher is illegal and
punishable by law. Please purchase only authorized electronic editions and do not participate in or encourage electronic piracy of copyrighted
materials. Your support of others' rights is appreciated.
U.S. Government License Rights; Restricted Rights: The Software and its documentation is commercial computer software developed at
private expense and is provided with RESTRICTED RIGHTS to the United States Government. Use, duplication, or disclosure of the Software
by the United States Government is subject to the license terms of this Agreement pursuant to, as applicable, FAR 12.212, DFAR
|
by the United States Government is subject to the license terms of this Agreement pursuant to, as applicable, FAR 12.212, DFAR
227.7202-1(a), DFAR 227.7202-3(a), and DFAR 227.7202-4, and, to the extent required under U.S. federal law, the minimum restricted rights
as set out in FAR 52.227-19 (DEC 2007). If FAR 52.227-19 is applicable, this provision serves as notice under clause (c) thereof and no other
notice is required to be affixed to the Software or documentation. The Government’s rights in Software and documentation shall be only those
set forth in this Agreement.
SAS Institute Inc., SAS Campus Drive, Cary, NC 27513-2414
September 2023
SAS® and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and
other countries. ® indicates USA registration.
Other brand and product names are trademarks of their respective companies.
v_041-P2:dplyml0phy0dkr
|
Contents
Chapter 1 /Pre-installation Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
ingress-nginx Vulnerability Mitigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Confidential Computing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Plan the Workload Placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Deploy the SAS Viya Platform Deployment Operator . . . . . . . . . . . . . . . . . . . . . . . . . 10
Retrieve Required Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Preparing for OpenShift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Directory Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Chapter 2 /Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Initial kustomization.yaml File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Common Customizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
|
Common Customizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Install the Orchestration Tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Create the SASDeployment Custom Resource . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Deploy the Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Readiness Service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Sign In as the sasboot User . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
Promotion and Migration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Chapter 3 /Post-Installation Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Configure Identities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
Configure Multi-tenancy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Configure the Connection to the Mail Service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
|
Configure the Connection to the Mail Service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
Configure Files Service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Configure Monitoring and Logging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Configure Guest Access . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Obtain and Run Hadoop Tracer Script . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
Configure Cloud Analytic Services (CAS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Configure Cloud Data Exchange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Configure Model Access . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
Configure SAS Asset and Liability Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Configure SAS Business Orchestration Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Configure SAS Data Engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
Configure SAS Data Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
|
Configure SAS Data Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
Configure SAS Dynamic Actuarial Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Deploy SAS Enterprise Session Monitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Configure SAS Event Stream Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Configure SAS Expected Credit Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
Configure SAS for Microsoft 365 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
Configure SAS Intelligent Decisioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
Configure SAS Model Manager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
Configure SAS Model Risk Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Configure SAS Risk Engine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Configure SAS Risk Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Configure SAS Stress Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
|
Configure SAS Stress Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Configure SAS Visual Analytics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Configure SAS Visual Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Configure SAS Viya . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
|
Configure SAS Viya Advanced . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Configure SAS Viya Enterprise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
Configure SAS Viya Programming . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
Configure SAS Viya with SingleStore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Configure SAS Workload Orchestrator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Configure SAS/CONNECT Spawner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Chapter 4 /Validating the Deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
The SAS Viya Platform: Deployment Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
The SAS Viya Platform and the SAS Operational Quality Tool . . . . . . . . . . . . . . . . 123
Chapter 5 /Modifying the Deployment Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
Modify Existing Customizations in a Deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
Chapter 6 /Uninstalling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
|
Uninstall with the SAS Viya Platform Deployment Operator . . . . . . . . . . . . . . . . . . 127
Remove the SAS Viya Platform Deployment Operator Only . . . . . . . . . . . . . . . . . . 133
Uninstall Deployments That Do Not Use the Deployment Operator . . . . . . . . . . . . 133
Appendix 1 /(Optional) Using a Mirror Registry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Create a Mirror Registry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Create and Populate a Mirror Registry in Microsoft Azure . . . . . . . . . . . . . . . . . . . . 147
Create and Populate a Mirror Registry in Amazon ECR . . . . . . . . . . . . . . . . . . . . . . 149
Create and Populate a Mirror Registry in Google Cloud Platform . . . . . . . . . . . . . . 150
Create and Populate a Mirror Registry in Red Hat OpenShift . . . . . . . . . . . . . . . . . 152
Create and Populate a Mirror Registry in JFrog Artifactory . . . . . . . . . . . . . . . . . . . 153
Create a Mirror Registry at a Dark Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
Appendix 2 /SAS Viya Deployment Operator Fields and Messages . . . . . . . . . . . . . . . . . . . . . . . . . 155
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
|
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Fields in the SASDeployment Custom Resource . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
Manage Updates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Environment Variables for the Operator Pod . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
Communications from the Operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
Appendix 3 /Change Deployment Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
Kubernetes Commands to SAS Viya Platform Deployment Operator . . . . . . . . . . 167
Kubernetes Commands to sas-orchestration Command . . . . . . . . . . . . . . . . . . . . . 169
SAS Viya Platform Deployment Operator to Any Other Method . . . . . . . . . . . . . . . 169
Appendix 4 /Troubleshooting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Deployment Operator Custom Resource Is Too Large . . . . . . . . . . . . . . . . . . . . . . . 171
Error Checking Database Connection Error Message . . . . . . . . . . . . . . . . . . . . . . . 172
Certframe Container Fails at init Stage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
|
Certframe Container Fails at init Stage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
Issues with External PostgreSQL Server Size and Connections . . . . . . . . . . . . . . . 173
SAS Viya Platform Deployment Operator Reconcile Failures . . . . . . . . . . . . . . . . . 174
Performing Updates Produces Many Error Messages . . . . . . . . . . . . . . . . . . . . . . . 174
Failure to Create Symbolic Link . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Long Delay for SAS/CONNECT Server Sign-on from SAS Studio . . . . . . . . . . . . . 175
"Operation not permitted" Error in SAS Configurator for Open Source Logs . . . . 176
Uninstall Leaves the Deployment Namespace in a Terminating State . . . . . . . . . . 176ivContents
|
1
Pre-installation Tasks
ingress-nginx Vulnerability Mitigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Confidential Computing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Plan the Workload Placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Assign Nodes by Class . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Deploy the SAS Viya Platform Deployment Operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Overview of the SAS Viya Platform Deployment Operator . . . . . . . . . . . . . . . . . . . . . . . 10
Retrieve the Files Required by the SAS Viya Platform Deployment Operator . . . . . . 11
Edit the transformer.yaml File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Edit the kustomization.yaml for Cluster-Wide Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Edit the kustomization.yaml for seccomp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
|
Edit the kustomization.yaml for seccomp . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Configure the Mirror Registry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Add an imagePullSecret . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Configure Proxy Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Apply the SAS Viya Platform Deployment Operator Resources to the Cluster . . . . . 18
Retrieve Required Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Preparing for OpenShift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Workload Node Placement Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Security Context Constraints and Service Accounts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Enable hostPath Mounts for CAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Networking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
|
Networking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
TLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Additional Security . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Directory Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
$deploy/sas-bases Directory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
$deploy/site-config Directory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Directories for the SAS Viya Platform Deployment Operator . . . . . . . . . . . . . . . . . . . . . 301
|
ingress-nginx Vulnerability Mitigation
The SAS Viya platform requires configuration of the ingress-nginx controller. By
default, ingress-nginx supports custom snippets, which the SAS Viya platform
requires. ingress-nginx has discovered that the default setting of "allow-snippets=true
" exposes the deployment to known vulnerability CVE-2021-25742.
A user with Kubernetes namespace administrator privileges could use custom
snippets to exfiltrate the ingress-nginx service account token and gain access to all
secrets in the cluster.
Note: For more information about the CVE, see CVE-2021-25742 Detail .
The SAS Viya platform supports a mitigation for CVE-2021-25742, which involves
the application of a block list. SAS recommends that you implement it as soon as
possible by following the steps described in this section.
Note: Enabling this mitigation affects all applications in the Kubernetes cluster that
are using the ingress-nginx controller.
1Ensure that your SAS Viya platform deployment and the cluster it is being
deployed in meets all of the following criteria before applying this mitigation:
nA supported version of the SAS Viya platform must either be deployed, or
you must be about to deploy a supported version.
nThe Kubernetes cluster does not have unsupported deployments of the SAS
Viya platform that lack the mitigation and that are sharing the ingress-nginx
controller.
nThe ingress-nginx controller is a supported version.
|
Viya platform that lack the mitigation and that are sharing the ingress-nginx
controller.
nThe ingress-nginx controller is a supported version.
nThe Kubernetes cluster does not have any additional applications using the
ingress-nginx controller that do not support using the "annotation-value-word-
blocklist" annotation.
If the cluster and your SAS Viya platform deployment do not meet these criteria,
do not proceed with applying the mitigation.
2Add the "annotation-value-word-blocklist" annotation to the ConfigMap for the
ingress-nginx controller in your Kubernetes cluster. Here is an example
command: kubectl patch cm ingress-nginx-controller -n ingress-nginx -p '{"data":{"allow-snippet-annotations":"true", "annotation-value-word-blocklist":"load_module,lua_package,_by_lua,location,root,proxy_pass,serviceaccount,{,},\\"}}';
In this example, the ConfigMap has the default name of ingress-nginx-controller
and is deployed in the ingress-nginx namespace. Modify this command as
required for your deployment.2Chapter 1 / Pre-installation Tasks
|
The suggested values for the blocklist are supported by the SAS Viya platform: "load_module,lua_package,_by_lua,location,root,proxy_pass,serviceaccount,{,},\"
Confidential Computing
Note: Confidential computing is only supported on Microsoft Azure at this time. If
you are not deploying on Microsoft Azure, skip this section.
Confidential computing helps you to protect your data and code from the time it is
created until it is destroyed. To support confidential computing SAS uses AMD SEV
technology, which supports a new set of security features that protects data-in-use
with full VM memory encryption.
To enable confidential computing on Microsoft Azure:
1Ensure that you have selected VM instance types that support confidential
computing with the SAS Viya platform. For more information, see “Requirements
for Confidential Computing” in System Requirements for the SAS Viya Platform .
2Ensure that the default (system) node pool and the user node pools contain VM
instance types that support confidential computing. For details, see Create
confidential VM on AMD in the Azure portal .
3If you are using the SAS Viya platform Infrastructure as Code (IaC) project to
provision your deployment, make the following changes to the terraform.tfvars
file in the IaC script:
aUpdate the Default Node pool VM type to one of the AMD SEV confidential
VMs. Here is an example:default_nodepool_vm_type = "Standard_DC8as_v5"
|
aUpdate the Default Node pool VM type to one of the AMD SEV confidential
VMs. Here is an example:default_nodepool_vm_type = "Standard_DC8as_v5"
bUpdate the node pool availability zone and location. Here is an example:default_nodepool_availability_zones = ["1", "2", "3"] location = "East US"
If you are unsure of the availability zone, set the value to empty ""
.
cUpdate the storage type. Using storage_type=standard
creates an NFS
server VM. Using storage_type=ha
creates Azure NetApp files.
