[jboss-jira] [JBoss JIRA] (WFLY-11750) Allow cluster to use DNS addresses instead of IP addresses

[ https://issues.jboss.org/browse/WFLY-11750?page=com.atlassian.jira.plugin... ] Tomasz Adamski edited comment on WFLY-11750 at 2/25/19 3:32 PM: ---------------------------------------------------------------- [~pferraro] But if I understand correctly HA features and correct transaction recovery are independent. Let's suppose we are on bare-metal and have a cluster of n nodes. At some point, one of the nodes is invoked by the client, takes part in the transaction with two-phase commit and after prepare stage the node fails. As a result, the client has a record in its permanent object store and tries to finish the transaction. As the record contains the IP address of the failed node the client will wait till the server is up again to finish the transaction. That node will have the new cluster identity. My understanding is that it has no effect on transaction recovery which will work fine. On the other hand, OpenShift does not guarantee the IP node identity - only DNS identity - so IMO the cluster will work fine but recovery won't. But the key thing here IMO is, the way remoting gets node information. When clustering is involved it gets raw IP addresses which won't work fine on OpenShift and my thought was that the simplest solution would be to use DNS addresses as base for cluster operation. was (Author: tomekadamski): [~pferraro] But if I understand correctly HA features and correct transaction recovery are independent. Let's suppose we are on bare-metal and have a cluster of n nodes. At some point, one of the nodes (let's call it node A) is invoked by the client, takes part in the transaction with two-phase commit and after prepare stage the node fails. As a result, the client has a record in its permanent object store and tries to finish the transaction. As the record contains the IP address of the failed node the client will wait till the server is up again to finish the transaction. That node will have the new cluster identity. My understanding is that it has no effect on transaction recovery which will work fine. On the other hand, OpenShift does not guarantee the IP node identity - only DNS identity - so IMO the cluster will work fine but recovery won't. But the key thing here IMO is, the way remoting gets node information. When clustering is involved it gets raw IP addresses which won't work fine on OpenShift and my thought was that the simplest solution would be to use DNS addresses as base for cluster operation. -- This message was sent by Atlassian Jira (v7.12.1#712002)