SCT failing to connect via public adddress with `scylla-driver==3.29.8`

[fruch]

SCT is running a one node db in artifact tests

and when working with 3.29.7

we could do that:

 node_ips = ['98.93.185.64'] # public address of the node
 load_balancing_policy = WhiteListRoundRobinPolicy(node_ips)
 kwargs = dict(contact_points=node_ips, port=port, ssl_context=ssl_context)
 cluster_driver = ClusterDriver(
 auth_provider=auth_provider,
 compression=compression,
 protocol_version=protocol_version,
 load_balancing_policy=load_balancing_policy,
 default_retry_policy=FlakyRetryPolicy(),
 connect_timeout=connect_timeout,
 **kwargs,
 )
 session = cluster_driver.connect()

with 3.29.8 this is breaking and we keep on getting session trying to open on the local address, which won't be ever working:

sdcm/cluster.py:4327: in cql_connection_patient
 return self.cql_connection(**kwargs)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
sdcm/cluster.py:4239: in cql_connection
 return self._create_session(
sdcm/cluster.py:4136: in _create_session
 session = cluster_driver.connect()
 ^^^^^^^^^^^^^^^^^^^^^^^^
cassandra/cluster.py:1735: in cassandra.cluster.Cluster.connect
 ???
cassandra/cluster.py:1810: in cassandra.cluster.Cluster._new_session
 ???
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

> ???
E cassandra.cluster.NoHostAvailable: ('Unable to connect to any servers', ['10.12.11.245'])

Reviewing the changes in 3.29.8, I suspect it might be #623

[sylwiaszunejko]

@fruch Could you provide driver logs?

[fruch]

this is the log from the failing test sct.log

should have the default drivers logs in it

[dkropachev]

@sylwiaszunejko , I already know where the problem is.

[fruch]

@dkropachev any news of this one ? it's blocking us from upgrading into new driver versions

[shyamvarre1]

@dkropachev
I had a similar issue with the connectivity. I almost spent half a day on this with the customer. Could you please fix this?
Glad I found this issue. thanks @fruch

ubuntu@scylla-cql:~$ cqlsh 192.168.2.7 14002 -u cassandra -p cassandra --no-compression

Warning: Using a password on the command line interface can be insecure.
Recommendation: use the credentials file to securely provide the password.

Connected to Oss Cluster at 192.168.2.7:14002
[cqlsh 6.0.33 | Scylla 5.2.19-0.20240519.2fc86cc24117 | CQL spec 3.3.1 | Native protocol v4]
Use HELP for help.
cassandra@cqlsh> exit
ubuntu@scylla-cql:~$ pip uninstall scylla-driver==3.29.7
Found existing installation: scylla-driver 3.29.7
Uninstalling scylla-driver-3.29.7:
 Would remove:
 /home/ubuntu/.local/lib/python3.10/site-packages/cassandra/*
 /home/ubuntu/.local/lib/python3.10/site-packages/scylla_driver-3.29.7.dist-info/*
 /home/ubuntu/.local/lib/python3.10/site-packages/scylla_driver.libs/libev-a3d0bd8c.so.4.0.0
Proceed (Y/n)? y
 Successfully uninstalled scylla-driver-3.29.7
ubuntu@scylla-cql:~$ pip install scylla-driver==3.29.8
Defaulting to user installation because normal site-packages is not writeable
Collecting scylla-driver==3.29.8
 Using cached scylla_driver-3.29.8-cp310-cp310-manylinux2014_aarch64.manylinux_2_17_aarch64.manylinux_2_28_aarch64.whl (4.0 MB)
Requirement already satisfied: pyyaml>5.0 in /usr/lib/python3/dist-packages (from scylla-driver==3.29.8) (5.4.1)
Requirement already satisfied: geomet>=1.1 in ./.local/lib/python3.10/site-packages (from scylla-driver==3.29.8) (1.1.0)
Requirement already satisfied: click in /usr/lib/python3/dist-packages (from geomet>=1.1->scylla-driver==3.29.8) (8.0.3)
Installing collected packages: scylla-driver
Successfully installed scylla-driver-3.29.8
ubuntu@scylla-cql:~$ cqlsh 192.168.2.7 14002 -u cassandra -p cassandra --no-compression

Warning: Using a password on the command line interface can be insecure.
Recommendation: use the credentials file to securely provide the password.

Connection error: ('Unable to connect to any servers', ['192.168.2.5'])