text
stringlengths 26
1.5k
|
|---|
same as the number of CPU cores on the machine. Use the --workers
option to
set the number of threads.
5Return to “Required Flags” on page 144 for information about the SAS Mirror
Manager commands. Consider whether you want to use any of the “Optional
Flags” on page 145. Then proceed to “Deploy the Software from the Registry” on
page 147 for additional deployment instructions.
Create and Populate a Mirror Registry in
Google Cloud Platform
SAS Mirror Manager requires additional configuration when the target container
registry is hosted in GCP.
Additional Requirements
In order to support Google Cloud Platform as the mirror destination, make sure you
have fulfilled the following additional requirements:150 Appendix 1 / (Optional) Using a Mirror Registry
|
nTo create a mirror registry in GCP, the GCP SDK is required. For instructions to
download and install the GCP SDK, see https://cloud.google.com/sdk/docs/
quickstart-linux .
nIn order to create a mirror registry in GCP using SAS Mirror Manager on a
Windows system, you must have Microsoft PowerShell.
Create the Registry
Create a container registry in your GCP account before you run SAS Mirror
Manager. Follow these instructions: https://cloud.google.com/container-registry/docs/
quickstart?hl=en_US .
Grant Permissions to the SAS Mirror Manager User
You must explicitly grant two storage-related permissions that enable SAS Mirror
Manager to access your GCP environment.
To change required storage permissions:
1Log in to your Google Cloud Platform account, and select Cloud Storage from
the side panel.
2Expand artifacts.$PROJECT.appspot.com and select permissions .
3In the Add Members field, enter the user name that you want to use to run SAS
Mirror Manager commands. This account can be your user account or a service
account.
4Grant the following permissions to the account that you specified:
nStorage Object Creator
nStorage Object Viewer
Using SAS Mirror Manager on GCP
The command to run SAS Mirror Manager is slightly different in GCP environments.
Use the gcloud
|
Using SAS Mirror Manager on GCP
The command to run SAS Mirror Manager is slightly different in GCP environments.
Use the gcloud
command to fetch your access token. Here is an example:mirror registry --deployment-data SASViyaV4_order-number_certs.zip --latest --destination gcr.io/$GCP_PROJECT/$NAMESPACE --username oauth2accesstoken --password $(gcloud auth print-access-token)
Note: The value for username
is always oauth2accesstoken
.
The value for $NAMESPACE
must start with a letter and can contain only lowercase
letters, numbers, hyphens, underscores, and forward slashes. Using a double
hyphen, an underscore, or a forward slash is not supported.Create and Populate a Mirror Registry in Google Cloud Platform 151
|
Return to “Required Flags” on page 144 for information about the SAS Mirror
Manager commands. “Optional Flags” on page 145. Consider whether you want to
use any of the “Optional Flags” on page 145. Then proceed to “Deploy the Software
from the Registry” on page 147 for additional deployment instructions.
Create and Populate a Mirror Registry in
Red Hat OpenShift
When the target container registry is hosted in Red Hat OpenShift, SAS Mirror
Manager creates a mirror of the SAS Viya platform software from your software
order in the OpenShift Container Platform internal image registry.
Note: Before the SAS mirror registry can be created, a cluster administrator must
configure the image registry for VMware vSphere and verify that a storage location
is available. For more information, see Configuring the registry for vSphere .
Take the following steps to create a mirror registry for an OpenShift deployment:
1Create a namespace on the OpenShift cluster. You must specify the namespace
where you will later deploy the SAS Viya platform: oc create namespace name-of-namespace
2A password is required to upload content to the image registry. Use one of the
following methods to retrieve the password:
nUsing the oc whoami
command
1Issue the following command: oc whoami -t
2Store the access token as an environment variable: export password=token
For token , substitute the token value that was returned in the previous
step.
nExtracting the password from an existing secret
|
For token , substitute the token value that was returned in the previous
step.
nExtracting the password from an existing secret
1When the namespace is created, several secrets should get created
automatically:oc -n name-of-namespace get secrets | grep builder-dockercfg
The command output should resemble the following:builder-dockercfg-##### kubernetes.io/dockercfg
2Run the following command to display the connection information in a
specified secret:kubectl get secret -n namespace builder-dockercfg-##### --output="jsonpath={.data.\.dockercfg}" | base64 --decode
152 Appendix 1 / (Optional) Using a Mirror Registry
|
For builder-dockercfg-##### , specify the name of the secret from the
previous step.
3Store the password token as an environment variable: export password=password
For password , use the long string value that was found for password in
the output from the previous step.