Note: If you are deploying using the viya4-deployment GitHub project, do not
attempt to configure confidential computing for the NFS server or Jump server.
This configuration is not supported. Use NetApp files for secure storage instead.
CAUTION
After you have deployed your software, if you restart your Microsoft Azure
cluster, the default netapp_protocols=["NFSv3"] might cause file lock
issues for SAS Configuration Server (Consul), Redis, and the logging/Confidential Computing 3
|
monitoring servers. These issues can result in a non-functional
deployment.
Plan the Workload Placement
Note: If you have used a GitHub project, SAS Viya Infrastructure as Code (IaC), for
one of the cloud providers to provision your infrastructure components, the
procedures described in this section have already been performed for you. You
might still want to verify the taints and labels with the following command:kubectl get nodes -o=custom-columns=NAME:.metadata.name,LABELS:.metadata.labels,TAINTS:.spec.taints
For more information about commands to validate the workload node placement in
your infrastructure, see “Place the Workload on Nodes” on page 9. For more
information about IaC, see “Help with Cluster Setup” in System Requirements for
the SAS Viya Platform .
Introduction
Before you deploy your SAS Viya platform, you should develop a plan to distribute
the workload across your deployment. Labeling and tainting the nodes prepares the
environment to use your plan when the manifest is applied during the deployment
process. This section explains how to distribute the workload across your
deployment according to several factors.
Assign Nodes by Class
Overview
The SAS Viya platform consists of multiple workload classes. Each class of
workload has a unique set of attributes that you must consider when planning your
deployment. When planning the placement of the workload, it is helpful to think
|
deployment. When planning the placement of the workload, it is helpful to think
beyond the initial deployment and to also consider the Kubernetes maintenance life
cycle.
IMPORTANT SAS strongly recommends labeling and tainting all your
nodes, especially the CAS nodes. Note, however, that the connect workload
class is only required if you are not using dynamically launched pods. For 4Chapter 1 / Pre-installation Tasks
|
more information about dynamically launched pods, see “connect” on page
6.
Workload Classes
Properties associated with the pods in your deployment describe the type of work
that they perform. By using the commands described in “Place the Workload on
Nodes” on page 9, you can configure where you want each class of pods to run,
thereby enabling you to manage the associated workload.
Here are the SAS Viya platform workload classes:
stateless
These workloads include web applications and microservices. Stateless
applications scale horizontally in order to handle increased workload and to
provide higher availability. By default, one replica of each stateless workload is
deployed. A high-availability (HA) scaling overlay is provided with the
deployment artifacts for each SAS Viya platform software order.
stateful
These workloads include open-source components that are used by the SAS
Viya platform in order to store critical information that is required to maintain
operational integrity. These workloads are deployed in an HA configuration by
default.
cas
SAS Cloud Analytic Services (CAS) is the primary compute engine in the SAS
Viya platform. Although CAS is a stateful process, it has unique attributes and
requirements within the Kubernetes infrastructure. CAS has been engineered to
use all available resources to complete each request in the shortest amount of
time. Therefore, CAS can quickly drive every available CPU core to 100%
|
time. Therefore, CAS can quickly drive every available CPU core to 100%
utilization. CAS also aggressively consumes RAM to process large volumes of
data as quickly as possible.
Because CAS was designed with performance as a primary objective, it must be
hosted on dedicated resources. Therefore, each CAS pod must be placed on its
own dedicated node. In addition, if you are not running with auto-resourcing
(which is enabled by default), SAS recommends that you configure the CAS
server with guaranteed quality of service (QoS) in order to prevent CAS pods
from being evicted by Kubernetes.
Note: The requirement to place each CAS pod on its own dedicated node does
not apply to the personal CAS server. For more information about the personal
CAS server, see “Create a Personal CAS Server” on page 64.
For increased granularity, you can also use the cascontroller and casworker
classes. The same restrictions described for the cas class also apply to
cascontroller and casworker.
compute
Compute processes are stateful processes that run to completion (that is, they
are not long-running services). They deliver multiple functions within a SAS Viya
platform deployment, including but not limited to batch jobs, interactive SAS Plan the Workload Placement 5
|
Studio sessions, and parallel compute processes that are executed as part of a
SAS Model Studio workflow.
connect
SAS/CONNECT provides the ability for clients to launch SAS compute
processes, the same processes described in “compute” on page 5, within a local
or remote SAS environment. The SAS/CONNECT Spawner is a SAS Viya
platform service that launches these processes on behalf of SAS/CONNECT
clients. The processes can be launched in their own pods (referred to as
“dynamically launched pods”) or in the SAS/CONNECT Spawner pod. SAS 9
clients from releases before SAS 9.4M7 and SAS Viya 3 clients from versions
before SAS Viya 3.5 do not support dynamically launched pods, so by default
their processes are launched in the SAS/CONNECT Spawner pod. For other
clients, dynamically launched pods are enabled.
IMPORTANT If you are using dynamically launched pods, you do not
need to create a connect workload class and should therefore skip the
rest of this description of the connect workload class.
If you are using external SAS/CONNECT clients with your SAS Viya platform
deployment, the SAS/CONNECT Spawner requires more resources in order to
avoid being evicted by Kubernetes. Like CAS, the SAS/CONNECT Spawner pod
requires a dedicated Kubernetes node, and SAS recommends that it be
configured with Kubernetes guaranteed QoS.
If you do use the connect workload class, you must add a transformer to the
|
configured with Kubernetes guaranteed QoS.
If you do use the connect workload class, you must add a transformer to the
base kustomization.yam file. For more details, see the README file at $deploy/sas-bases/examples/sas-connect-spawner/README.md
(for Markdown format)
or at $deploy/sas-bases/docs/configure_sasconnect_spawner_in_the_sas_viya_platform.htm
(for HTML
format).
Default Node Pool Configuration
SAS recommends that you use a node pool that does not run SAS workloads in the
cluster. This node pool is referred to as the default node pool . The default node pool
is useful for ingress controllers and for hosting LDAP and monitoring stacks. SAS
also recommends that the default node pool be labeled with kubernetes.azure.com/mode=system
so that SAS pods are not scheduled there.
This label is automatically applied to the default Microsoft Azure Kubernetes Service
system node pool. If you are provisioning the cluster with the SAS-provided IaC
tooling, the label is automatically applied to the equivalent default node pool for
other infrastructures. If you are deploying on an infrastructure other than Microsoft
Azure Kubernetes Service and you are not provisioning using IaC, you must apply
the label yourself with the following command:kubectl label nodes node-name kubernetes.azure.com/mode=system
Note: The node-name can be the name of a single node or it can be a list of node
names separated by a space. The term azure
is applied to all of the supported
|
names separated by a space. The term azure
is applied to all of the supported
cloud platforms because it represents a default SAS Viya platform deployment
setting that is not vendor-specific.6Chapter 1 / Pre-installation Tasks
|
Considerations and Default Settings for
Workload Placement
The Kubernetes cluster administrator will recognize that each of the SAS Viya
platform workload classes should be managed differently. In situations where the
deployment has rigorous availability requirements and the Kubernetes cluster
software is being upgraded, the different requirements might resemble the following:
nNodes that host the stateless and stateful workload class can be individually
drained and updated with no additional management concerns as long as the
HA transformer was used for the deployment.
nNodes that host the compute workload class should be cordoned. The workloads
that run on those nodes should be allowed to run to completion in order to avoid
interrupting work in progress. Examples are a SAS Studio session, a batch job,
or a SAS Model Studio session. When all compute processes have completed,
the node can be drained so that maintenance tasks can be performed.
Each node that hosts compute class workloads can be cordoned, drained, or
updated in a manner that does not disrupt end-user activity as long as each
compute session is allowed to run to completion.
Note: SAS Workload Orchestrator requires at least one node labeled for the
compute workload class. Therefore, unless you intend to disable SAS Workload
Orchestrator, you must have at least one node labeled for the compute workload
class.
nNodes that host workloads for the cas and connect classes must be terminated
|
class.
nNodes that host workloads for the cas and connect classes must be terminated
and started on different nodes before the underlying nodes can be maintained as
required.
Some aspects of workload placement are set by default during the SAS Viya
platform deployment, including both nodeAffinity and tolerations. In addition, SAS
Viya platform workloads are not scheduled on the system node pool in Microsoft
Azure, the default node group in Amazon Web Services (AWS), or the default node
pool in Google Cloud Platform (GCP).
The following table summarizes additional implications of the SAS Viya platform
default deployment for each workload class:Table 1.1 Default Taints and nodeAffinity Settings per Class Workload ClassDefault SettingsStateless workloadsPrefer to schedule on nodes that are labeled workload.sas.com/class=stateless
Tolerate the following taints:
nworkload.sas.com/class=stateless:NoSchedule
nworkload.sas.com/class=stateful:NoSchedule
Stateful workloadsPrefer to schedule on nodes that are labeled workload.sas.com/class=stateful
Plan the Workload Placement 7
|
Workload ClassDefault SettingsTolerate the following taints:
nworkload.sas.com/class=stateful:NoSchedule
nworkload.sas.com/class=stateless:NoSchedule
Compute workloadsPrefer to schedule on nodes that are labeled workload.sas.com/class=compute
Tolerate the taint workload.sas.com/class=compute:NoSchedule
CAS workloadsPrefer to schedule on nodes that are labeled workload.sas.com/class=cas
Tolerate the taint workload.sas.com/class=cas:NoSchedule
Connect workloads
Note: This workload
class is not required if you
are using dynamically
launched pods.Prefer to schedule on nodes that are labeled workload.sas.com/class=connect
Tolerate the taint workload.sas.com/class=connect:NoSchedule
To obtain further granularity for scheduling your CAS controllers and CAS workers,
divide your CAS workloads between cascontroller and casworker workload groups.Table 1.2 Default Taints and nodeAffinity Settings for Cascontroller and Casworker Classes Workload ClassDefault SettingsCascontroller workloadsPrefer to schedule on nodes that are labeled workload.sas.com/class=cascontroller
Tolerate the following taints:
nworkload.sas.com/class=cas:NoSchedule
nworkload.sas.com/class=cascontroller:NoSchedule
Casworker workloadsPrefer to schedule on nodes that are labeled workload.sas.com/class=casworker
Tolerate the following taints:
nworkload.sas.com/class=cas:NoSchedule
nworkload.sas.com/class=casworker:NoSchedule
By default, the CAS operator uses auto-resourcing to utilize the full resources of a
|
nworkload.sas.com/class=casworker:NoSchedule
By default, the CAS operator uses auto-resourcing to utilize the full resources of a
node. When a CAS workload is scheduled to run on a specific node, no other
application pods can be scheduled to run on that node. If you want to disable auto-
resourcing and modify resource requests manually, consider configuring guaranteed
QoS on these pods. For more information, see the README file located at sas-bases/overlays/cas-server/auto-resources/README.md
(for Markdown format) 8Chapter 1 / Pre-installation Tasks
|
or $deploy/sas-bases/docs/auto_resources_for_cas_server_for_the_sas_viya_platform.htm
(for HTML
format).
If you deploy the SAS Viya platform into a cluster with untainted nodes, the
Kubernetes scheduler may attempt to schedule CAS onto those untainted nodes
instead of the nodes that have been labeled with workload.sas.com/class=cas
. An
overlay that ensures that CAS servers are only scheduled onto nodes that have the
dedicated label is available. For more information, see the “Optional CAS Server
Placement Configuration” section of the README file located at $deploy/sas-bases/overlays/cas-server/README.md
(for Markdown format) or at $deploy/sas-bases/docs/mpp_cas_server_for_sas_viya.htm
(for HTML format).
Multi-tenancy
If you are enabling multi-tenancy in your deployment, the tenants will share most of
the nodes with the provider tenant. However, because each tenant has its own CAS
server, the total number of nodes required for CAS for the full deployment is greater
than that for a non-multi-tenant deployment. The number of additional CAS nodes
required per tenant depends on whether the tenant is deploying SMP or MPP CAS.
Place the Workload on Nodes
SAS requires that you identify the node or nodes on which CAS pods should be
scheduled. SAS further recommends that you identify the nodes on which all
classes should be scheduled. The following commands place the workload on the
nodes in your deployment:
To list the names of the nodes:kubectl get nodes
|
nodes in your deployment:
To list the names of the nodes:kubectl get nodes
To assign a class of pods to a specific node:kubectl taint nodes node-name workload.sas.com/class=class-name:NoSchedule --overwritekubectl label nodes node-name workload.sas.com/class=class-name --overwrite
Note: The node-name can be the name of a single node, or it can be a list of node
names separated by spaces.
Here is an example that assigns only CAS pods to a node named node1.kubectl taint nodes node1 workload.sas.com/class=cas:NoSchedule --overwritekubectl label nodes node1 workload.sas.com/class=cas --overwrite
To verify the nodes that have been tainted:kubectl get nodes -o=custom-columns=NAME:.metadata.name,TAINTS:.spec.taints
To view node labels and taints at the same time:
Note: Because of the length of the command and the margin of the page, this
command appears as more than one line. The command should be entered as a
single line.Plan the Workload Placement 9
|
kubectl get nodes -o=custom-columns=NAME:.metadata.name,LABELS:.metadata.labels,TAINTS:.spec.taints
After you have assigned the pods to nodes, to check the assignments:
Note: Because of the length of the command and the margin of the page, this
command appears as more than one line. The command should be entered as a
single line.kubectl -n name-of-namespace get pod -o=custom-columns=NAME:.metadata.name,NODE:.spec.nodeName | sort
Deploy the SAS Viya Platform
Deployment Operator
Overview of the SAS Viya Platform Deployment
Operator
The SAS Viya Platform Deployment Operator provides an automated method for
deploying and updating your SAS Viya platform deployment. It runs in the
Kubernetes cluster and watches for declarative representations of SAS Viya
platform deployments in the form of custom resources (CRs) of the type
SASDeployment. The operator can watch for SASDeployments in the namespace
where it is deployed (namespace mode) or in all the namespaces in a cluster
(cluster-wide mode). Using namespace mode means that the operator and the SAS
Viya platform deployment are located in the same namespace. Operators in cluster-
wide mode have their own namespaces.
Note: SAS recommends using only one mode of operator in a cluster.
When a new SASDeployment is created or an existing SASDeployment is updated,
the operator updates the SAS Viya platform deployment to match the state that is
described in the CR. The operator determines the release of the SAS Viya platform
|
described in the CR. The operator determines the release of the SAS Viya platform
that it is working with. It also uses the appropriately versioned tools for that release
while deploying and updating the SAS Viya platform software. Thus, a single
instance of the operator can manage all SAS Viya platform deployments in the
cluster.
The SAS Viya Platform Deployment Operator requires updates occasionally but not
at the same frequency as the SAS Viya platform.
If you want to consider using the SAS Viya Platform Deployment Operator to
manage your SAS Viya platform deployments, continue reading this section for its
deployment steps. If you want to deploy and update your software using Kubernetes
commands or the sas-orchestration tool, skip this section and go to “Retrieve
Required Files” on page 18.10 Chapter 1 / Pre-installation Tasks
|
Retrieve the Files Required by the SAS Viya
Platform Deployment Operator
Note: If you plan to use a mirror registry for the deployment of your SAS Viya
platform software, you should download and install SAS Mirror Manager before you
begin your deployment. For more information, see “Create a Mirror Registry” on
page 139.
The required deployment assets are delivered in a TGZ file that is accessible from
the my.sas.com portal.
1If you plan to use a cluster-wide mode SAS Viya Platform Deployment Operator
and you have not already created a namespace that is dedicated to the
deployment of the SAS Viya Platform Deployment Operator, create that
namespace now. If you plan to use a namespace mode SAS Viya Platform
Deployment Operator and you have not created a namespace for your SAS Viya
platform deployment, create that namespace now.
Note: This namespace is referred to as sasoperator
in the examples that
follow.
2Create a directory on the kubectl machine or on a machine that can be reached
by your kubectl machine:mkdir directory-name
Note: If you created a mirror registry, the new directory should be parallel to
$deploy.
SAS recommends that you name the directory operator-deploy
, but you should
use a name that is meaningful to you. The directory is referred to as $operator-
deploy in this guide. Replace $operator-deploy with the directory name that you
prefer.
3Click the Get Started link that is provided in your Software Order Email (SOE).
The my.sas.com page opens.
|
prefer.
3Click the Get Started link that is provided in your Software Order Email (SOE).
The my.sas.com page opens.
4Examine the order information. Ensure that you are working with the order you
expect.
5Select the Downloads tab. The line above the Asset Type table indicates the
release cadence and the version of SAS Viya platform software that you will
deploy. If you want to deploy a different version, find the listing for that version on
the page.
Note: If you are changing your deployment method to use the deployment
operator prior to performing an update, be sure to download the deployment Deploy the SAS Viya Platform Deployment Operator 11
|
assets for the version you intend to update to. The target version should be used
for deploying the deployment operator.
6Select Deployment Assets from the Asset Type list, then click Download to
download the files that are required to deploy your software.
7Save the TGZ file from the my.sas.com page to the directory that you created in
step 2.
8Extract the files from the TGZ file in the same directory:tar xvfz file-name.tgz
The result is a directory structure that looks like this:$operator-deploy/└── sas-bases/ ├── base/ ├── components/ ├── docs/ ├── examples/ ├── extras/ └── overlays/
9Copy the operator files to the top level of the $operator-deploy directory and
make them writable:cd $operator-deploycp -r sas-bases/examples/deployment-operator/deploy/* .chmod +w site-config/transformer.yaml
If you are deploying the deployment operator in cluster-wide mode or into a
cluster that does not support seccomp, the operator kustomization.yaml file
should also be made writable. In such a scenario, this would be the last line in
the previous command:chmod +w site-config/transformer.yaml kustomization.yaml
The result is a directory structure that looks like this:$operator-deploy/├── kustomization.yaml├── operator-base/├── sas-bases/└── site-config/ ├── cluster-wide-transformer.yaml ├── remove-seccomp-transformer.yaml └── transformer.yaml
The kustomization.yaml file in the $operator-deploy directory is referred to as the
|
The kustomization.yaml file in the $operator-deploy directory is referred to as the
operator kustomization.yaml file throughout the documentation.
Note: Verify that you are using the correct directory for your changes: $operator-deploy/site-config
or $deploy/site-config
.12 Chapter 1 / Pre-installation Tasks
|
Edit the transformer.yaml File
Replace the variables in the $operator-deploy/site-config/transformer.yaml
file as indicated in that file. Two patches must be updated.
nThe first patch sets the name of the ClusterRoleBinding resource that is needed
by the operator. The SAS Viya Platform Deployment Operator requires cluster-
wide permissions even when the operator is deployed in namespace mode
because the SAS Viya platform deployment requires the deployment of cluster
resources. Because ClusterRoleBinding is a cluster-wide resource, it must be
unique within the cluster. Each SAS Viya Platform Deployment Operator
instance must have a unique ClusterRoleBinding.
nThe second patch sets the name of the namespace in which the operator is
being deployed, and therefore where the RBAC resources, such as
ServiceAccount, are deployed
Here is an example of the transformer.yaml file. It contains the name sasoperator for
the ClusterRoleBinding and the namespace sasoperator .
Note: If you are deploying the operator in namespace mode, the namespace value
is also used in the base kustomization.yaml file since you are also deploying the
|
is also used in the base kustomization.yaml file since you are also deploying the
SAS Viya platform into it.---apiVersion: builtinkind: PatchTransformermetadata: name: patch-transformer-sas-deployment-operatorpatch: |- # name: MUST BE PROVIDED BY USER AND UNIQUE IN CLUSTER. - op: replace path: /metadata/name value: "sasoperator" # namespace: MUST BE PROVIDED BY USER. THIS MUST BE THE NAMESPACE INTO WHICH # THE SAS DEPLOYMENT OPERATOR IS DEPLOYED. - op: add path: /subjects/0/namespace value: "sasoperator"target: annotationSelector: sas.com/component-name=sas-deployment-operator # Some required components for SAS Viya are scoped to the cluster. In order to # deploy your software, the operator requires the ClusterRoleBinding, regardless # of whether the intended scope is to the namespace or the cluster. kind: ClusterRoleBinding
Deploy the SAS Viya Platform Deployment Operator 13
|
Edit the kustomization.yaml for Cluster-Wide Mode
If the operator is being deployed in namespace mode, skip this section. However, if
the operator is being deployed in cluster-wide mode, edit the operator
kustomization.yaml to uncomment the reference to site-config/cluster-wide-transformer.yaml
. Here is an example of the file when deploying in cluster-wide
mode.resources:- operator-basetransformers:- site-config/transformer.yaml## Uncomment the following inclusion if you are deploying the## operator in cluster-wide mode.- site-config/cluster-wide-transformer.yaml
Edit the kustomization.yaml for seccomp
If you are using the operator in a cluster that does not support secure computing
mode (seccomp), uncomment the relevant line in the operator kustomization.yaml
file. Red Hat OpenShift does not support seccomp.
Here is an example:transformers: ... ## Uncomment the following inclusion if you are deploying the ## operator in a cluster that does not support seccomp. - site-config/remove-seccomp-transformer.yaml
Configure the Mirror Registry
Note: If you not using a mirror registry for your deployment, skip this section.
