jexp/store-utils

Running the tool with no downtime

Opened this issue · 0 comments

redserpent7 commented

Hi,

I am currently

I have a neo4j 3.5.23 cluster that consists of 3 core servers and 2 reader servers.

Recently we deleted close to 12M nodes with their relations and the db size increased dramatically, the db is currently around 120GB with 54M nodes and 456M relations.

I tried copying the DB on one of the reader instances with keep-node-ids set to false and it took around 3 hours to complete.

Now my understanding is that the entire database (all cluster nodes) need to be offline in order to do copy the store, not to mention re-applying the indexes which basically means that our production environment will be down for over 3 hours.

I was wondering if there is a way to mitigate this? what are my options to prevent going completely offline or at the most minimize the downtime to say below 30 mins if at all possible?