3Create the registry using SAS Mirror Manager. Run the command and append
the token variable. The command should resemble the following example: ./mirrormgr mirror registry --destination path-to-openshift-image-registry/project-name --flatten --deployment-data SASViyaV4_order-ID_certs.zip --latest --username user-name --password $password --path /path-to-local-image-cache
If you generated an access token, specify your login ID as the user-name .
If you extracted the password from the secret, specify the user name that was
provided in the output. For example, specify serviceaccount for the user-name .
Note: The --flatten
flag is required for a deployment into OpenShift.
Additional flags are optional. For example, you could add the --deployment-assets
or --cadence
flags in order to mirror only the content that you require, rather than
mirroring the latest versions of all supported cadences. For more information, see
“Optional Flags” on page 145.
A transformer that is applicable only to "flattened" registries has been provided in
your deployment assets. You must add it to your base kustomization.yaml file just
after overlays/required/transformers.yaml
. Proceed to “Deploy the Software
|
after overlays/required/transformers.yaml
. Proceed to “Deploy the Software
from the Registry” on page 147 for instructions.
If you are using the SAS Viya Platform Deployment Operator to manage your SAS
Viya platform deployment, verify that the data from the local mirror copy is
accessible by the operator.
Create and Populate a Mirror Registry in
JFrog Artifactory
Artifactory is a supported option for a private or internal Docker registry. SAS has
tested using the local docker registry
option from a self-hosted Artifactory.
Authentication to the registry can use either a user ID and password combination or
authentication tokens.
1Generate an API token through the JFrog Platform Access Token UI.
2Export the token as an environment variable:export accessToken=token-string
3Run SAS Mirror Manager to create the mirror registry:Create and Populate a Mirror Registry in JFrog Artifactory 153
|
mirrormgr mirror registry --remove-after-upload --platform platform-filter --destination registry-name.mydomain.com --username user-name --password $accessToken --deployment-data path-to-certs-zip-file --deployment-assets SASViyaV4_order-number_cadence_release_deploymentAssets_order-ID.tgz --path /path-to-repositories
4Return to “Required Flags” on page 144 for information about the SAS Mirror
Manager commands. Consider whether you want to use any of the “Optional
Flags” on page 145. Then proceed to “Deploy the Software from the Registry” on
page 147 for additional deployment instructions.
Create a Mirror Registry at a Dark Site
You can use SAS Mirror Manager to enable deployment of the SAS Viya platform to
an air-gapped network or to a network where systems are not allowed to
communicate with hosts on the internet. SAS Mirror Manager enables you to
download content to a local directory, from which you can copy the content to the
isolated network. From there, you can upload SAS Viya platform content to an
internal registry.
To deploy the SAS Viya platform to a dark site:
1Follow the steps in “Download Required Files” on page 142 to obtain the SAS
Viya platform software and entitlements and to download SAS Mirror Manager.
2Mirror the software to a local directory. mirrormgr mirror registry --path ./sas_repos --deployment-data SASViyaV4_order-number_certs.zip
If you do not provide a destination to the mirrormgr mirror registry
|
If you do not provide a destination to the mirrormgr mirror registry
command, all files are downloaded to the directory that is specified by the --path
flag. If you do not specify a path, ~/sas_repos
is used by default.
TIP Be sure to include any cadence version filters that you want to use
in order to limit the software that is downloaded. For more information,
see “Optional Flags” on page 145.
3Mirror the local directory to an internal registry. mirrormgr mirror registry --deployment-data SASViyaV4_order-number_certs.zip --path ./sas_repos --destination registry.example.com--username registry-user-name --password registry-password --push-only
Specifying the --destination
flag along with the --push-only
flag mirrors the
local copy to the specified registry without connecting to the SAS Entitlement
Service or to the SAS Container Registry.
Return to “Required Flags” on page 144 for more information about the SAS Mirror
Manager commands. Consider whether you want to use any of the “Optional Flags”
on page 145. Then proceed to “Deploy the Software from the Registry” on page 147
for additional deployment instructions.154 Appendix 1 / (Optional) Using a Mirror Registry
|
Appendix 2
SAS Viya Deployment Operator
Fields and Messages
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Fields in the SASDeployment Custom Resource . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
Primary Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
User File Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Secret Key Selector Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Repository Warehouse Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
Deployment User Content Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
Manage Updates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Environment Variables for the Operator Pod . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
Communications from the Operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
|
Communications from the Operator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
Status Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
Deployment Condition Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
Deployment Release Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
Deployment Release Details Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
Overview
This appendix contains lists of fields and variables that can be included in the files to
configure the SAS Viya Platform Platform Deployment Operator or that are included
in its status messages.155
|
Fields in the SASDeployment Custom
Resource
Primary Fields
The following fields can be added to the SASDeployment custom resource in the
spec (specification) block.