1Copy the mirror configuration yaml file from $operator-deploy/sas-bases/examples/mirror/
to the $operator-deploy/site-config
directory. If you are
using any cloud provider except Red Hat OpenShift, copy the mirror.yaml file. If
you are using Red Hat OpenShift, copy the mirror-flattened.yaml file.
|
you are using Red Hat OpenShift, copy the mirror-flattened.yaml file.
2Revise the mirror configuration file based on the infrastructure you are using for
your deployment.
nIf you are deploying on any cloud provider other than Red Hat OpenShift,
open the mirror.yaml file and replace each instance of {{ MIRROR-HOST }}
with the fully qualified domain name (FQDN) of the mirror registry in this
format:14 Chapter 1 / Pre-installation Tasks
|
FQDN/registry-namespace/sasdeployment
Here is an example:example.company.com/registry-namespace/sasdeployment
After replacing all instances of {{ MIRROR-HOST }}
, save and close the file.
nIf you are deploying on Red Hat OpenShift, modify the mirror-flattened.yaml
file.
1Find the image registry name and port:oc get svc -n name-of-registry-namespace
The output looks like this:Name Type Cluster-IP External-IP Ports Ageimage-registry ClusterIP 172.30.146.74 <none> 5000/TCP 174m
2Open the mirror-flattened.yaml file and replace each instance of {{ MIRROR-HOST }}
with the internal route to the registry from inside the
cluster followed by the name of the SAS Viya platform namespace. Use
this format:service-name.name-of-registry-namespace.svc:port/platform-namespace
The service-name and port values are in the output from the command in
the previous step. The name-of-registry-namespace is the same one that
you used in the command itself. The platform-namespace is the
namespace where you will deploy the SAS Viya platform. Here is an
example using the sample output from the previous step:image-registry.openshift-image-registry.svc:5000/myviya
3After replacing all instances of {{ MIRROR-HOST }}
, save and close the
mirror-flattened.yamlfile.
3In the operator kustomization.yaml file, add the path to mirror configuration yaml
file as the last entry in the transformers section. For example, if you are
|
file as the last entry in the transformers section. For example, if you are
deploying on any cloud provider except Red Hat OpenShift and you copied the
mirror.yaml file to the top level of the $operator-deploy/site-config
directory,
the entry would look like this:transformers:- site-config/transformer.yaml- site-config/mirror.yaml## Uncomment the following inclusion if you are deploying the...
If you are deploying on Red Hat OpenShift and you copied the mirror-
flattened.yaml file to the top level of the $operator-deploy/site-config
directory, the entry would look like this:transformers:- site-config/transformer.yaml- site-config/mirror-flattened.yaml## Uncomment the following inclusion if you are deploying the...
Deploy the SAS Viya Platform Deployment Operator 15
|
Add an imagePullSecret
Note: If you are not using a mirror registry or if you are using a mirror registry but
authentication is not required, skip this section.
If SAS content has been mirrored, and that mirror requires authentication, you must
configure an imagePullSecret and patch that secret into the SAS Viya Platform
Deployment Operator's Deployment resource.
All Cloud Providers
For information about creating an imagePullSecret, see Pull an Image from a
Private Registry . The secret should be added to the SAS Viya Platform Deployment
Operator's Deployment resource in the operator kustomization.yaml file. Here is an
example of the resulting kustomization.yaml file:resources:- operator-basetransformers:- site-config/transformer.yaml## Uncomment the following inclusion if you are deploying the## operator in clusterwide mode.- site-config/cluster-wide-transformer.yamlsecretGenerator:- name: site-image-pull-secret type: kubernetes.io/dockerconfigjson literals: - .dockerconfigjson={{ MY-IMAGEPULLSECRET }}patches:- patch: |- - op: add path: /spec/template/spec/imagePullSecrets value: - name: site-image-pull-secret target: kind: Deployment name: sas-deployment-operator
Red Hat OpenShift Alternative
If you are deploying on Red Hat OpenShift, you can use an existing secret that
OpenShift includes instead of creating a new one.
1Find the name of the secret:kubectl -n name-of-namespace get secret | grep default-dockercfg
|
OpenShift includes instead of creating a new one.
1Find the name of the secret:kubectl -n name-of-namespace get secret | grep default-dockercfg
16 Chapter 1 / Pre-installation Tasks
|
The output looks like this:default-dockercfg-#####
2Edit the operator kustomization.yaml file by adding a patch that contains a
reference to the secret:patches:- patch: |- - op: add path: /spec/template/spec/imagePullSecrets value: - name: output-of-step-1 target: kind: Deployment name: sas-deployment-operator
Configure Proxy Information
Note: If you are not using a forward proxy for your cluster, skip this section. If you
would like to use a proxy but limit it to the deployment you are working on, add
proxy variables to your SASDeployment Custom Resource. For details, see “Revise
the Custom Resource for Proxy Information” on page 84.
To define the same proxy information for all operator-based SAS Viya platform
deployments, you should modify the deplyment operator manifest file. The manifest
file is located at $operator-deploy/operator-base/deployment.yaml
.
The deployment.yaml file is initially organized like this: containers: - args: - reconcile command: [] env:
In the env
section, add the following environment variables: containers: - args: - reconcile command: [] env: - name: HTTP_PROXY value: "proxy-URL-for-HTTP-requests" - name: HTTPS_PROXY value: "proxy-URL-for-HTTPS-requests" - name: NO_PROXY value: "do-not-proxy-list"...
|
The do-not-proxy-list is a comma-separated list of host names, fully qualified host
names, and IP addresses that the proxy should ignore.
Here is an example of the deployment.yaml file with values:Deploy the SAS Viya Platform Deployment Operator 17
|
containers: - args: - reconcile command: [] env: - name: HTTP_PROXY value: "http://webproxy.example.com:5000" - name: HTTPS_PROXY value: "http://webproxy.example.com:5000" - name: NO_PROXY value: "localhost,noproxy.example.com,apiserver.example.com,kubernetes.default.svc,10.96.0.1"...
Note: The following values must be included in the list of values for the
NO_PROXY variable:
nkubernetes.default.svc (the Kubernetes API server)
nthe value of the KUBERNETES_SERVICE_HOST environment variable for the
cluster
Apply the SAS Viya Platform Deployment Operator
Resources to the Cluster
Run the following command from the $operator-deploy directory to deploy the SAS
Viya Platform Deployment Operator into Kubernetes:kustomize build . | kubectl -n name-of-deployment-operator-namespace apply -f -
Retrieve Required Files
Note: If you plan to use a mirror registry for the deployment of your SAS Viya
platform and you have not already done so, you should download and install SAS
Mirror Manager before beginning your deployment. For more information, see
“Create a Mirror Registry” on page 139.
1Deploying the SAS Viya platform requires a separate namespace for each
deployment of your software. If you are using the SAS Viya Platform Deployment
Operator in namespace mode, the SAS Viya platform deployment must be
performed in the same namespace that was used for the operator.
|
Operator in namespace mode, the SAS Viya platform deployment must be
performed in the same namespace that was used for the operator.
If you have not already done so, create a namespace for your SAS Viya platform
deployment.
2Create the directory for your deployment assets.18 Chapter 1 / Pre-installation Tasks
|
nIf you have created a mirror registry, you have already created the $deploy
directory and should skip to step 3.
nCreate a directory on the kubectl machine or on a machine that can be
reached by your kubectl machine:mkdir directory-name
Note: If you have deployed the SAS Viya Platform Deployment Operator, the
new directory should be parallel to $operator-deploy.
SAS recommends that you name the directory deploy
, but you should use a
name that is meaningful to you. The directory is referred to as $deploy in this
guide. Replace $deploy with the directory name that you prefer.
If you are performing multiple deployments of the SAS Viya platform in a
cluster, each $deploy directory should be named in a way that allows you to
associate a directory with a specific deployment.
Note: If you are deploying with the SAS Viya Platform Deployment Operator,
you can use a GitOps repository to store your deployment assets rather than
create a directory. If you choose to use GitOps, the instructions for $deploy
throughout the documentation apply to the GitOps directory as well.
3Click the Get Started link that is provided in your Software Order Email (SOE).
The my.sas.com page opens.
4Examine the order information. Ensure that you are working with the order you
expect.
5Select the Downloads tab. The line above the Asset Type table indicates the
release cadence and the version of SAS Viya platform software that you will
|
release cadence and the version of SAS Viya platform software that you will
deploy. If you want to deploy a different version, find the listing for that version on
the page.
6Select Deployment Assets from the Asset Type list, then click Download to
download the files that are required to deploy your software.
Save the TGZ file to the directory that you created in step 2.
7Extract the files from the TGZ file in the same directory:tar xvfz file-name.tgz
The result is a directory structure that looks like this:$deploy/└── sas-bases/ ├── base/ ├── components/ ├── docs/ ├── examples/ └── overlays/
For a description of each subdirectory, see “$deploy/sas-bases Directory” on
page 27.
8If you are performing a deployment with Kubernetes commands, skip this step.Retrieve Required Files 19
|
aCreate a new directory parallel to the $deploy directory in which to store your
license and *-certs.zip files.mkdir directory-name
You should use a name that is meaningful to you. The directory is referred to
as $license in this guide. Replace $license with the directory name that you
prefer.
bCopy the *-certs.zip file and the license file, identified by the .jwt extension, to
the $license directory.
Note: If you plan to have multiple SAS Viya platform deployments in your
cluster, you should organize the $license directory as you see fit. Whatever
strategy you use to organize, ensure that you can easily differentiate the
license and certs.zip files by order.
9Parallel to the sas-bases
subdirectory, create a new directory named site-config
.
For a description of the $deploy/site-config
directory, see “$deploy/site-config
Directory” on page 28.
Preparing for OpenShift
Note: If you are not deploying the SAS Viya platform in a Red Hat OpenShift
environment, skip this section.
Deployment steps for OpenShift are different from those required for other
infrastructures.This section describes those differences.
Workload Node Placement Considerations
Although the workload placement documentation recommends labelling and tainting
all worker nodes, you must have at least two untainted default nodes to allow for the
default ingress controller replica count. For more information, see Ingress controller
configuration parameters .20 Chapter 1 / Pre-installation Tasks
|
Security Context Constraints and Service Accounts
Overview
A deployment in OpenShift requires multiple custom security context constraints
(SCCs) to provide permissions to SAS Viya platform services. SCCs are required in
order to enable the Pods to run. (Pods provide essential SAS Viya platform
components.) In addition, some SCCs can be customized to meet your unique
requirements.
A security context acts like a request for privileges from the OpenShift API. In an
OpenShift environment, each Kubernetes Pod starts up with an association with a
specific SCC, which limits the privileges that Pod can request. An administrator
configures each Pod to run with a certain SCC by granting the corresponding
service account for that pod access to the SCC. For example, if Pod A requires its
own SCC, an administrator must grant access to that SCC for the service account
under which Pod A is launched. Use the OpenShift OC administrative command-line
tool to grant or remove these permissions.
Note: For additional details about SCC types, see “SCCs and Pod Service
Accounts” in System Requirements for the SAS Viya Platform .
