The replacement data directory could not be started

[Wwangzhengyang]

I changed the data_directory from /data03/redpanda to /data04/redpanda and redpanda couldn't start, so I used yum remove redpanda,I yum install redpanda and set data_directory to /data04. At this time, there was no file under /data04/redpanda, and I re-executed rpk redpanda mode production. rpk redpanda tune all, rpk redpanda config bootstrap, etc.
Can't start now.node01, node02,node3 long wait until time out.I can see that each node sets up LISTEN on port 33145, but there is no communication between them. Not kafka_api set up

The following log messages on node 233:
cluster - cluster_discovery.cc:143 - Requesting node ID for node UUID {xxxxxx-4d18-b3b6-00005e4bd68d} from {host: 232, port: 33145}
cluster - cluster_discovery.cc:143 - Requesting node ID for node UUID {xxxxxx-4d18-b3b6-00005e4bd68d} from {host: 223, port: 33145}
The following log messages on node 232:
cluster - cluster_discovery.cc:143 - Requesting node ID for node UUID {xxxxxx-4d18-b3b6-00005e4bd68d} from {host: 223, port: 33145}
cluster - cluster_discovery.cc:143 - Requesting node ID for node UUID {xxxxxx-4d18-b3b6-00005e4bd68d} from {host: 233, port: 33145}

other node 223:
Failure during startup: std::runtime_error (cluster_discovery.cc:469 - Cannot bootstrap a cluster due to seed servers configuraton mismatch, check the errors above in the log for details. Brokers: {{id: 0, kafka_advertised_listeners: {{:{host: 233, port: 9093}}}, rpc_address: {host: 233, port: 33145}, rack: {nullopt}, properties: {cores 40, mem_available 116, disk_available 7392}, membership_state: active}, {id: 1, kafka_advertised_listeners: {{:{host: 232, port: 9093}}}, rpc_address: {host: 232, port: 33145}, rack: {nullopt}, properties: {cores 40, mem_available 116, disk_available 7392}, membership_state: active}, {id: 2, kafka_advertised_listeners: {{:{host: 223, port: 9093}}}, rpc_address: {host: 223, port: 33145}, rack: {nullopt}, properties: {cores 40, mem_available 118, disk_available 7392}, membership_state: active}})

what should i do?

[emaxerrno]

There is some left over state on disk somewhere. @jcsp maybe node id persisted ?

[emaxerrno]

a cluster due to seed servers configuraton mismatch, check the errors above in the log for details.

Any more logs related to bootstrap server list

[jcsp]

There should be ERROR logs earlier relating to the "Cannot bootstrap a cluster..." message.