I have a few questions about the IPaddr2 resource agent used in a clone cluster. They probably relate to the iptables CLUSTERIP feature, too. But, I hope someone can answer them here.
What happens if one of the node in the clone cluster goes down for an extended time? It seems to me that the iptables CLUSTERIP will continue to work, but messages which hash to the absent node will simply be dropped by all the other nodes. Is there any recovery for this? Seems to me that you’d need to reconfigure the cluster to be n-1 nodes so that iptables CLUSTERIP will hash correctly.
Is there a way to know which node an iptables CLUSTERIP will hash to? Say the hash mode is ‘sourceip’ and I have the source IP. Can I know which node would get messages from that source IP? I know that cluster node number and iptables CLUSTERIP node number can be different. I’d like to know which node so it can be ‘prepared’ for new traffic from a client on that IP. All the clients will all be configured with the clone cluster IP. The client traffic would be load-balanced over the cluster.
May you live in interesting times, may you come to the attention of important people and may all your wishes come true.
THIS COMMUNICATION MAY CONTAIN CONFIDENTIAL AND/OR OTHERWISE PROPRIETARY MATERIAL and is thus for use only by the intended recipient. If you received this in error, please contact the sender and delete the e-mail and its attachments from all computers.