Keeping the business application running 24/7 is a number one priority these days and going offline for few hours to even few minutes during planned or unplanned software upgrade is not an acceptable option anymore.
With current release of Oracle NoSQL Database you can upgrade your NoSQL cluster in a rolling fashion. There is a smart administration utility that calculates the safest combination of storage nodes that can be brought down in parallel and upgraded, keeping all the shards in the database available for read/write operations at all the time.
For example say you have deployed 1x3 database cluster (i.e. 1 shards, with 3 replicas each shard) on three physical storage nodes and you want to find out what is the safest order to upgrade your cluster. You will log on to administration console and from the prompt you will run ‘show upgrade’ command, which will show you storage node IDs in multiple rows, something like this:
kv-> show upgrade
Calculating upgrade order, target version: 22.214.171.124.8, prerequisite: 126.96.36.199.23
Storage nodes (SNs) that can all be upgraded in parallel will be represented in the same row and multiple rows represent the order in which upgrade should take place. There is also an example shell script that can be used to automate the whole process of upgrade.<KVROOT>/examples/upgrade/onlineUpgrade
Figure 1: Showing Storage Node 1 getting upgraded and all the shards in the cluster still available for read and write operation during the process.