Originally from the User Slack
@Sabin: Hi, whenever I try to add a cluster on Scylla Manager with --username and --password, I am greeted with this error message:
Error:
Trace ID: (grep in scylla-manager logs)
On logs, I see this:
scylla-manager[15404]: {"L":"ERROR","T":"2024-07-03T14:42:46.671Z","N":"http","M":"Panic","panic":"no valid connect address for host: [HostInfo hostname=\"0.0.0.0\" connectAddress=\"0.0.0.0\" peer=\"<nil>\" rpc_address=\"<nil>\" broadcast_address=\"<nil>\" preferred_ip=\"<nil>\" connect_addr=\"0.0.0.0\" connect_addr_source=\"invalid\" port=9042 data_centre=\"\" rack=\"\" host_id=\"\" version=\"v0.0.0\" state=UP num_tokens=0]. Is your cluster configured correctly?"
But if I remove --username and --password, it is added without any errors. I am trying to restore backup from S3 repository and restore fails if I do not provide the credentials:
Error: create restore target, units and views: create worker: get CQL cluster session: cluster requires CQL authentication but username/password was not set
Trace ID: _0CgXIGBR4OIVhFO2QmN7g (grep in scylla-manager logs)
I have configured the default credentials on Scylla Manager’s configuration file.
@Felipe_Cardeneti_Mendes: what’s the full command line you are trying to run? 0.0.0.0 isn’t a valid address you know
@Sabin: This is the command that fails:
sctool cluster add --host 10.1.169.163 --name test-cluster --username cassandra --password cassandra --auth-token "vqHy2zbGLX7kB9cCyFtliLv2H7HV8iU6rY1WZW0UYJpmTpT6nm4wf0...sc"
and this one works fine except I see CQL TIMEOUT
sctool cluster add --host 10.1.169.163 --name test-cluster --auth-token "vqHy2zbGLX7kB9cCyFtliLv2H7HV8iU6rY1WZW0UYJpmTpT6nm4wf0...sc"
For the last command, the output looks like this:
@Felipe_Cardeneti_Mendes: does any of your nodes have rpc_address: 0.0.0.0 ? show all rpc_* entries on your scylla.yaml definitions
@Sabin: Hmm. All of them have that:
listen_on_broadcast_address is set to true however and it looks like this on one of the node
broadcast_address: 10.1.169.163
listen_on_broadcast_address: true
broadcast_rpc_address: 0.0.0.0
@Felipe_Cardeneti_Mendes: that’s the problem then. You need to set broadcast_rpc_address to the actual node IP, not 0.0.0.0
@Sabin: okay let me try 
Thanks @Felipe_Cardeneti_Mendes it worked. I do not understand why RPC played such a role there? Is there any document that I can read to improve my knowledge upon this?
@Felipe_Cardeneti_Mendes: this is the IP broadcasted to drivers when they connect to the node. if its 0.0.0.0 the drivers dont know how to route traffic to it