I am working at a customer where we have the problem that one very large application disturbs the performance for other smaller applications. For arguments sake lets call the large application App A and the small application App B.
App A: size 1 GB, used by 5 users
App B: Size 50 MB, used by 200 users
Our idea is to set up clustered QVS machines (QVS 1 and QVS 2) where App A is loaded on QVS 1 and App B is loaded on QVS 2, so far no problems.
Question 1: What happens if QVS 1 fails, will App A then load on QVS 2 although we have specified that App A is only loaded on QVS 1 or will App A upon not finding QVS 1 now load on QVS 2?
Question 2: Is there any documentation where these failover scenarios are described?
yes, so goes Q cluster works. Really wanted to be that in case of failure of the first node of the cluster, the application will still run on the second node.
I did so.
preloading the application you chose for me the necessary node. And in web server chosen by balancing "Preloaded document". In the case when both nodes are available all works well. But in case of failure of such a situation comes.
AppA reload in QVS1. All node run
If we connect to AccessPoint appA open in QVS1.
appA reconnect and open in QVS2.
appA avalable preloaded in QVS1 and QVS2. load balance "Preloaded document"
in this case, on which server the document opens?
The experiments showed that opens on the server where this point was open session or most open sessions.
But then that did not work out as we need to. All documents are to work on the second server.
I have looked into all kind of reference documents, I can not find an answer to my question there 😞
The whole purpose with having a clustered environment in this case is to run the heavy application on one QVS so that the rest of the applications can run on the other QVS .. I don't want to put everything on the same QVS unless one of the QVS machines fail.
So the question is: considering my setup, what will happen if either QVS 1 or 2 fails?
I think the way its architecture was mounted will not get the desired effect for you. I see no other way out, but put both apps on the same server and improve performance with optimization techniques such as installation only services required for each server.
Huh? It is designed and recommended to sometimes run applications only on a specific QVS in clustered environment.
In my case my QlikView environment will consist of three machines, a backend server where QDS, DSC and QMC is running and two (clustered) machines running QVS and QVWS
Can you explain exactly how that answers my specific question on failover?
The architecture is from the QlikView system admin certification but when else than for that test would you ever set up an environment like that with the QVS and the QDS on the same machine?
Question 1: What happens if QVS 1 fails, will App A then load on QVS 2 although we have specified that App A is only loaded on QVS 1 or will App A upon not finding QVS 1 now load on QVS 2? Yes, the document will load on the second server if the other server is down.
Question 2: Is there any documentation where these failover scenarios are described? I don't know of any.