Is it possible to run a Neo4j cluster with strong consistency?

The Neo4j docs indicate that when working in HA mode you get possible consistency. This is a quote from this page:

All updates will, however, be distributed from the master to the other subordinates in the end, the recording from one subordinate cannot be immediately visible to all other subordinates

My question is: is there a configuration that will allow me to write a cluster with a strong sequence, of course, due to a decrease in performance? I am looking for some kind of configuration with an active passive failover cluster.

+6
source share
1 answer

There is such a configuration option. ha.tx_push_factor determines how many slave transactions need to be translated synchronously. By setting this parameter to ha.tx_push_factor=<clustersize>-1 you will have immediate full consistency.

+11
source

Source: https://habr.com/ru/post/978028/


All Articles