I am working on a new SolrCloud infrastructure - ZooKeeper.
Some background information:
- all other services (mainly website infrastructure) are distributed between two data centers , with active configurations.
- at the network level, servers are configured for extended LANS, with dark fiber through data centers. Thus, latency is minimal.
- Most of these applications will use the SolrCloud infrastructure - ZooKeeper.
I got Solr Cloud and the ZooKeeper ensemble. The implementation at this level is fine.
But I am wondering how to distribute my ZooKeeper servers. I should have an odd number of servers, but I only have 2 data centers. If someone fails, I have a 50-50 chance that I will lose most.
Any ideas? So far I have been thinking:
a third data center request (unlikely to happen, $$$!)
host 2 to the data center and 1 to the external cloud provider (Amazon or ...?). Again $$$
set an odd number in data center 1 and use an observer on site 2. What happens if site 1 fails? Can SolrCloud work with only one observer?
Thanks for any idea and comments.
Nick
source
share