Note: Apply the indention rules from the preceding examples to new resources that
you createTable A2.1 Fields in the spec Block FieldDescriptioncaCertificateOptional. Information about the certificate
authority. See “User File Options” on page 157
for more information about the format.clientCertificateOptional. Information about the client certificate.
See “User File Options” on page 157 for more
information about the format.cadenceNameRequired. The name of the cadence to deploy.cadenceReleaseOptional. The release of the cadence to deploy.cadenceVersionRequired. The version of the cadence to deploy.deploymentTypeOptional. The type of deployment to use.imageRegistryOptional. The image registry to deploy from.licenseOptional. The license to apply to the deployment.
For more information about the format, see
“User File Options” on page 157repositoryWarehouseOptional. Information about the repository
warehouse. See “Repository Warehouse
Options” on page 158 for more information about
the options.156 Appendix 2 / SAS Viya Deployment Operator Fields and Messages
|
FieldDescriptionuserContentRequired. The user customizations. See
“Deployment User Content Options” on page
158 for information about the options.
User File Options
A field that requires a user file can have one of three values:Table A2.2 User File Options FieldDescriptioncontentA nested map.secretKeyRefA secret in the namespace of this SASDeployment.
See “Secret Key Selector Fields” on page 157 for more
information.urlA go-getter URL from which to fetch content. See
GitHub for more information about go-getters.
Note: If you use a go-getter URL for any of the values
in this command, that go-getter should reference a Git
repository. If it is a private Git repository, the URL must
include the user ID and the personal access token
(PAT). If it is a public Git repository, the user ID and
PAT are not required. The URL must use the git::https
protocol.
Do not use a go-getter URL to refer to local files.
Secret Key Selector Fields
A field that requires a secret key selector has two parts:Table A2.3 Secret Key Selector Fields FieldDescriptionnameRequired. The name of the secret.keyRequired. The name of the secret key. Fields in the SASDeployment Custom Resource 157
|
Here is an example of the format:spec:... caCertificate: secretKeyRef: name: secret-name key: key-name...
Repository Warehouse Options
The repository warehouse field has two options. For an example of the format, see
“Manage Updates” on page 159.Table A2.4 Repository Warehouse Options FieldDescriptionupdatePolicyOptional. The policy by which the operator
automatically applies content changes to the
repository warehouse. See “Manage Updates” on
page 159 for details of the possible values.urlOptional. The repository warehouse URL from which
to deploy your SAS Viya platform. The default is
https://ses.sas.download/ses .
Deployment User Content Options
Deployment user content must use one of these two options:Table A2.5 Deployment User Content Options FieldDescriptionfilesA nested map.urlA go-getter URL from which to fetch content. See GitHub
for more information about go-getters.
Note: If you use a go-getter URL for any of the values in
this command, that go-getter should reference a Git
repository. If it is a private Git repository, the URL must
include the user ID and the personal access token (PAT).
If it is a public Git repository, the user ID and PAT are not
required. The URL must use the git::https
protocol.158 Appendix 2 / SAS Viya Deployment Operator Fields and Messages
|
FieldDescriptionDo not use a go-getter URL to refer to local files.
Manage Updates
The SAS Viya Platform Deployment Operator can automatically download and apply
updates to your SAS Viya platform with the addition of the updatePolicy
field in the
custom resource....spec: cadenceName: "LTS" cadenceVersion: "2020.1" cadenceRelease: "20200817.1597696671121" repositoryWarehouse: updatePolicy: {{ VALUE }}
Replace the {{ VALUE }}
variable with one of the following values:
Never
The operator performs no updates. This is the default. Alternatively, if you do not
want the operator to perform updates, you can omit the updatePolicy
field
entirely.
Releases
The operator automatically applies any cadenceRelease updates within the
selected cadenceName and cadenceVersion. This setting restricts the updates
to software with two decimal places in their version numbers. For example, if you
choose "Releases" with the version/name/release in the example above, then
the operator would update the deployment for each stable release in 2020.1
(2021.1.1, 2021.1.2, and so on) and any patch releases of these versions. It
would not update to 2020.2.
