Open antejavor opened 1 month ago
Hi,
Memgraph must be configured to listen on all ipv6 addresses by setting --bolt-address=::
.
To run without explicit docker network, use:
docker run --name memgraph -p 7687:7687 memgraph/memgraph:2.20.0 --log-level=TRACE --also-log-to-stderr --telemetry-enabled=False --bolt-address=::
For the host parameter in mgconsole or uri in Python driver, you can use :: (corresponding to IPv4 0.0.0.0) or ::1 (corresponding to Ipv4 127.0.0.1).
If you create an explicit ipv6 docker network, you can then also use network's ipv6 address to connect in addition to :: and ::1.
Nice, then this is a documentation issue, will transfer this issue on the docs, so we can properly document this.
Please edit your answer here 🙏 : https://stackoverflow.com/questions/74025244/how-to-connect-using-neo4j-driver-over-ipv6
We also need to document/test similar for monitoring server.
Create a IPv6 network:
docker network create --ipv6 --subnet 2001:db8::/64 ip6net
Start Memgraph in IPv6 network:
docker run --name memgraph --network ip6net -p 7687:7687 memgraph/memgraph:2.20.0 --log-level=TRACE --also-log-to-stderr --telemetry-enabled=False
Keep in mind that now Bolt port works both for IPv4 and IPv6 networking.
This can be checked via:
docker container inspect memgraph --format "{{json .NetworkSettings.Ports}}"
The problem is that connection to the
[::]
[::1]
or IPv6 address from the container"GlobalIPv6Address": "2001:db8::2",
does not work in any way, and results in the following errors, depending how you change the driver or URI config:Neo4j driver:
Snippets:
Context: