As a SAS Administrator for SAS Viya  two command line tools are essential: sas-viya and the kubectl. Both commands help you to do common tasks. This post will be about using kubectl for some common tasks.   The article only introduces commands and options as they are used for the examples shown. Consult the kubectl documentation for all the available commands and options.   The following topics are covered: Basics on kubectl List all pods running a SAS compute server List all pods running either as compute server, connect server or batch job List log from a pod Basics on kubectl   The kubectl command is your interface to the Kubernetes cluster's control plane, it uses the the Kubernetes API for communication.   In order to connect you will need a profile. By default the profile is stored at this location: ~/.kube in the file named config. Ask your Kubernetes administrator for the profile. You can use the kubectl config view command to see the current config.   To get a list of all Kubernetes  resources that you can manage use the command kubectl api-resources . The output looks like this:   Select any image to see a larger version. Mobile users: To view the images, select the "Full" version at the bottom of the page.   To get a list of resources of a specific kind you can use kubectl get NAME|SHORTNAMES where NAME|SHORTNAMES is the value under the respective column headings in the output above.   Setting a default namespace   A SAS Viya environment is contained within a Kubernetes namespace. To get a list of all namespaces use the command kubectl get namespaces or kubectl get ns .   As we can see in the image above many resources need a namespace to be specified (column NAMESPACED is true). You can specify the namespace name with each command like so: kubectl get pods -n edu , or you can set a default namespace with this command kubectl config set-context --current --namespace edu ,to avoid typing the -n namespace option with every command.   List all pods running a SAS compute server   SAS Studio users use a compute server to execute the SAS program code. The compute server is running within a pod in the Kubernetes cluster. In this example we will look how we can list all the compute server pods and which user is using it.   First we need to get some information on a compute server pod. Make sure you have a compute server running. Luckily for us all the compute server pods follow this naming convention: sas-compute-server-<some-key>. So we can use the command kubectl get pods | grep sas-compute-server to get a list. The list will look similar to this:   sas-compute-server-1c685360-a73c-4d16-874e-253267158530-1997 2/2 Running 0 37m sas-compute-server-4156241b-ff03-44f6-a413-208678c8166a-2002 2/2 Running 0 20s   Each pod has some labels attached that we can use for selecting specific pods and/or display the label information. Lets see what labels are available for a compute server pod. For that we use the command kubectl describe pod <name-from-list-above> . The result will look like this (only the top portion is displayed):     Some labels of interest are: launcher.sas.com/job-type telling us the type of server and launcher.sas.com/username giving us the username. An alternative to display labels of pods is by using the --show-labels option with the kubectl get pods command.   The kubectl get pods command supports two options that help us:   Option name Description -l, --selector filter pods that have a given label, we can filter on more than one label, all conditions must be true -L, --label-columns define a comma separated list of labels to display as additional columns   Example:   $ kubectl get pods -l sas.com/created-by=sas-launcher -L launcher.sas.com/username NAME                                                         READY STATUS  RESTARTS AGE   USERNAME sas-compute-server-4156241b-ff03-44f6-a413-208678c8166a-2002 2/2   Running 0        6h26m eric sas-compute-server-d944163a-dafe-4109-a397-2e39ae843af2-2043 2/2   Running 0        92m   christine   You can add the option -o wide to the above command to see on which node the pod is executing on. If you want to sort the listing by username, you can use the --sort-by option. You must write the field to sort by as a JSONPath expression. Example:   $ kubectl get pods -l sas.com/created-by=sas-launcher -L launcher.sas.com/username --sort-by '.metadata.labels.launcher\.sas\.com/username' NAME                                                         READY STATUS  RESTARTS AGE   USERNAME sas-compute-server-d944163a-dafe-4109-a397-2e39ae843af2-2043 2/2   Running 0        94m   christine sas-compute-server-4156241b-ff03-44f6-a413-208678c8166a-2002 2/2   Running 0        6h29m eric   Note the the JSONPath must be specified in single quotes. the period in the label name must be escaped.  List all pods running either as compute server, connect server or batch job Beside a compute server pod there may be other types of SAS servers running, to get a list of all servers we can use a similar command as before with a different -l option and displaying two labels:  $ kubectl get pods -l 'sas.com/created-by=sas-launcher' -L launcher.sas.com/username,launcher.sas.com/job-type NAME                                                      READY STATUS  RESTARTS AGE   USERNAME JOB-TYPE sas-compute-server-659c96bf-8ceb-4fc2-a1e1-f97d2f4b16d2-5 2/2   Running 0        6m56s geladm   compute-server sas-connect-server-9b2aa9bb-dc88-47a8-881f-e54d6e3d6c0d-6 2/2   Running 0        6m35s geladm   sas-connect-server sas-batch-server-e2015140-ce29-41c0-bec9-b33a6a309117-7   2/2   Running 0        17s   Frank    sas-batch-job   As we can see, beside the compute server we also have a pod running that was created by a SAS/CONNECT Server as well as a pod create by running a SAS program in batch using the sas-viya batch jobs submit-pgm command.   List log from a pod   Another typical operation might be to look at the log of a given service or pod running.   The command to get the log of a specific container in a pod looks like this:   kubectl logs <pod-name> <container-name>   This command will return all the log entries available. The container name is optional.   A service of interest to the administrator is the sas-readiness service. It keeps track of the services not ready. During the start of the SAS Viya environment we can easily follow the log of this service using this command:   kubectl logs -l app=sas-readiness --tail -1 | jq   The kubectl logs command can read the logs of a container within a pod. Options used explained:   Option Description -l app=... We only want to list the log of the sas-readiness service. Using this option we do not need to specify the pod name. --tail -1 Since we use the -l to select a specific service only the last 10 lines are returned by default. So we use the --tail -1 to get all log lines -f Optionally you can specify this option to "follow" the log as new log entries are created.   The jq command will display the log lines, they are JSON format, in a more readable form. The image below shows the last two log entries from the sas-readiness service.     As we can see from the image above a log entries contains a lot of information which we might not always be interested in. We can use functionality of the jq command to only display certain fields from the JSON. The example below shows the log of the default CAS server.   kubectl -n edu logs sas-cas-server-default-controller | jq -R -r '. as $line | try (fromjson| "\(.timeStamp) \(.level | ascii_upcase) \(.message)" ) catch $line'   Since sometimes there are plain text messages produced by a command we use the try, catch functionality of the jq command to either test for a valid JSON or just print the plain text message.     As we can see from the image above, since we did not specify a container name, the kubectl logs command choose on. Since we have the try/catch in our jq command this line is printed as is. For the JSON based log lines only the fields we asked for are printed.   Summary   This article described three use cases for the kubectl command. We learned about the possibility to use the -l option to subset the listing of pods based on a label assigned to the pod. Likewise we learned about the -L option to display additional information. Finally we have seen some examples to list the log of a pod in different ways.   The kubectl command has many more options and commands to do many more things. So be curious and start exploring. ... View more

A SAS Studio user can create his own shortcuts either for SAS content folders or for directories on the filesystem.   In this article we learn about, how an administrator can define shortcuts to SAS content folders or a directory on the filesystem for all users of SAS Studio.   The following topics are covered: What are global shortcuts How to define them Use cases What are global shortcuts   As already mentioned beside user defined shortcuts, an administrator can define global shortcuts for all users. Since this definition is on the SAS Studio level it will apply to all users of SAS Studio independent of the compute context being used.   Lets look how the SAS Studio user sees it:   Select any image to see a larger version. Mobile users: To view the images, select the "Full" version at the bottom of the page.   As we can see from the image above, global shortcuts are decorated with a globe, so it is easy to distinguish them from the user defined shortcuts. These shortcut definitions (user defined or global) will be seen in the Explorer as well in the Open/Save dialog windows. They provide an easy way to folders/directories that are deeply nested.   How to define them   Using SAS Environment Manager   We first look at how we can define/update global shortcut definitions.   In SAS Environment Manager: We go to the Configuration page Search for the "SAS Studio" service Selected the SAS Studio service On the right side we collapse all configuration instance Click on the pencil icon for the "sas.studio" configuration instance     A new window opens where we can see all the properties. We scroll down until we find the property "globalShortcuts".       In the text field we can enter/update our shortcut definitions. Each global shortcut is defined by two lines. The lines are separated by a newline control character (\n). If you enter text in the above text field the proper newline is inserted automatically.   The first line is the label of the shortcut, this is the text seen in the Explorer pane in SAS Studio. The second line is the path: For a filesystem path simply enter the directory name. Example: "/workshop". For a SAS content path the line looks like "sascontent:<folderURI>".  Example: "sascontent:/folders/folders/3f6ab1b0-af34-46d0-8ce2-f735a4bc0a36". To determine the folderURI of a folder in SAS Content, see the URI field in SAS Drive, or use the Content page in SAS Environment Manager.   You cannot have any empty lines.   In order to see the new global shortcuts, Sign out/Sign in to SAS Studio.   Using the sas-viya CLI   The globalShortcuts property can also be set using the sas-viya configuration configurations update  command.   Follow the steps below.   Get the id of the configuration instance   We do need this later to update a JSON file.   configId=$(sas-viya --output json configuration configurations list --service SASStudio --definition-name sas.studio | jq -r '.items[] | .id')   Prepare the needed JSON file for the update   Note the configId env var is used in the file. Please make sure that do you do not have any line breaks other than the "\n" in the text for the key globalShortcuts   tee ~/sas-studio-globalshortcuts.json > /dev/null << EOF { "version": 2, "accept": "application/vnd.sas.configuration.config+json", "name": "configurations", "items": [ { "id": $configId, "metadata": { "isDefault": false, "mediaType": "application/vnd.sas.configuration.config.sas.studio+json;version=21", "services": [ "SASStudio", "studio" ] }, "version": 1, "globalShortcuts": "A global (sc)\nsascontent:/folders/folders/a665eb32-5e39-43a5-a1f8-6e7823272a8f\nworkshop (fs)\n/workshop" } ] } EOF   How does one know what the JSON file should look like? You can use this command sas-viya configuration configurations download --service SASStudio --definition-name sas.studio to download the complete configuration instance. Above I have only specified the minimum elements needed. You can edit the downloaded file and use it in the next step to set the new values. In this case the step to read the configuration id is not necessary.   Apply the change sas-viya configuration configurations update --file ~/sas-studio-globalshortcuts.json   If everything goes according to plan, the following response is returned: "PATCH" "/configuration/configurations" complete from "/home/christine/sas-studio-globalshortcuts.json"   For debugging purposes you can add the --verbose option to the sas-viya command   Use cases   Here are a few use cases where global shortcuts provide benefits:   Make directories available for use even when the property "fileNavigationRoot" is set to USER meaning the SAS Studio user will only see his home directory. Even when the filesystem access is switched off (property showServerFiles), a global shortcut gives you access to the defined directory. Provide simplified navigation to common directories or SAS content folders that are otherwise nested several levels.   Summary   You have to be a SAS administrator to define global shortcuts. These shortcuts apply to all SAS Studio users. There is currently now way to define them for instance based on group membership or for an individual compute server context. We have seen some uses cases where global shortcuts do have some benefits. The shortcuts are displayed independent of permissions for the "path". So you might get an error message if you try to access a "path" where you do not have the proper permission.   References   SAS Studio configuration properties make sure to select the version that matches your SAS Viya.   ... View more