Updates are checked once every 24 hours by a cronjob that is created in the SAS
Viya platform namespace at deployment time. The exact start time is randomized
with each deployment in an attempt to keep all of the deployments in a given time
zone from updating at the same time. The schedule can be modified for all SAS
|
zone from updating at the same time. The schedule can be modified for all SAS
Viya platform deployments managed by an operator by setting the
AUTOUPDATE_SCHEDULE environment variable on the SAS Deployment
Operator’s Deployment resource. The schedule can also be modified for a specific
deployment by setting the environment.orchestration.sas.com/AUTOUPDATE_SCHEDULE
annotation on the SASDeployment custom resource for that
deployment.
To prevent updates from occurring when building and applying changes to the
SASDeployment custom resource, set the custom resource‘s cadence variables to
their current values. When you are ready to perform an update or re-enable
automatic updates, modify and apply the custom resource again to remove these
settings. An example of when this strategy is useful is making permanent changes
to the base kustomization.yaml file after having made other, ad hoc, changes.Manage Updates 159
|
Environment Variables for the Operator
Pod
The following variable is required in the pod specification of the file that deploys the
operator.
SERVICE_ACCOUNT_NAME
Specifies the name of the service account under which the operator is running.
This value is typically provided by the Downward API in Kubernetes.
The following variables are optional in the pod specification of the file that deploys
the operator.
AUTOUPDATE_SCHEDULE
Specifies the schedule for the automatic update job, in CRON format. The
default is daily. SAS recommends not setting this schedule to more often than
once a day.
DEBUG_LOG_LOCATION
Enables debug logging to a file at the specified path. The directory that houses
the file must exist and be writable. The special values stdout
and stderr
enable
debug logging to standard output and standard error, respectively.
DISABLE_APPLY
Disables the application of the results of generation when the value is true
.
When combined with DISABLE_CLEANUP, this variable enables the user to test
the operator results without making any changes to their cluster.
DISABLE_AUTO_RBAC
Disables the creation of the service accounts and associated RBAC resources
that the operator normally creates for the autoupdate
and reconcile
workloads.
When the variable is set to true
, the Kubernetes administrator is responsible for
ensuring that the service accounts for these two workloads have been created
|
, the Kubernetes administrator is responsible for
ensuring that the service accounts for these two workloads have been created
and have the appropriate permissions. This variable is usually used with
SERVICE_ACCOUNT_AUTOUPDATE_NAME and
SERVICE_ACCOUNT_RECONCILE_NAME to define the names of the service
accounts the administrator created.
DISABLE_CLEANUP
Disables cleanup of the work directory when the variable is set to true
. The
operator writes temporary files to the work directory. Users can mount persistent
storage to this location and set this variable to true
to test operator results.
DISABLE_LEADER_ELECTION
Disables leader election when the variable is set to true
. The user must ensure
that no more than one instance of the operator is managing the same
namespace.
DISABLE_LIVENESS_PROBE
Disables the liveness endpoint when the variable is set to true
. Otherwise, the
operator starts a liveness endpoint and creates a Service resource in
Kubernetes for that endpoint.160 Appendix 2 / SAS Viya Deployment Operator Fields and Messages
|
DISABLE_METRICS_SERVICE
Disables the metrics service when the variable is set to true
. Otherwise, the
operator starts a metrics endpoint and creates a Service resource in Kubernetes
for that endpoint.
FLATTENED_IMAGE_REGISTRY
Excludes the deployable unit's unit path in the calculation of the image path
when the variable is set to true
.
HTTP_PROXY
When a forward proxy server is in use, the HTTP_PROXY environment variable
defines the proxy URL for HTTP requests.
HTTPS_PROXY
When a forward proxy server is in use, the HTTPS_PROXY environment
variable defines the proxy URL for HTTPS requests.
KUBECONFIG
When the operator is running outside of a Kubernetes cluster, this variable can
be set to identify and provide access to a Kubernetes cluster. By default, the
~/.kube/config file is used.
MINIMUM_UPDATE_INTERVAL
This variable is deprecated with version Stable 2020.1.1. While this variable is
still supported, you should use AUTOUPDATE_SCHEDULE
instead. See
“AUTOUPDATE_SCHEDULE ” on page 160 for more information.
NO_PROXY
When a forward proxy server is in use, the NO_PROXY environment variable
defines requests that the proxy should ignore.
SAS_UPDATE_CHECK_INTERVAL
This variable is deprecated with version Stable 2020.1.1. While this variable is
still supported, you should use AUTOUPDATE_SCHEDULE
instead. See
“AUTOUPDATE_SCHEDULE ” on page 160 for more information.