Apply and Bind the Security Context
Constraints
After you have downloaded the deployment assets as described in “Retrieve
Required Files” on page 18, apply the following SCCs that pertain to your
deployment and then bind the SCC to the appropriate service account.
sas-cas-server
Every deployment on OpenShift must apply one of the following SCCs for the CAS
|
sas-cas-server
Every deployment on OpenShift must apply one of the following SCCs for the CAS
server.
Why the SCC is needed: CAS relies on SETUID, SETGID, and CHOWN
capabilities. CAS is launched by a SETUID root executable called caslaunch. By
default, caslaunch starts the CAS controller running under the runAsUser/
runAsGroup values and a root process named launchsvcs. Caslaunch connects
these processes with a pipe. The SETUID and SETGID capabilities are required by
launchsvcs in order to launch session processes under the user’s operating-system
(host) identity instead of launching them using runAsUser/runAsGroup values. The
CHOWN capability is necessary to support Kerberos execution, which requires
modification of the ownership of the cache file that is created by a direct Kerberos
connection. By default, the cache file is owned by runAsUser/runAsGroup identities,
but in order to support Kerberos, it must be owned by the user’s host identity.Preparing for OpenShift 21
|
Therefore, at a minimum, the cas-server-scc SCC must be applied. If the custom
group "CAS Host Account Required" is used, apply the cas-server-scc-host-launch
SCC. If the CAS server is configured to use a custom SSSD, apply the cas-server-
scc-sssd SSC. For more information on enabling SSSD, see “Use Kerberos
Connections to Connect to the CAS Server” on page 65.
1Apply the SCC with one of the following commands:kubectl apply -f sas-bases/examples/cas/configure/cas-server-scc.yamlkubectl apply -f sas-bases/examples/cas/configure/cas-server-scc-host-launch.yamlkubectl apply -f sas-bases/examples/cas/configure/cas-server-scc-sssd.yaml
2Bind the SCC to the service account with the command that includes the name
of the SCC that you applied:oc -n name-of-namespace adm policy add-scc-to-user sas-cas-server -z sas-cas-serveroc -n name-of-namespace adm policy add-scc-to-user sas-cas-server-host -z sas-cas-serveroc -n name-of-namespace adm policy add-scc-to-user sas-cas-server-sssd -z sas-cas-server
sas-connect-spawner
By default, no SCC is required for SAS/CONNECT and you can skip this item. The
SCC is required only if you intend to launch your SAS/CONNECT servers in the
Spawner pod, rather than in their own pods. For more information, see the
README file at $deploy/sas-bases/examples/sas-connect-spawner/openshift/README.md
(for Markdown format) or at $deploy/sas-bases/docs/granting_security_context_constraints_on_an_openshift_cluster_for_sasconnect.htm
(for HTML format).
|
(for HTML format).
Why the SCC is needed: The SAS/CONNECT Launcher must be able to launch the
SAS/CONNECT Server under end user identity.
1Apply the SCC with the following command:kubectl apply -f sas-bases/examples/sas-connect-spawner/openshift/sas-connect-spawner-scc.yaml
2Bind the SCC to the service account with the following command:oc -n name-of-namespace adm policy add-scc-to-user sas-connect-spawner -z sas-connect-spawner
sas-esp-project
To determine if your deployment includes SAS Event Stream Processing, look for it
in the "License Information" section of your Software Order Email (SOE) for the list
of products that are included in your order. If your SOE is unavailable, look for $deploy/sas-bases/examples/sas-esp-operator
in your deployment assets. If
that directory exists, then your deployment includes SAS Event Stream Processing.
If it does not exist, skip this SCC.
To run SAS Event Stream Processing projects with a user other than "sas", you
must bind the sas-esp-project service account to the nonroot SCC.
The nonroot SCC is a standard SCC defined by OpenShift. For more information
about the nonroot SCC, see Managing SCCs in OpenShift .22 Chapter 1 / Pre-installation Tasks
|
1There is no SCC to apply.
2Bind the SCC to the service account with the following command:oc -n name-of-namespace adm policy add-scc-to-user nonroot -z sas-esp-project
sas-microanalytic-score
To determine if the sas-microanalytic-score SCC is needed for your deployment,
check for a README file in your deployment assets at $deploy/sas-bases/overlays/sas-microanalytic-score/service-account/README.md
. If the
README file is present, then the SCC is available for deployments on OpenShift.
Why the SCC is needed: Security context constraints are required in an OpenShift
cluster if the sas-micro-analytic-score pod needs to mount an NFS volume. If the
Python environment is made available through an NFS mount, the service account
requires NFS volume mounting privileges.
For more information, see the README file at $deploy/sas-bases/overlays/sas-microanalytic-score/service-account/README.md
(for Markdown format) or at $deploy/sas-bases/docs/configure_sas_micro_analytic_service_to_add_service_account.htm
(for
HTML format).
1Apply the SCC with the following command:kubectl apply -f sas-bases/overlays/sas-microanalytic-score/service-account/sas-microanalytic-score-scc.yaml
2Bind the SCC to the service acount with the following command:oc -n name-of-namespace adm policy add-scc-to-user sas-microanalytic-score -z sas-microanalytic-score
sas-model-publish-kaniko
To determine if the sas-model-publish-kaniko service account exists in your
|
sas-model-publish-kaniko
To determine if the sas-model-publish-kaniko service account exists in your
deployment, check for a README file in your deployment assets at $deploy/sas-bases/examples/sas-model-publish/kaniko/README.md
. If the README file is
present and you plan to publish models with SAS Model Manager or SAS Intelligent
Decisioning to containers using kaniko, you must bind the sas-model-publish-kaniko
service account to the anyuid SCC. The anyuid SCC is a standard SCC defined by
OpenShift. For more information about the anyuid SCC, see Managing SCCs in
OpenShift . For more information about publishing models with kaniko, see the
README at $deploy/sas-bases/examples/sas-model-publish/kaniko/README.md
(for Markdown format) or at $deploy/sas-bases/docs/configure_kaniko_for_sas_model_publish_service.htm
(for HTML format).
1There is no SCC to apply.
2Bind the service account with the following command:oc -n name-of-namespace adm policy add-scc-to-user anyuid -z sas-model-publish-kaniko
sas-model-repository
To determine if the sas-model-repository SCC is needed for your deployment, check
for a README file in your deployment assets at $deploy/sas-bases/overlays/sas-model-repository/service-account/README.md
. If the README file is
present, then the SCC is available and might be required for deployments on
OpenShift.Preparing for OpenShift 23
|
Why the SCC is needed: The sas-model-repository pod requires a service account
with privileges if the Python environment is made available through an NFS mount.
NFS volumes are not permitted in the restricted SCC, so an SCC that has NFS in
the allowed volumes section is required.
For more information, see the README file at $deploy/sas-bases/overlays/sas-model-repository/service-account/README.md
(for Markdown format) or at $deploy/sas-bases/docs/configure_sas_model_repository_service_to_add_service_account.htm
(for
HTML format).
1Apply the SCC with the following command:kubectl apply -f sas-bases/overlays/sas-model-repository/service-account/sas-model-repository-scc.yaml
2Bind the SCC to the service account with the following command:oc -n name-of-namespace adm policy add-scc-to-user sas-model-repository -z sas-model-repository
sas-opendistro
Every SAS Viya platform deployment contains sas-opendistro. If your deployment
uses an internal instance of OpenSearch, the sas-opendistro is required.
Deploying OpenSearch on OpenShift requires changes to a few kernel settings. The
sysctl-transformer.yaml file can apply the necessary sysctl parameters to configure
the kernel, but it requires the following special privileges: allowPrivilegeEscalation
option enabled, allowPrivilegedContainer
option
enabled, and runAsUser
set to RunAsAny
.
Therefore, before you apply the sas-opendistro-scc.yaml file, you must modify it to
|
option
enabled, and runAsUser
set to RunAsAny
.
Therefore, before you apply the sas-opendistro-scc.yaml file, you must modify it to
enable these privileges. For the instructions to modify them, see the “Modify sas-
opendistro-scc.yaml for sysctl-transformer.yaml” section of the README file at $deploy/sas-bases/examples/configure-elasticsearch/internal/openshift/README.md
(for Markdown format) or at$deploy/sas-bases/docs/opensearch_on_red_hat_openshift.htm
(for HTML format).
Why the SCC is needed: For optimal performance, deploying OpenSearch software
requires the vm.max_map_count
parameter to be set on the Kubernetes nodes
running the stateful workloads to ensure there is adequate virtual memory available
for use with mmap for accessing the search indices. To provide a method to set this
argument automatically,the SAS Viya platform includes an optional transformer as
part of the internal-elasticsearch overlay that adds an init container to automatically
set this parameter. The init container must be run at a privileged level (using both
the privileged = true and allowPrivilegeEscalation = true security context options)
since it modifies the kernel parameters of the host. The container terminates after it
sets the kernel parameter, it terminates, and the OpenSearch software will then
proceed to start as a non-privileged container.
|
sets the kernel parameter, it terminates, and the OpenSearch software will then
proceed to start as a non-privileged container.
1Apply the SCC with the following command:kubectl apply -f sas-bases/examples/configure-elasticsearch/internal/openshift/sas-opendistro-scc.yaml
2Bind the SCC to the service account with the following command:oc -n name-of-namespace adm policy add-scc-to-user sas-opendistro -z sas-opendistro
24 Chapter 1 / Pre-installation Tasks
|
SAS Watchdog
SAS Watchdog is included in every SAS Viya platform deployment. If you choose to
deploy SAS Watchdog, the SCC is required. For more information, see the
README file at $deploy/sas-bases/overlays/sas-programming-environment/watchdog/README.md
(for Markdown format) or at $deploy/sas-bases/docs/configuring_sas_compute_server_to_use_sas_watchdog.htm
(for HTML format).
Why the SCC is needed: SAS Watchdog monitors processes to ensure that they
comply with the terms of LOCKDOWN system option. It emulates the restrictions
imposed by LOCKDOWN by restricting access only to files that exist in folders that
are allowed by LOCKDOWN. It therefore requires elevated privileges provided by
the custom SCC.
1Apply the SCC with the following command:kubectl apply -f sas-bases/examples/sas-programming-environment/watchdog/sas-watchdog-scc.yaml
2Bind the SCC to the service account with the following command:oc -n name-of-namespace adm policy add-scc-to-user sas-watchdog -z sas-programming-environment
sas-programming-environment
The SAS Watchdog service account includes the same permissions as the sas-
programming-environment service account described here. If you have already
bound the SAS Watchdog service account as described above, you should skip this
step.
1If your deployment does not require hostPath volume mounts, bind the nonroot
|
step.
