QV - Clustering Best Practices

[jtanajewski]

Looking for information on building a QV clustered environment.  Currently running QV11 on single server, looking to build new 2 server cluster running QV 11R3, and then migrate existing dashboards to new cluster.

Does anyone have document/best practices on building 2 server QV cluster?

Appreciate info, and any past experiences.

[jtanajewski]

Come on no replies?  Posted over week ago, and no input?  Someone out there has to have some info/best practices Clustering servers?  Community only good as active members.

Need help, appreciate info.

[magavi_framsteg]

Hi Jim.

If you are a partner you can look at the Enterprise Framework.
It contains loads of scalability documents and whatnots.

https://eu1.salesforce.com/sfc/#version/069D0000000jeWy?retURL=https%3A%2F%2Fc.eu1.visual.force.com%...

Kind regards

Magnus Åvitsland

BI Architect Consultant

Framsteg Business Intelligence Corp.

[#4]

Hi Jim,

Have a look at this document. http://support.ipc-global.net/index.php?/Knowledgebase/Article/GetAttachment/69/120.

It provides enough detail about how to create clustered server.

What I have seen in my experience is that if you have complex calculation with big dataset in Qlikview dashboard and you are clustering your Qlikview servers, then all the cache built by users are also distributed to different servers depends upon where the user session is created and your Qlikview server will not truly take the advantage of cached memory calculations. This is why Qliktech suggest to pinned the application to specific server so that its cached is built on the same server.

But if you do that you are diverting from the purpose of clustering your Qlikview servers but only get a single console to manage them.

May be in future Qliktech should implement a way to utilize clustered cache irrespective of the server where user is diverted in clustered environment.

Hope this help.

Anosh

[Tony_Strano]

First you will need a QV cluster license to support 2 nodes.      Basically a file share that is available to both nodes and from the QMC, the share contains a folder for the root folder so both nodes can see it.     From the QMC under system, setup, qlikview servers, general tab,  add the second server.     Remember to shutdown the QMC on the second node.     You never want to run two QMCs to support 1 QV environment.     The logs,  documents, and root folder must be on the share and the QV service account has full access.  

[Anonymous]

Jim

As Anthony says above you can only run one QMC [QlikView Management Console] at a time, i.e. it cannot be clustered.

I cluster the other QlikView services primarily for resilience, as I work in a 24*7 continuous operational environment across 6 continents [we do not operate in Antarctica]  and their various timezones.  You do not mention why you are planning to cluster QV.

This left the QMC as a single point of failure, albeit one that would only affect administration - but still one that could cause pain.

So last week I moved our QMC onto a Virtual Machine, with the resilience virtualisation inherently provides.

[Note: Don't virtualise any of the other QV services].

The documents you have been given links to  above, should provide the Clustering details you require, but if you have any specific questions then feel free to ask.

Best Regards,     Bill Markham