On 05/18/09 09:51, Gino wrote: > hello, > no way :( > > > node1# clnode remove node2 > clnode: (C656493) Node is in cluster mode. > node1# scconf -r -h node=node2 > scconf: Failed to remove node (node2) - quorum could be compromised. > scconf: All two-node clusters must have at least one shared quorum device. > > > node2# /usr/cluster/bin/cluster status > cluster: (C152734) This node is not in cluster mode. >
Hi Gino, Looks like node1 is in cluster mode and node2 is in non-cluster mode. Which node do you want to remove? The node to be removed must be in non-cluster mode. So if you want remove node1, you should bring it to non-cluster mode first. Here is the brief procedure: - While the nodes are still in cluster mode, do "claccess allow-all". - Also make sure no SVM device groups has reference to the node to be removed. If the node is still used in SVM DGs, use the Solaris commands (metaset etc) to remove the node from those DGs. - Bring the node to be removed to non-cluster mode. Then run "clnode remove -F" on that node. - If above "clnode remove -F" failed for some reasons, on any active node, run "clnode clear -F <node-to-remove>". If you want to remove all the nodes, do above procedure on all the nodes (except the last step since there will be no active node). When you try to add the nodes back to cluster again, use "scinstall" and choose the option of adding a node into existing cluster. We recently just found that centralized installation does not work after removing all the nodes, but adding node one by one should work. If you hit any problems let me know. Thanks, - Lucia > gino >