1If your deployment does not require hostPath volume mounts, bind the nonroot
SCC to the sas-programming-environment service account using this command:oc -n name-of-namespace adm policy add-scc-to-user nonroot -z sas-programming-environment
2If your deployment requires hostPath volume mounts, bind the hostmount-anyuid
SCC to the sas-programming-environment service account using this command:oc -n name-of-namespace adm policy add-scc-to-user hostmount-anyuid -z sas-programming-environment
Note: The nonroot and hostmount-anyuid SCCs are standard SCCs defined by
OpenShift. For more information about them, see Managing SCCS in OpenShift .
Enable hostPath Mounts for CAS
Note: Using a hostPath volume for storage is supported, but it potentially entails
multiple security risks. SAS recommends that you use hostPath volumes only if you
are knowledgeable of those risks. For more information, see the "hostPath" section
of the Kubernetes Storage volume documentation .
If you want to use hostPath mounts for CAS in your OpenShift deployment, perform
the following steps:Preparing for OpenShift 25
|
1Modify the CAS SCC you chose to use at “Security Context Constraints and
Service Accounts” on page 21.
aSet allowHostDirVolumePlugin
to true.
bAdd hostPath
to the list of volume types. Here is an example:volumes:- configMap- downwardAPI- emptyDir- persistentVolumeClaim- projected- secret- nfs- hostPath
2Run the following command on the CAS worker node with the directory to be
used as a hostPath:sudo chcon -t container_file_t -R directory-path
Replace directory-path with the directory path to be used as a hostPath mount.
This directory must not be located in the file system’s root directory (/).
Note: If you prefer, you can use a MachineConfig to apply the same chcon
change instead. Consult your cluster administrator to determine which option is
best for your deployment.
Networking
OpenShift uses routes for its ingress controllers. The initial kustomization.yaml file
does not account for routes, so you must change the resources block. Here is the
original kustomization.yaml content:resources:- sas-bases/base- sas-bases/overlays/network/networking.k8s.io
Here is the revision for OpenShift:resources:- sas-bases/base- sas-bases/overlays/network/route.openshift.io
For details, see the README file at $deploy/sas-bases/examples/security/README.md
(for Markdown format) or at $deploy/sas-bases/docs/configure_network_security_and_encryption_using_sas_security_certificate_framework.htm
(for HTML format).26 Chapter 1 / Pre-installation Tasks
|
TLS
Because OpenShift uses routes instead of Ingress, the security component listed in
the initial kustomization.yaml file must be revised. Here is the original
kustomization.yaml content:components:- sas-bases/components/security/core/base/full-stack-tls- sas-bases/components/security/network/networking.k8s.io/ingress/nginx.ingress.kubernetes.io/full-stack-tls
Here is the revision for OpenShift:components:- sas-bases/components/security/core/base/full-stack-tls- sas-bases/components/security/network/route.openshift.io/route/full-stack-tls
For details, see the README file at $deploy/sas-bases/examples/security/README.md
(for Markdown format) or at $deploy/sas-bases/docs/configure_network_security_and_encryption_using_sas_security_certificate_framework.htm
(for HTML format).
Additional Security
You must perform two additional tasks to apply security for OpenShift:
nremove the Secure Computing (seccomp) profile
nupdate the fsGroup field
For the instructions to perform both tasks, see the README file at $deploy/sas-bases/examples/security/container-security/README.md
(for Markdown
format) or $deploy/sas-bases/docs/modify_container_security_settings.htm
(for HTML format).
Directory Structure
$deploy/sas-bases Directory
The $deploy/sas-bases
directory contains the files that are provided by SAS to
deploy your software. Here is the structure:sas-bases/├── base/├── components/
Directory Structure 27
|
├── docs/├── examples/├── extras/└── overlays/
The examples subdirectory contains files that are organized by categories, such as
CAS and security, that control the configuration of your deployment. README.md
files describe specific configuration settings and the files and modifications that are
required to enable those settings. Example files contain the variables that control
configuration settings, such as the number of workers in an MPP deployment of
CAS. If you decide to enable the configuration setting that is described in a
README.md file, you should copy the example file that is listed in the README.md
file to the site-config directory, and replace the variables with the values that you
want for your deployment.
The overlays subdirectory also contains files, organized by category, that affect the
configuration of your deployment. However, unlike the examples directory, the files
in the overlays directory do not need to be modified before they are used. Therefore,
they also do not have to be copied to the site-config directory. To use an overlay,
you add a reference to the specific file to your base kustomization.yaml file.
The docs subdirectory contains HTML versions of the README files in your
software order with an HTML index for them.
Depending on the software in your order, you might have an extras subdirectory. It
contains configuration files for special deployment options. It is also subdivided into
|
contains configuration files for special deployment options. It is also subdivided into
examples and overlays subdirectories that act in the same way as the higher level
examples and overlays directories.
The base and components subdirectories do not contain any files that you should
revise.
$deploy/site-config Directory
Overview
If you are following the instructions in this guide, you have created a directory
structure for your SAS Viya platform deployment that looks like this:$deploy/├── sas-bases/└── site-config/
The $deploy/sas-bases
directory contains the files that SAS provides to deploy
your software. During the process of deploying your software, you will be instructed
to edit files in the sas-bases directory. You should not edit these files directly.
Instead, copy the file to the $deploy/site-config
directory, ensure that the copied
file is writeable, and make your modifications to the copied file.
You are being directed to modify copied files because future updates of your
software involving replacing the sas-bases directory with an updated one. If you
store modified files that control the installation and configuration of your software in
sas-bases, those files are lost during the update. Updates do not change the files in $deploy/site-config
, although you might be asked to modify a file that exists
there.28 Chapter 1 / Pre-installation Tasks
|
Organization
Because the $deploy/site-config
directory contains files that you modify, you
should organize the directory according to your needs. However, it is important that
you plan the organization before you continue the deployment process. You should
organize your files in the site-config directory so that you and others who work on
the deployment can locate them easily.
Because the structure that you choose for your site-config directory depends on
your choices, explicit instructions on where to save specific files cannot be provided.
Instead, only general instructions can be offered. An example of a general
instruction is “Save the file in your site-config directory.” Suggestions for the
organization, however, can be provided.
By Function
You can organize the files according to the functional area of the deployment that
they are associated with. When you are working with the files, it should be clear
from this guide or a README what aspect of the deployment they relate to. Here is
an example:$deploy/├── sas-bases/└── site-config/ ├── authentication/ └── cpu/ └── storage/
By Type
Your deployment is based on a file that you create named kustomization.yaml,
which contains five main blocks. When you modify an example file or use an overlay
file, you will also be instructed to add a reference to that file in the
kustomization.yaml file. Therefore, you can consider organizing the site-config
|
kustomization.yaml file. Therefore, you can consider organizing the site-config
directory according to the blocks in kustomization.yaml. Here is an example:$deploy/├── sas-bases/└── site-config/ ├── configMapGenerator/ ├── patches/ ├── resources/ ├── secretGenerator/ └── transformers/
Directory Structure 29
|
Directories for the SAS Viya Platform Deployment
Operator
Note: If you are deploying your SAS Viya platform using Kubernetes commands or
the sas-orchestration command, skip this section.
Deploying with the SAS Viya Platform Deployment Operator requires a more
extensive directory structure that is based on the structure for the $deploy directory
that is described in a preceding section.├── $license/| ├── *-certs.zip| └── *.jwt├── $operator-deploy/| ├── kustomization.yaml| ├── operator-base/| ├── sas-bases/| └── site-config/| ├── cluster-wide-transformer.yaml| ├── remove-seccomp-transformer.yaml| └── transformer.yaml└── $deploy/ ├── kustomization.yaml ├── sas-bases/ └── site-config/
The $license directory contains the .jwt file that licenses your software and the *-
certs.zip file that contains the entitlement and CA certificates. If you have more than
one order of the SAS Viya platform, there will be multiple license and *-certs.zip
files.
The $operator-deploy directory contents, including the operator kustomization.yaml
file, are used only for configuring and deploying the SAS Viya Platform Deployment
Operator.
The $deploy directory is organized as described in a preceding section.
Configurations and customizations for your SAS Viya platform deployment are made
in the $deploy/sas-bases
and $deploy/site-config
subdirectories. You should
have a uniquely named $deploy directory for each SAS Viya platform deployment.
|
and $deploy/site-config
subdirectories. You should
have a uniquely named $deploy directory for each SAS Viya platform deployment.
Note: Do not put additional files in the $deploy directory unless directed to do so by
SAS documentation or SAS Technical Support. 30 Chapter 1 / Pre-installation Tasks
|
2
Installation
Initial kustomization.yaml File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Create the File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Common Customizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Using a Mirror Registry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Add the Update Checker to Your Deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Add a sitedefault File to Your Deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Configure TLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Add Forward Proxy Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
|
Add Forward Proxy Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Configure Container Security Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Enable Multi-tenancy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Configure SAS Image Staging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Deploy SAS Startup Sequencer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Configure High Availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Configure PostgreSQL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Specify SMP or MPP CAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Configure CAS Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Use Kerberos Connections to Connect to the CAS Server . . . . . . . . . . . . . . . . . . . . . . . 65
Configure OpenSearch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
|
Configure OpenSearch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Configure SAS/CONNECT Settings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Configure SAS Programming Run-Time Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Configure SAS Workload Orchestrator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
Configure Redis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Set Default SAS LOCALE and ENCODING in SAS Launcher Service . . . . . . . . . . . . 75
Change the Location of the NFS Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Enable Access Methods Through LOCKDOWN System Option . . . . . . . . . . . . . . . . . . 75
Specify PersistentVolumeClaims to Use ReadWriteMany StorageClass . . . . . . . . . . 76
Configure Open Source Integration Points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Configure SAS/ACCESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Install the Orchestration Tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
|
Install the Orchestration Tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Create the SASDeployment Custom Resource . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Add an imagePullSecret for the SAS Viya Platform Namespace . . . . . . . . . . . . . . . . . 78
Run the create sas-deployment-cr Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Revise the Custom Resource for Proxy Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Revise the Custom Resource for Red Hat OpenShift . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8531
|
Deploy the Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Deployment Using the SAS Viya Platform Deployment Operator . . . . . . . . . . . . . . . . . 85
Deployment Using the sas-orchestration Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Deployment Using Kubernetes Commands . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Save the $deploy Directory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Readiness Service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Checks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
Usage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
Review the Logs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Customize the Readiness Check Period . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Sign In as the sasboot User . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
|
Sign In as the sasboot User . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
Promotion and Migration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Initial kustomization.yaml File
Overview
The base kustomization.yaml file ($deploy/kustomization.yaml
) is where you
customize your Kubernetes deployment and allocate resources. The manifest that is
used to deploy the software is based on the declarations in the kustomization.yaml
file.
Note: Do not rename the kustomization.yaml file. The name of the file must be
kustomization.yaml so that it is recognized by the Kustomize tool.
In this section, you are instructed to create a base kustomization.yaml file that
deploys SAS with CAS in SMP mode and with an internal instance of PostgreSQL. It
is intended to deploy an initial basic SAS Viya platform to help you understand how
the base kustomization.yaml file works before adding more customizations. After
you have deployed using the initial kustomization.yaml file, you can update it with
more customizations for a new deployment or to change your initial deployment, or
you can create your own kustomization.yaml file from scratch.
