Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
We have 235 Gb source data and about 250 concurrent users. As a result of sizing we need 454 Gb of RAM and about 38 cores. But we have only servers with such configuration: 16 cores with 192 Gb RAM. Will be 2 servers plus half resources of second enough for my purposes? I mean, am I wrong, saying by using clusters we just increase amount of possible concurrent users, so each node should be able to hold qvw file in memory? Or its a good approach to size as I did: just 454/192 = ~2.4 servsers ?
Any ideas will be appreciated.
so each node should be able to hold qvw file in memory?
Yes.
WoW, I knew that. That`s really important - thank you Gysbert Wassenaar
Can I store different qvw files on different servers, but build dashboards using them inside one application?
235 Gb source data can be misleading, as depending on how you script and design your applications the resulting QVW files might be quite small.
Also the larger the QVW file the more resources are required to expand this into RAM and the more required per concurrent user.
Clustering also allows you to specify a specific server to load a specific document, useful if your servers are of different size.
As your probably aware you'll need a QVS licence per node -unlike Qlik Sense where you can have as many nodes as your want with no extra cost.
Peter Turner - But can I store files on different servers, but build dashboards using them inside one application?
Yes. But probably not in the way you mean. The final dashboard application will need to contain all the data.