Each server started task is a single-threaded processor. For multi-threaded processing, multiple servers can be configured and run simultaneously. The optimal number of servers required at a site depends on the number of users and the frequency with which they simultaneously run CA Endevor SCM Integration for the Natural Environment online sessions.
The server is a stateless server, meaning that it treats each request as an independent transaction that is unrelated to any previous request. Each request is processed on a first-in-first-out basis without regard to which user submitted it. Due to the nature of the server processing, a dedicated server task is not required for each user.
There is no limit to the number of users a single server can support. If 100 users are running concurrent sessions, all of their list requests and action requests are processed. However, users do experience a relatively slower response to their requests than if the concurrent sessions were significantly fewer in number. The response time can vary widely from minute to minute, depending upon whether requests are submitted simultaneously. For example, if all 100 users never submit a request at the same time that another user’s request is being processed, then there is no delay in processing. In this case, it is as if each user has the server all to themselves.
If your site has multiple concurrent users who believe that their response time is unacceptably slow and a review of the server logs shows that a large number of requests are regularly processed in close succession during certain times (suggesting that multiple requests are often simultaneously queued up), then you may want to improve session response times. To do this you can alter the class or priority of the server task, add additional server tasks, or both.
Administrators can change the server and API timeout values in the startup user exit, UEXIT00. The timeout values limit the time for the server to begin processing a request and for the API to complete processing of the request. If the timeout intervals are exceeded, control is returned to the user with the option of continuing to wait for the request to complete or return to CA Endevor SCM Integration for the Natural Environment. If the user chooses not to wait for the request to complete, the request is not terminated. The server continues to wait for the request to complete, because there is no way for it to cancel a request sent to the API after it has begun. All records associated with the request are automatically cleared from the message queue file at a later time.
|
Copyright © 2014 CA.
All rights reserved.
|
|