search

neo4j / apoc

Apache License 2.0
95 stars 28 forks source link

Problem with storing data into systemdb in a Cluster env #52

Open neo-technology-build-agent opened 2 years ago

neo-technology-build-agent commented 2 years ago

Issue by conker84 Thursday Sep 02, 2021 at 14:11 GMT Originally opened as https://github.com/neo4j-contrib/neo4j-apoc-procedures/issues/2194

Related to (#2193) in a cluster env the systemdb can have a LEADER in a different instance where the actual call apoc.trigger.add happened.

We have a couple of options here but I need to discuss them with @fbiville @JMHReif & @jexp

neo-technology-build-agent commented 2 years ago

Comment by AdrienLemaire Tuesday Dec 07, 2021 at 06:43 GMT

@conker84 @fbiville @JMHReif @jexp Hi, I've just stumbled against this issue. After setting a trigger in my local environment, I failed to deploy this feature to the neo4j cluster environment, and realized that only the system leader allows for apoc.trigger.add when my db leader is on another node.

I wish the documentation mentioned that triggers aren't yet supported on clusters. What are the couple of options available ? Is there anything I can do to get triggers to work in my cluster and not throw away all the work involving neo4j triggers I'm trying to release ?

Running neo4j 4.1 on the GCP marketpace neo4j enterprise causal cluster. Thank you for your support

Edit: As a temporary solution, I removed the core replicas, replacing them with one standalone leader and read replicas.

ncordon commented 2 years ago

Hi @AdrienLemaire, sorry for the delay, we are working on a solution for this. We'll keep you posted