2 weeks ago
2 weeks ago
Since the spawner (along with the metadata server) is usually started with the ID of the SAS installation user, I'd check the constraints of that userID.
OTOH, you might have developed a memory leak. See how long the spawner has been up, and stop/start it. You might consider to set up a periodic restart (eg every night).
With processes where only a single instance will be started on a given machine, you can be liberal with regards to memory consumption, while processes with many instances (workspace servers) should have settings that make sure all expectable concurrent instances fit into available memory.
a week ago
normally, that kind of a message is just a reflection of other problems, normally caused by the amount , the frequency or the duration of certain sessions. Or its incapability to close orphaned SAS sessions.
An example: Often, this is caused by long stored process sessions, or amount of spanwed STP sessions, which are meant to have short duration. The solution for those is to extend the pool of STP and improve the code for the Stored Processes, or memory settings.
Another example: if object spawner is incapable of closing orphaned SAS Workspace Server sessions, then it has to handle unuseful SAS sessions.
There are many more. I suggest to investigate the behaviour of your spawned SAS sessions. It is very rare the requirement to extend the memory settings of the SAS Object Spawner, even on really BIG deployments. It is a service very optimised.