SERVICE_ACCOUNT_AUTOUPDATE_NAME
Defines the name of the service account that is used by the autoupdate
|
SERVICE_ACCOUNT_AUTOUPDATE_NAME
Defines the name of the service account that is used by the autoupdate
job
launched by the operator.
SERVICE_ACCOUNT_NAMESPACE
When the operator is running outside of a Kubernetes cluster, this variable can
be set to specify the namespace of the service account under which the operator
is running.
SERVICE_ACCOUNT_RECONCILE_NAME
Defines the name of the service account that is used by the reconcile
job
launched by the operator.
WATCH_NAMESPACE
Specifies the namespace to manage. If this variable is not set, or is set to an
empty string, then the operator functions across the cluster and watches all
namespaces.
WORK_DIRECTORY
Specifies an override for the work directory that is used by the operator.
To revise these settings, you should modify the deplyment operator manifest file.
The manifest file is located at $operator-deploy/operator-base/deployment.yaml
. The variables should be listed in the env
block using the
indentation already present in the file. Here is an example using the proxy variables: containers:
Environment Variables for the Operator Pod 161
|
- 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"...
Communications from the Operator
Overview
When the operator performs tasks, descriptions of actions and results are added to
the status block in the SASDeployment custom resource. The following tables
describe the format and content of the status block.
Here is an example of a status block for reference:apiVersion: orchestration.sas.com/v1alpha1kind: SASDeployment...status: conditions: - lastTransitionTime: "2020-10-06T13:19:24Z" message: Reconcile succeeded reason: ReconcileSucceeded status: "True" type: Ready displayName: Stable latest: releaseForCadence: cadenceName: stable cadenceRelease: "20201006.1601989747737" cadenceVersion: "2020" displayName: Stable matchesDeployedRelease: true supportEnd: "2030-03-15T04:00:00.000Z" supportLevel: SUPPORTED releaseForVersion: cadenceName: stable cadenceRelease: "20201006.1601989747737" cadenceVersion: "2020" displayName: Stable matchesDeployedRelease: true supportEnd: "2030-03-15T04:00:00.000Z"
162 Appendix 2 / SAS Viya Deployment Operator Fields and Messages
|
supportLevel: SUPPORTED messages: - "" - "" - |- role.rbac.authorization.k8s.io/sas-deployment-operator-mickey-test-a-sas-deployment created...
Status Fields
The following fields can be added to the status block of the SASDeployment custom
resource (referred to as resource in the table)..Table A2.6 Deployment Status Fields FieldDescriptionconditionsThe current state of this SASDeployment. See
“Deployment Condition Fields” on page 163 for more
information.displayNameThe display name of the cadence for this
SASDeployment.messagesThe list of messages from the last operation that was
performed by the operator on this SASDeployment.stateThe state of this SASDeployment. Valid values are PENDING
, RECONCILING
, SUCCEEDED
, or FAILED
.supportEndThe date on which support for this SASDeployment
ends.supportLevelThe support level for this SASDeployment.
Deployment Condition Fields
The following fields are used to describe the condition of the SASDeployment
custom resource.Table A2.7 Deployment Condition Fields FieldDescriptiontypeThe type of this condition. Communications from the Operator 163
|
FieldDescriptionstatusThe status of this condition.lastTransitionTimeThe timestamp that corresponds to the last status
change of this condition.reasonThe reason for this condition's last transition.messageThe description of the last transition.
Deployment Release Fields
The following fields are used to describe the cadence release and version.Table A2.8 Deployment Release Fields FieldDescriptionreleaseForVersionThe latest available release for the version of this
SASDeployment.releaseForCadenceThe latest available release for the cadence of this
SASDeployment.
For more information about either of these fields, see “Deployment Release Details
Fields” on page 164.
Deployment Release Details Fields
The following fields are used to provide details about the cadence release and
version.Table A2.9 Deployment Release Details Fields FieldDescriptioncadenceNameThe name of this available release.cadenceReleaseThe release of this available release.cadenceVersionThe version of this available release.displayNameThe displayName of this available release. 164 Appendix 2 / SAS Viya Deployment Operator Fields and Messages
|
FieldDescriptionmatchesDeployedReleaseThe value is True
if this available release is currently
deployed by this SASDeployment.supportEndThe date on which support for this available release
ends.supportLevelThe support level of this available release. Communications from the Operator 165
|
166 Appendix 2 / SAS Viya Deployment Operator Fields and Messages
|
Appendix 3
Change Deployment Methods
Kubernetes Commands to SAS Viya Platform Deployment Operator . . . . . . . . . . . 167
Deploy the SAS Viya Platform Deployment Operator . . . . . . . . . . . . . . . . . . . . . . . . . . 167
Create the $license Directory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
Create the SASDeployment Custom Resource . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
Deploy the Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
Kubernetes Commands to sas-orchestration Command . . . . . . . . . . . . . . . . . . . . . . . 169
SAS Viya Platform Deployment Operator to Any Other Method . . . . . . . . . . . . . . . . 169
Kubernetes Commands to SAS Viya
Platform Deployment Operator
Note: Changing your deployment method requires an outage for your SAS Viya
platform software.
