Error seeding cluster

[bastian]

Hi.

I've set up a 4.4.5 causal cluster on 3 AWS EC2s following the neo4j documentation. I want to seed it with a dump or export from our current (non-clustered) neo4j 4.1.

So I've done this using offline dumps/loads:

1. dumped the original db
2. upgraded to 4.2 => 4.3 => 4.4.5 using local docker images as described in the docs. I've tested the 4.4.5 DB and it works fine.
3. dumped the result
4. followed the neo4j documentation to seed the new cluster

But after loading to each node and then creating the db in neo4j (this step isn't mentioned in the offline seeding docs, but in other places about loading a dump), SHOW DATABASES again shows the currentStatus as "dirty" , and error "An error occurred! Unable to create new database (...)"

I have also tried making a backup from the 4.4.5 docker instead of a dump and using the online seeder node method. This then shows currentStatus as "dirty" , and error "An error occurred! Unable to create new database (...)" on the seeder node, and currentStatus as "offline" , and error "An error occurred! "An error occurred! Unable to create new database (...)" on the other nodes.

Any help would be appreciated.

NB: When creating the cloudformation stack, using VERSION=4.4.4 as described in the docs doesn't work because the S3 link for 4.4.4 returns ACCESS DENIED. So I created 4.4.3, and then upgraded the neo4j in each EC2 to 4.4.5 . Will the template be made available to 4.4.5 and a newer Ubuntu version? Ubuntu 18's regular maintenance ends this year...

NB2: I've also posted about this on Discord but there doesn't seem to be much help there.

[bastian]

Managed to import by exporting via apoc.export.cypher.all and then importing with cypher-shell. It was slower, but it worked. It can probably also be used to export/import across bigger neo4j version increments, so that might make it actually more convenient in many cases.