Skip to main content
cancel
Showing results for 
Search instead for 
Did you mean: 
Not applicable

How are you handling DR for your QlikView prod environment?

Hello,

Curious to hear how other enterprise users are handling Disaster Recovery for their QlikView production environment. We currently have an environment with 2 large Server & IIS boxes and 1 Publisher box at our Primary data center. At our Secondary Data Center we have the box for our development environment. It is sized like one of the production Server boxes but it has Server and Publisher installed on it.

That was fine when we first started, but QlikView has really taken off so we need a better DR environment. If we were to have a disaster at our PDC we could not run all (200+ in prod) applications on the server at the SDC. So we purchased 2 new Server boxes and another Publisher box to put at our SDC. We planned on using Qlustering to use all the boxes in normal operating conditions. If we did have a disaster at either data center, we would at least have a system still available to handle our applications (maybe in degraded performance).

However, when we reviewed our plan with Qlik, they told us that we would not be supported unless the two data centers are on the same subnet. We've asked them many times for a technical reason why and we have yet to get a clear answer. They have mentioned possible corruption of PGO and shared files, but that seemed to be under the assumption that connection speeds between the two data centers would be bad. That is not the case for us. Our data centers are fairly close (<20 miles apart) geographically and the speeds are very, very fast.

Our operations people tell us that building out a new subnet is not a trivial undertaking.

So I'm wondering if anyone else has run into this and how you've handled it.

Thanks in advance for any response.

David

2 Replies
Anonymous
Not applicable
Author

Qlik are very vague about this and told me that you need zero latency between the cluster nodes !!!

I asked if <1ms latency was ok and they insisted it has to be zero latency.  If you ask your network people for a zero latency network connection and say <1ms latency is not acceptable they may well ridicule you.

Not applicable
Author

Right. They kind of did. The latency between the data centers is almost non existant but I don't know how you could achieve ZERO between servers even if they're in same data center.