Deploy the SAS Viya Platform Deployment
Operator
If you have not already deployed the SAS Viya Platform Deployment Operator,
follow the instructions in “Deploy the SAS Viya Platform Deployment Operator” on
page 10.167
|
Create the $license Directory
The SAS Viya Platform Deployment Operator needs an explicit location for the
license file and the zip file that contains the certificates for your software order. To
create this directory, see step 6 in “Retrieve Required Files” on page 18.
Create the SASDeployment Custom Resource
To create the custom resources required by the SAS Viya Platform Deployment
Operator, follow the instructions in “Create the SASDeployment Custom Resource”
on page 78.
When you reach the step to issue the create sas-deployment-cr
command, you
must enter the correct values for the --cadence-*
fields. To find those values, run
the following command:kubectl -n name-of-namespace get cm -o yaml | grep ' SAS_CADENCE'
Use the output of this command to determine the correct values for the create sas-deployment-cr
command.
Note: If you have manually modified any of the resources since the SAS Viya
platform was last deployed, those changes are overwritten and replaced with their
original values when this command is run.
Deploy the Software
To avoid updating your SAS Viya platform software while changing the deployment
method, ensure that the updatePolicy
field in the custom resource is set to “never”
or that the cadence variables are set to their current values. For more information,
see “Manage Updates” on page 159.
Follow the instructions in “Deployment Using the SAS Viya Platform Deployment
|
see “Manage Updates” on page 159.
Follow the instructions in “Deployment Using the SAS Viya Platform Deployment
Operator” on page 85 to deploy the software using the SAS Viya Platform
Deployment Operator.
If you did not update your software while changing deployment methods, after you
have confirmed that the operator is working as you expect, follow the instructions at
“Update to a New Version Using the Deployment Operator” in SAS Viya Platform
Operations: Updating Software to update your software.168 Appendix 3 / Change Deployment Methods
|
Kubernetes Commands to sas-
orchestration Command
Note: Changing your deployment method requires an outage for your SAS Viya
platform software.
Follow the instructions at “Deployment Using the sas-orchestration Command” on
page 88 to deploy or update your software using the sas-orchestration command.
Be sure to install the orchestration tool, described at that link, as part of the process
of changing deployment methods.
SAS Viya Platform Deployment Operator
to Any Other Method
To change deployment methods from the SAS Viya Platform Deployment Operator
to anything else requires dissociating the SAS Viya deployment from the operator.
1Dissociate the deployment from the operator by following the steps in the
README located at $deploy/sas-bases/examples/deployment-operator/disassociate/README.md
(for Markdown format) or at $deploy/sas-bases/docs/disassociate_a_sas_viya_platform_deployment_from_the_sas_viya_platform_deployment_operator.htm
(for HTML format).
2(Optional) If all the deployments deployed by the operator have been dissociated
from the operator, you can remove the operator. For the steps to remove the
operator, see “Remove the SAS Viya Platform Deployment Operator Only” on
page 133.
3After the dissociation, administer and update your deployment according to the
procedures described for your chosen deployment method.SAS Viya Platform Deployment Operator to Any Other Method 169
|
170 Appendix 3 / Change Deployment Methods
|
Appendix 4
Troubleshooting
Deployment Operator Custom Resource Is Too Large . . . . . . . . . . . . . . . . . . . . . . . . . 171
Error Checking Database Connection Error Message . . . . . . . . . . . . . . . . . . . . . . . . . . 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 . . . . . . . . . . 176
Deployment Operator Custom Resource
Is Too Large
Error
When you are performing a deployment with the SAS Viya Deployment Operator
and you attempt to apply the SASDeployment custom resource, either of the
following error messages is displayed:Error from server (RequestEntityTooLarge)The SASDeployment "sas-viya" is invalid: metadata.annotations: Too long: must have at most 262144 bytes
Explanation
|
Explanation
Applying the custom resource requires that it not exceed a size limit. The RequestEntityTooLarge
and Too long
errors indicate that the custom resource
exceeds that limit.171
|
Resolution
When the custom resource is created, it consumes the content of the $deploy
directory. To minimize the size of the custom resource, ensure that only the files
that are needed are included in the $deploy directory. For example, site.yaml
files from earlier manual deployments and old or duplicate kustomization.yaml
files should be moved from the $deploy directory or deleted.
