You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I have set "--wsrep_node_address", "ist.recv_addr", "--wsrep-sst-receive-address" with DNS name of the node. I have set bind address to 0.0.0.0. And I have proper host aliases in each node to resolve the DNS name to public IP (except for itself, for self the DNS will resolve to private IP).
Cluster starts fine and works for a while. When one nodes goes down, the other nodes start hitting the private IP of the died node. I.e. the wsrep is somehow sending private IP to other nodes. When that node comes back up, then it doesn't connect, it goes in crash loop saying duplicate uuid.
Is there a config to say not to send the listen_address to other nodes?
I have set "--wsrep_node_address", "ist.recv_addr", "--wsrep-sst-receive-address" with DNS name of the node. I have set bind address to 0.0.0.0. And I have proper host aliases in each node to resolve the DNS name to public IP (except for itself, for self the DNS will resolve to private IP).
Cluster starts fine and works for a while. When one nodes goes down, the other nodes start hitting the private IP of the died node. I.e. the wsrep is somehow sending private IP to other nodes. When that node comes back up, then it doesn't connect, it goes in crash loop saying duplicate uuid.
Is there a config to say not to send the listen_address to other nodes?
why there is empty string here for the remote_address - https://github.com/codership/galera/blob/4.x/gcomm/src/gmcast.cpp#L524 ?
The text was updated successfully, but these errors were encountered: