@RPoumarede thank you very much for this valuable information and indeed it is a great work of SAS R&D. We have checked inside our organization and know that for the next Kubernetes update to version 1.30 we should be fine. Pls place an update when there is a solution in place for this issue. ... View more

@JanStienstra  according to SAS documentation the sitedefault.yaml file, 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. The question is, is the sitedefault.yaml used only at initial deployment or also later when new values have been added? ... View more

I am not sure if I will be able to attend this workshop as I am in another time zone, probably I will watch it later. Could you please provide answer to a question about using the full functionality of SWO in user interface. From my research I thought you can kill and restart jobs only when using batch server via CLI. We prefer however to schedule jobs through user interface, it has more features than batch context in CLI. Has something changed here about killing and restarting jobs. We are considering to use in the future Airflow. Pls advise about the possibilities to use the full functionality of SWO. regards Karolina ... View more

Yes, we are missing this feature as well. Is this explanation also applying to Viya 4 on kubernetes? ... View more

@BrunoMueller thank you very much for this post, we have been looking for this option as there are always users who start to use the SAS Viya platform before it is ready after the update. Is there also a way to adjust the message displayed to users to a customized one when putting Viya into offline mode? regards Karolina T ... View more

Thank you for your answer ... View more

hello Wilbram, valid point, in the Viya connection you have only compute server to your disposal and through that cas server. I have tried your code, it works fine, I have only shortened the libref to 8 signs. I will not use code caslib _all_ assign anymore. Thank you ... View more

Thank you for useful command to see the node labels and suggestion to keep all the workload for compute server in one place! We will consider to try it out in our cluster. Looking forward to next blogs about how many compute nodes do you need or finding answer to questions like: do you need another stateless node or another compute node if there is significant programming workload, however based on past experience there is also a need for stateless/stateful node with more CPU than for compute node. One thing I was missing in the blog is, where can you find the default deployment for stateful/stateless nodes except of site.yaml , with the nodes affinity and tolerations requirements? ... View more

Well explained concepts, it makes clear why one needs to implement the taints. I love the scenarios, they make it easy to follow blog ... View more

I liked the part about cleaning midtier and cache. It was very clear explained. Do you need to do that in M8 as well? ... View more

hello @Wilbram-SAS , I have followed your workshop and have successfully established connection with Viya 4. I am now able to see the SAS 9.4 folders from Viya in the left panel of EG and also my mounted path in Viya. However when assigning caslibs they are not visible in the left panel. Is this correct or am I missing something? regards Karolina T ... View more

@NicolasRobert  hello Nicolas, thank you for checking. We have just updated to version 2024.04. Our users have put the libraries in different tab Resources, and then you cannot start the compute session.   ... View more

Thank you for this article. If adding more compute nodes, for example for SAS Studio, is it correct that you need more cores than memory? We are reconfiguring our resources in order to use SWO. Pls advise if possible the dependencies between CPU and memory or provide a link. ... View more

@NicolasRobert Thank you for providing clarity on this topic. I have a question: we have defined domains for oracle databases schemas. When we add such oracle database library to the compute context, then users who are not in the domain are not able to start the compute session in SAS Studio at all. Is this how the context supposed to work? We thought it will load the compute session without the library. We have added this library to the autoexec of  the compute server. Then we have no problems and users can start compute session without this library if they have no permission for it. regards Karolina  ... View more