Error Checking Database Connection
Error Message
Error
The following error message is displayed:Error checking database connection: x509: certificate signed by unknown authority
Explanation
The SAS Viya platform cannot connect securely with a database during
deployment.
Resolution
For instructions to provide CA certificates to the SAS Viya platform, see the
"Incorporating Additional CA Certificates into the SAS Viya Deployment" section
of 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).
Certframe Container Fails at init Stage
Error
The certframe container on one or more pods fails with the following log entries:[ERROR] - Certificate failed to show a status of 'Ready' with a new resource version after retrying for 300 seconds
or[ERROR] - Could not find a CertificateRequest in 'Ready=False' state associated with Certificate
Explanation
A pod has not completed initialization, and the certframe init container is moving
|
Explanation
A pod has not completed initialization, and the certframe init container is moving
into a failed state.
Resolution
1If a pod has not completed pod initialization due to the certframe container,
search the logs of the certframe container for an error message such as the
following:[ERROR] - Could not find a CertificateRequest in 'Ready=False' state associated with
172 Appendix 4 / Troubleshooting
|
Certificate sas-risk-management-8d2d37cz42-p7rj9
2If you find the error message, check the certificaterequests resource to
determine whether there are multiple requests.kubectl -n name-of-namespace --kubeconfig kubeconfig get certificaterequests | grep search-term
For search-term , use part of the pod name. From the example error
message, risk
or management
is a good search term. Any part of the pod
name can be used, but the more uncommon the search term is, the fewer
pod names there are to search through. Here is an example command that
uses the sample error message in a namespace named test1:kubectl -n test1 --kubeconfig kubeconfig get certificaterequests | grep risk
Here is sample output from the command:sas-risk-management-app-8d2d37cz42-p7rj9-5xpr7 True 123msas-risk-management-app-8d2d37cz42-p7rj9-j6g5c True 126m
3If you find multiple certificaterequests, delete the oldest ones until only one is
left.kubectl -n name-of-namespace --kubeconfig kubeconfig delete certificaterequest full-pod-name
Here is an example that uses output from the preceding steps.kubectl -n test1 --kubeconfig kubeconfig delete certificaterequest sas-risk-management-app-8d2d37cz42-p7rj9-j6g5c
4Check the pod status.
Issues with External PostgreSQL Server
Size and Connections
Note: If your deployment uses an internal instance of PostgreSQL, skip this topic.
Error
Deployments with an external instance of PostgreSQL have any of the following
messages in the logs.
|
Error
Deployments with an external instance of PostgreSQL have any of the following
messages in the logs.
nError messages associated with sas-transfer that include the following:GenericJDBCException: Unable to acquire JDBC Connection
nError messages associated with sas-backupjob:The backup agent encountered an error for the "postgres" data source.The "backup" job for the "postgres" data source finished with the status "Failed".
nThe sas-readiness service shows the sas-database-ready check going in and
out of the ready state with the following message:Issues with External PostgreSQL Server Size and Connections 173
|
Connection failed: pq: remaining connection slots are reserved for non-replication superuser connections
Explanation
These error messages indicate that the PostgreSQL database, which is required
to support SAS Infrastructure Data Server, has exhausted its pool of available
connections.
Resolution
Verify that you have met all the requirements for an external PostgreSQL
database in “External PostgreSQL Requirements” in System Requirements for
the SAS Viya Platform .
An external PostgreSQL server should support max_connections and
max_prepared_transactions of at least 1024. However, in some environments,
particularly multi-tenant SAS Viya platform deployments, this value might not be
sufficient. For more information, see “PostgreSQL Requirements for a Multi-
Tenant Deployment” in System Requirements for the SAS Viya Platform .
SAS Viya Platform Deployment Operator
Reconcile Failures
Error
After you apply the SASDeployment custom resource, checking the status
results in a FAILED
state.
Explanation
The state of the content modified for deployment, such as the
kustomization.yaml file or any of the example YAML files used for configuration,
is not what the SAS Viya Platform Deployment Operator expects. The result is a
conflict that causes the deployment to fail.
