Dear LazyWeb: HELP, is anyone doing *real* high-availability enterprise architecture for Jive 6.0 please? #jive #jivesoftware #gartner

Hey all,

Ceci n’est pas une enterprise architecture:

…it is neither layer2 not layer3; possibly it is a layer7 diagram breaking down an architecture into hypothetical components but it then seems to take great liberties in connecting “Nodes” (centre) straight into databases (bottom right) without any intervening server.

It is certainly not n-tier.

I am brand new to Jive – but I am looking at its H/A and am faintly horrified – it seems to involve setting up a cluster of nodes – behind a load balancer, which makes everything highly available, right? – and then wiring them to multiple SPOF databases and services like “document conversion”.

And don’t get me started on NFS failover, if the reference to “/NFS” means what I think it means.

So, questions for everyone:

  1. Is it unfair that my spider sense is telling me that Jive as documented is a H/A nightmare?
  2. Is anyone doing it properly, in a sense that an ex-SunPS EA would consider proper?

Did I mention that Jive is in Gartner’s magic quadrant this quarter?  Robin, this is all your fault…

 

Leave a Reply