“Common Customizations” on page 36 contains the most commonly used
configurations for deployments. The full list of configurations that are available for
|
configurations for deployments. The full list of configurations that are available for
your order can be accessed in either Markdown or HTML. To see the index of
READMES available for your order in Markdown, go to /$deploy/sas-bases/README.md
. To see the index of READMEs available for your order in HTML, go to $deploy/sas-bases/docs/index.htm
. Failure to review the READMEs in your
order and perform the tasks described in them will lead to failures in deployment
and usage.
IMPORTANT Before using an existing base kustomization.yaml file for
deployments of subsequent versions of the SAS Viya platform, review the
deployment notes for changes that could invalidate the continued use of that 32Chapter 2 / Installation
|
file for the current version of the software. Although the deployment notes
focus on updates, the same issues apply to a base kustomization.yaml file
(and other .yaml files) that was created for the previous version of the
software but is being used for the current version of the software.
Note: The README files that are written in the Markdown language can be read as
plain text, but they are more useful if they are rendered as Markdown. SAS
recommends that you use a tool that renders Markdown when using the README
files.
Create the File
1Change to the directory where you saved the deployment assets in “Retrieve
Required Files” on page 18:cd /$deploy
2Create a kustomization.yaml file. Copy the following code example as the basis
of the kustomization.yaml file. For each variable that is enclosed in braces ({ }),
replace the entire variable (including braces, text, and spaces) with the value
that is appropriate for your deployment. For example, {{ NAME-OF-NAMESPACE }}
should be replaced with a namespace name, such as deploy
|
should be replaced with a namespace name, such as deploy
.namespace: {{ NAME-OF-NAMESPACE }} 1resources:- sas-bases/base- sas-bases/overlays/network/networking.k8s.io 2- site-config/security/openssl-generated-ingress-certificate.yaml 3- sas-bases/overlays/cas-server- sas-bases/overlays/crunchydata/postgres-operator- sas-bases/overlays/postgres/platform-postgres# If your deployment contains SAS Viya Programming, comment out the next line- sas-bases/overlays/internal-elasticsearch- sas-bases/overlays/update-checker- sas-bases/overlays/cas-server/auto-resources 4configurations:- sas-bases/overlays/required/kustomizeconfig.yamltransformers:# If your deployment does not support privileged containers or if your deployment# contains SAS Viya Programming, comment out the next line- sas-bases/overlays/internal-elasticsearch/sysctl-transformer.yaml- sas-bases/overlays/required/transformers.yaml- sas-bases/overlays/cas-server/auto-resources/remove-resources.yaml 4# If your deployment contains SAS Viya Programming, comment out the next line- sas-bases/overlays/internal-elasticsearch/internal-elasticsearch-transformer.yaml# Mount information# - site-config/{{ DIRECTORY-PATH }}/cas-add-host-mount.yamlcomponents:- sas-bases/components/crunchydata/internal-platform-postgres 5- sas-bases/components/security/core/base/full-stack-tls 6- sas-bases/components/security/network/networking.k8s.io/ingress/nginx.ingress.kubernetes.io/full-stack-tls 6
Initial kustomization.yaml File 33
|
patches:- path: site-config/storageclass.yaml 7 target: kind: PersistentVolumeClaim annotationSelector: sas.com/component-name in (sas-backup-job,sas-data-quality-services,sas-commonfiles,sas-cas-operator,sas-pyconfig)# License information# secretGenerator:# - name: sas-license# type: sas.com/license# behavior: merge# files:# - SAS_LICENSE=license.jwtconfigMapGenerator:- name: ingress-input behavior: merge literals: - INGRESS_HOST={{ NAME-OF-INGRESS-HOST }}- name: sas-shared-config behavior: merge literals: - SAS_SERVICES_URL=https://{{ NAME-OF-INGRESS-HOST }}:{{ PORT }} 8 # - SAS_URL_EXTERNAL_VIYA={{ EXTERNAL-PROXY-URL }} 91
If you are deploying with the SAS Viya Platform Deployment Operator, the
name of the namespace is not required.2
If you are deploying on Red Hat OpenShift, replace this line with - sas-bases/overlays/network/route.openshift.io
.3
This line is appropriate for deployments using openssl to generate the
ingress certificate. For information about using cert-manager instead of
openssl, see the README file at $deploy/sas-bases/examples/security/README.md
(for Markdown format) or at $deploy/sas-bases/docs/configure_network_security_and_encryption_using_sas_security_certificate_framework.htm
(for HTML format).4
The nodes in the CAS node pool require taints in order to prevent Kubernetes
from scheduling non-CAS workloads on them. If you enable the CAS auto-
resources option, your designated CAS nodes must already be labeled. For
|
from scheduling non-CAS workloads on them. If you enable the CAS auto-
resources option, your designated CAS nodes must already be labeled. For
more information, see “Plan the Workload Placement” on page 4. If the nodes
are not labeled, you must disable auto-resources and use hardcoded default
resource settings. To disable this option, remove the following line from the
resources section:- sas-bases/overlays/cas-server/auto-resources
And remove the following line from the transformers section:- sas-bases/overlays/cas-server/auto-resources/remove-resources.yaml5
Any lines added to the components block that are not for TLS should be
placed before the TLS lines. For example, this line for PostgreSQL comes
before the two TLS lines.6
These lines are appropriate for ingress-nginx on a full-stack TLS deployment
using openssl as the certificate generator. For the other options available for
these lines, see “Configure TLS” on page 43.34Chapter 2 / Installation
|
7
For information about the storageclass.yaml file, see “Specify
PersistentVolumeClaims to Use ReadWriteMany StorageClass” on page 76.8
The SAS_SERVICES_URL
variable is used to provide the value for the
SERVICESBASEURL option in compute server sessions. It should point to
the host name and port of the cluster's Ingress for this deployment. The
Ingress must be reachable from pods running inside the cluster. The {{ NAME-OF-INGRESS-HOST }}
must match the value of the INGRESS_HOST
variable.
If you are working with external reverse proxies or application gateways, such
as Azure Application Gateway, SAS_SERVICES_URL
and INGRESS_HOST
can be
set to a load balancer. The load balancer can be for the Ingress or external.9
The SAS_URL_EXTERNAL_VIYA
variable is optional. Use it to define an external
proxy that should be used for access from outside your internal network (for
example, in generated email messages or text messages). If you choose to
use the option, you must uncomment the line by removing the number sign
(#).
3Because this kustomization.yaml deploys Full-stack TLS using openssl as the
certificate generator, follow the steps at “Configure TLS” on page 43 for
certificates for Ingress provided by openssl.
4Because this kustomization.yaml deploys an internal instance of PostgreSQL,
follow the steps for internal instances at “Configure PostgreSQL” on page 47 to
modify the appropriate .yaml file.
5If you want to make changes to your configuration, see “Common
|
modify the appropriate .yaml file.
5If you want to make changes to your configuration, see “Common
Customizations” on page 36, which is a description of the most frequently used
configuration options. To see a list of all the configuration options for your
software, go to/$deploy/sas-bases/README.md
for Markdown versions of the
README files or $deploy/sas-bases/docs/index.htm
for HTML versions of
the README files.
Note: The README files written in the Markdown language can be read as
plain text, but they are more useful if they are rendered as Markdown. SAS
recommends that you use a tool that renders Markdown when using the
README files.
6To deploy your software:
nIf you are deploying your software with the SAS Viya Platform Deployment
Operator, first install the orchestration tool using the instructions at “Install the
Orchestration Tool” on page 77, then continue to “Create the
SASDeployment Custom Resource” on page 78.
nIf you are deploying your software with the sas-orchestration command, go to
“Deployment Using the sas-orchestration Command” on page 88.
nIf you are deploying your software with Kubernetes commands, go to
“Deployment Using Kubernetes Commands” on page 91.Initial kustomization.yaml File 35
|
Common Customizations
Overview
Configuring your SAS Viya platform deployment, including setting environment
variables, is performed by modifying example and overlay files in your deployment
assets and then adding references to those files to the base kustomization.yaml file.
This section describes the most common deployment-wide configuration decisions
that need to be made for the deployment and how to enact them.
IMPORTANT This section contains customizations generic tro the SAS Viya
platform. Additional product-specific configuration tasks are described in the
README files included in your deployment assets. To see the full list of
configurations that are available for your order, including the ones described
in this section, go to /$deploy/sas-bases/README.md
for README files in
Markdown language or $deploy/sas-bases/docs/index.htm
for README
files in HTML. Failure to review the READMEs in your order and perform the
tasks described in them will lead to failures in deployment and usage.
IMPORTANT Although the steps to configure the deployment are based on
Kubernetes tasks and the Kustomize tool, the decisions behind them are
based on how you want the SAS Viya platform to run. Successfully
configuring and deploying your software to run with the features you want
requires conversation between the Kubernetes administrator and the SAS
administrator. The Kubernetes administrator for your site should inform the
|
requires conversation between the Kubernetes administrator and the SAS
administrator. The Kubernetes administrator for your site should inform the
SAS administrator of the options available for the deployment based on the
README files in the deployment assets, perhaps even sharing the README
files and their descriptions of the options. The Kubernetes administrator
should then use the instructions in the README files to enable the options
that the SAS administrator wants to use.
For more information about example and overlay files, see “$deploy/sas-bases
Directory” on page 27.
Using a Mirror Registry
Note: For introductory information about mirror registries, including how to create
them, see “Create a Mirror Registry” on page 139.36Chapter 2 / Installation
|
Add a Mirror Registry to Your SAS Viya
Platform Deployment
Note: Complete this task only if you are deploying your SAS Viya platform with
Kubernetes commands or the sas-orchestration command. If you are deploying with
the SAS Viya Platform Deployment Operator, you have already completed this task
while deploying the operator.
Any Cloud Provider Except Red Hat OpenShift
1Copy the mirror.yaml file from sas-bases/examples/mirror/
to the site-config
directory.
2Open the mirror.yaml file and replace each instance of {{ MIRROR-HOST }}
with
the fully qualified domain name (FQDN) of the mirror registry and the registry
namespace in this format:FQDN/registry-namespace/sasdeployment
Here is an example:example.company.com/registry-namespace/sasdeployment
After replacing all instances of {{ MIRROR-HOST }}
, save and close the file.
3In the base kustomization.yaml file, add the path to the mirror.yaml immediately
to the transformers block, immediately after the entry for sas-bases/overlays/required/transformers.yaml
. Here is an example:transformers:...- sas-bases/overlays/required/transformers.yaml- site-config/mirror.yaml
4Add the following content to the configMapGenerator block in the base
kustomization.yaml file:configMapGenerator:- name: input behavior: merge literals: - IMAGE_REGISTRY=registry-host:registry-port/registry-namespace...