Resolution
See “Initial Troubleshooting” on page 86 for information about determining the
cause of failure. The .status.messages
field contains the actual messages from
|
cause of failure. The .status.messages
field contains the actual messages from
Kustomize and may prove the most useful in diagnosing problems.
Performing Updates Produces Many
Error Messages
Error
During the update process, you see many errors messages in the logs.
Explanation
A number of error messages come from timing issues that resolve themselves in
the course of the update. Some might come from third-party software and are 174 Appendix 4 / Troubleshooting
|
more like warnings or information messages. But since they originate with third-
party software, SAS cannot change the content of the messages.
Resolution
The error messages should begin to diminish after 30 minutes. If more errors
continue to appear in the logs, then you need to investigate.
Failure to Create Symbolic Link
SAS Configurator for Open Source requires a PVC for Python builds. If you want to
use Microsoft Azure File Services (azurefile) as a storage class for the PVC, mfsymlinks
must be supplied as a mount option to the azurefile storage class in
order to enable the update functionality. Here is an example:kubectl describe storageclass azurefile -n productionName: azurefileIsDefaultClass: NoAnnotations: noneProvisioner: kubernetes.io/azure-file Parameters: skuName=Standard_LRS AllowVolumeExpansion: True MountOptions: mfsymlinks actimeo=30 ReclaimPolicy: Delete VolumeBindingMode: Immediate Events: none
If the option is not supplied, a non-fatal message is generated in the SAS
Configurator for Open Source pod log that reports a failure to create a symbolic link
on the file system. The message provides the target of the link to be referenced.
The target can be used to reference the Python install.
Long Delay for SAS/CONNECT Server
Sign-on from SAS Studio
You might see a long (2–minute) delay during sign-on to a SAS/CONNECT server
from an internal SAS Viya platform client, such as SAS Studio. You can work around
this issue by setting TCPNOIPADDR
to 1.
|
from an internal SAS Viya platform client, such as SAS Studio. You can work around
this issue by setting TCPNOIPADDR
to 1.
This problem has been observed on Microsoft Azure, AWS, and GCP. The delay is
caused by an errant connection back to the SAS/CONNECT spawner. The spawner
hangs until time-out, at which point a successful connection is made to the launched
SAS/CONNECT server.
A fix is planned. In the meantime, run your programs with TCPNOIPADDR=1
to work
around the issue that causes the delay.Long Delay for SAS/CONNECT Server Sign-on from SAS Studio 175
|
"Operation not permitted" Error in SAS
Configurator for Open Source Logs
Error
When a pod ends up in an error state, searching the logs reveals that the
following fatal error is included for SAS Configurator for Open Source:install: cannot change permissions of '/opt/sas/viya/home/sas-pyconfig/...': Operation not permitted
Explanation
SAS Configurator for Open Source performs a number of operations in the PVC.
To use Microsoft Azure File Services (azurefile) as a storage class for the PVC,
add the noperm
mount option to the azurefile storage class. The noperm
mount
option allows client-side permissions checks to pass if options such as file_mode
or dir_mode are set on the PVC mount in the pod.
Resolution
Add the noperm
mount option to the azurefile storage class:Name: azurefileIsDefaultClass: NoAnnotations: noneProvisioner: kubernetes.io/azure-file Parameters: skuName=Standard_LRS AllowVolumeExpansion: True MountOptions: noperm actimeo=30 ReclaimPolicy: Delete VolumeBindingMode: Immediate Events: none
Uninstall Leaves the Deployment
Namespace in a Terminating State
Note: If you have uninstalled your software using the SAS Viya Platform
Deployment Operator, you should skip this note. Also, if your deployment did not
have an internal instace of PostgreSQL, you should skip this note.
Error
When you perform an uninstall, the namespace in which your software was
deployed is left in an unending Terminating
state.176 Appendix 4 / Troubleshooting
|
Explanation
When you uninstalled your software, the PostgreSQL resources were not
removed prior to uninstalling your deployment.
Resolution
1Create a list of the PostgresClusters in your deployment:kubectl -n name-of-namespace get postgresclusters
2Run the following command for each PostgresCluster returned in step 1,
replacing PostgresCluster-name with the name of the PostgresCluster.kubectl -n name-of-namespace patch postgrescluster PostgresCluster-name -p '{"metadata":{"finalizers":[]}}' --type=merge
Uninstall Leaves the Deployment Namespace in a Terminating State 177
|
178 Appendix 4 / Troubleshooting
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.