In the value for IMAGE_REGISTRY
:
nregistry-host is required.
|
In the value for IMAGE_REGISTRY
:
nregistry-host is required.
nregistry-port is optional. If registry-port is omitted, the colon (:) that precedes
it should be omitted also.
nregistry-namespace is optional. registry-namespace is not the same as the
Kubernetes namespace. If registry-namespace is omitted, the slash (/) that
precedes it should be omitted also.Common Customizations 37
|
Using Kubernetes Commands on Red Hat
OpenShift
1Copy the mirror-flattened.yaml file from sas-bases/examples/mirror/
to the
site-config directory.
2Modify the mirror-flattened.yaml file.
aFind the image registry name and port:oc get svc -n name-of-registry-namespace
The output looks like this:Name Type Cluster-IP External-IP Ports Ageimage-registry ClusterIP 172.30.146.74 <none> 5000/TCP 174m
bOpen the mirror-flattened.yaml file and replace each instance of {{ MIRROR-HOST }}
with the internal route to the registry from inside the cluster followed
by the name of the SAS Viya platform namespace. Use this format:service-name.name-of-registry-namespace.svc:port/platform-namespace
The service-name and port values are in the output from the command in the
previous step. The name-of-registry-namespace is the same one that you
used in the command itself. The platform-namespace is the namespace
where you will deploy the SAS Viya platform. Here is an example using the
sample output from the previous step:image-registry.openshift-image-registry.svc:5000/myviya
cAfter replacing all instances of {{ MIRROR-HOST }}
, save and close the file.
3In the base kustomization.yaml file, add the path to the mirror-flattened.yaml file
to the transformers block, immediately after the entry for sas-bases/overlays/required/transformers.yaml
|
to the transformers block, immediately after the entry for sas-bases/overlays/required/transformers.yaml
. Here is an example:transformers:...- sas-bases/overlays/required/transformers.yaml- site-config/mirror-flattened.yaml
4Add the following content to the configMapGenerator block in the base
kustomization.yaml file:configMapGenerator:- name: input behavior: merge literals: - IMAGE_REGISTRY=service-name.name-of-registry-namespace.svc:port/platform-namespace...
The value for IMAGE_REGISTRY=
is the same value that you used for {{ MIRROR-HOST }}
in step 2 above.38Chapter 2 / Installation
|
Using the sas-orchestration Tool on Red Hat
OpenShift
Add content to the configMapGenerator block in the base kustomization.yaml file.
1Find the image registry name and port:oc get svc -n name-of-registry-namespace
The output looks like this:Name Type Cluster-IP External-IP Ports Ageimage-registry ClusterIP 172.30.146.74 <none> 5000/TCP 174m
2Add the following content to the base kustomization.yaml fileconfigMapGenerator:- name: input behavior: merge literals: - IMAGE_REGISTRY=service-name.name-of-registry-namespace.svc:port/platform-namespace...
The service-name and port values are in the output from the command in the
previous step. The name-of-registry-namespace is the same one that you used
in the command itself. The platform-namespace is the namespace where you will
deploy the SAS Viya platform. Here is an example using the sample output
above:configMapGenerator:- name: input behavior: merge literals: - IMAGE_REGISTRY=image-registry.openshift-image-registry.svc:5000/myviya...
Use ImagePullSecrets to Access the Mirror
Registry
If the target registry that you deposit the images in requires imagePullSecrets to
access them, you must configure an imagePullSecret and add a reference to it in
the base kustomization.yaml file.
All Cloud Providers
1Create a representation of an imagePullSecret with kubectl’s --dry-run
option.
For the steps to do so, see Pull an Image from a Private Registry .Common Customizations 39
|
2Here is the command to obtain the unencoded value of the .dockerconfigjson
field of imagePullSecret required for {{ MY-IMAGEPULLSECRET }}
. The
credentials used in these examples vary by cloud provider and by local cloud
provider conventions adopted by your organization.kubectl create secret docker-registry --dry-run=client regcred \ --docker-server=cloud-provider-specific-registry-location \ --docker-username=Docker-username \ --docker-password=Docker-password \ --docker-email=youremail@example.com \ --output="jsonpath={.data.\.dockerconfigjson}" | base64 --decode
Note: Use the following for the Docker-username , based on your cloud provider:
nFor Microsoft Azure, use 00000000-0000-0000-0000-000000000000
.
nFor Amazon ECR, use AWS
.
nFor Google Cloud Platform, use oauth2accesstoken
.
nFor Red Hat OpenShift, use the user name of the user. If you are unsure of
the user name, use the command oc whoami
.
You should be familiar with the cloud-provider-specific-registry-location because
you used it to create your mirror registry. For example, if you are using Microsoft
Azure as your cloud provider, cloud-provider-specific-registry-location is myregistry.azurecr.io
.
3Add the following content to the secretGenerator block:secretGenerator:- name: sas-image-pull-secrets behavior: replace type: kubernetes.io/dockerconfigjson literals: - .dockerconfigjson={{ MY-IMAGEPULLSECRET }}...{{ MY-IMAGEPULLSECRET }}
should be replaced with the output of the command
|
should be replaced with the output of the command
in step 2.
You can also optionally save .dockerconfigjson to a local file and reference it in
your secretGenerator with the “files” keyword instead of “literals”.
Red Hat OpenShift Alternative
If you are deploying on Red Hat OpenShift, you can use the contents of an existing
secret to generate the imagePullSecret instead of creating your own secret.
1To get the contents of the docker-dockercfg secret, find the name of the secret:kubectl -n name-of-namespace get secret | grep default-dockercfg
The output looks like this:default-dockercfg-#####
2Create a file that contains the contents of the secret:40Chapter 2 / Installation
|
kubectl -n name-of-namespace get secret output-from-step-1 --output="jsonpath={.data.\.dockercfg}" | base64 --decode > name-of-namespace.default.dockercfg.json
3Edit the secretGenerator block of the base kustomization.yaml file:secretGenerator:- name: sas-image-pull-secrets behavior: replace type: kubernetes.io/dockercfg files: - '.dockercfg=name-of-namespace.default.dockercfg.json'...
Add the Update Checker to Your Deployment
Overview
The Update Checker cron job builds a report comparing the currently deployed
release with available releases in the upstream repository. The report is written to
the stdout of the launched job pod and indicates when new content related to the
deployment is available.
If you use the initial kustomization.yaml file from this guide, the Update Checker is
deployed by default. If you do not want to deploy the Update Checker, remove the - sas-bases/overlays/update-checker
line from the resources block of the initial
kustomization.yaml file.
For more information about using the Update Checker, see “Update Checker
Report” in SAS Viya Platform Operations: Updating Software .
(Optional) Define Proxy Environment
Variables
The Update Checker cron job is defined in the sas-bases/overlays/update-checker/cronjobs.yaml
file. If you are using a proxy server for your SAS Viya
platform deployment, these environment variables must be defined on the Update
Checker cron job:
nHTTP_PROXY
nHTTPS_PROXY
nNO_PROXY
|
platform deployment, these environment variables must be defined on the Update
Checker cron job:
nHTTP_PROXY
nHTTPS_PROXY
nNO_PROXY
1Copy $deploy/sas-bases/examples/update-checker/proxy_transformer.yaml
file to $deploy/site-config/update-checker/proxy_transformer.yaml
.Common Customizations 41
|
2In the copied file, replace the value for HTTP_PROXY
with the host name, the fully
qualified host name, or the IP address of the HTTP proxy. Replace the value for HTTPS_PROXY
with the host name, the fully qualified host name, or the IP address
of the HTTPS proxy. Replace the value for NO_PROXY
with a comma-separated
list of host names, fully qualified host names, and IP addresses that the proxy
server should ignore.
3Save and close proxy_transformer.yaml.
4In the transformers section of the base kustomization.yaml file, add a reference
to the proxy_transformer.yaml file:transformers:...- site-config/update-checker/proxy_transformer.yaml
Add a sitedefault File to Your Deployment
Note: The extension for the sitedefault file for SAS Viya 4 is different from the file
extension used in SAS Viya 3.x.
nSAS Viya 3.x uses sitedefault.yml
.
nSAS Viya 4 uses sitedefault.yaml
.
Instructions in this section include references to both sitedefault files, depending on
which version is being referred to. Take note of the file extension when it is used.
The sitedefault.yaml file, which is located in the $deploy/sas-bases/examples/configuration
directory, is used primarily for the bulk loading of configuration
values. After the initial deployment, the sitedefault.yaml file cannot be used to
modify an existing value and to deploy the software again. The sitedefault.yaml file
can be used only to set new property values that have not already been set.
|
can be used only to set new property values that have not already been set.
Therefore, SAS recommends that you do not use the sitedefault.yaml file for the
initial deployment of your SAS Viya platform, except where specifically described in
this guide. You can, however, use a sitedefault file from a previous SAS Viya
release.
To load a SAS Viya 3.x sitedefault.yml file in your deployment:
1Ensure that the sitedefault.yml file is in the site-config directory.
2Add the following content to the kustomization.yaml file in the secretGenerator
block:- name: sas-consul-config behavior: merge files: - SITEDEFAULT_CONF={{ DIRECTORY-PATH }}/sitedefault.yml
42Chapter 2 / Installation
|
Configure TLS
Follow these steps to deploy using openssl as the certificate generator.
Note: The example kustomization.yaml file, located at “Create the File” on page 33,
represents a deployment with NGINX TLS for full-stack TLS mode using openssl to
generate the ingress certificate.
1Copy the $deploy/sas-bases/examples/security/openssl-generated-ingress-certificate.yaml
file to $deploy/site-config/security/openssl-generated-ingress-certificate.yaml
.
2In the resources block of the base kustomization.yaml, add a reference to the
openssl-generated-ingress-certificate.yaml file:resources:...- site-config/security/openssl-generated-ingress-certificate.yaml...
3Add the following lines to the components block of the base kustomization.yaml
file. Create the components block if it does not already exist.
aAdd a reference to the appropriate subdirectory in $deploy/sas-bases/components/security/core/base/
:components:- sas-bases/components/security/core/base/TLS-mode
Replace TLS-mode with one of the following based on the TLS mode you are
deploying:
nfront-door-tls
nfull-stack-tls
ntruststores-only
Note: If you use truststores-only
as the TLS-mode , skip part b of this
step.
bAdd a reference to the appropriate subdirectory in $deploy/sas-bases/components/security/network/
:components:- sas-bases/components/security/network/ingress-technology/TLS-mode
Replace ingress-technology with the technology that handles traffic entering
your cluster:
|
Replace ingress-technology with the technology that handles traffic entering
your cluster:
nnetworking.k8s.io/ingress/nginx.ingress.kubernetes.io
nroute.openshift.io/route
Replace TLS-mode with one of the following based on the TLS mode you are
deploying:Common Customizations 43
|
End of preview. Expand
in Data Studio
No dataset card yet
- Downloads last month
- 4