Data Backup and Recovery Discussions
Please verify following steps in create shared luns in a 2-nodes cluster environment, ignored some of details in options.
- igroup create igroup_node1 wwn_node1
- igroup create igroup_node2 wwn_node2
- vol create
- lun create shared_lun1
- lun map shared_lun1 igroup_node1
- lun map shared_lun1 igroup_node2
If I want to create 2nd shared lun then add the following:- lun creaet shared_lun2
- lun map shared_lun2 igroup_node1 # map to the same igroup
- lun map shared_lun2 igroup_node2
then on two nodes, disk scan to discover the new created shanred lun.
Are above steps seem correct?
I prefer to use single initiator group for all nodes in cluster, that ensure that every LUN is always mapped to all nodes. But your way will work as wll, and is actually used by SnapDrive.
Yes they use separate initiators for SDW. Also if you want to map the same LUN on one node using iSCSI and other using FCP, then you would need separate igroups as above.
Got your point.
A follow-up questions: Will you create a igroup for each one of shared LUN's and mapped to all nodes, or create a single igroup for all lun's and mapped to all nodes?
Single group for all LUNs and all nodes in a cluster.
Thanks for your time and patience.
One more question:
What benifits I am getting by creating a single igroup for them all, instead of the way I initially wrote?
It ensures that any change in configuration is always applied to the whole cluster. When you add LUN, it is ensured that all nodes see it. When you a node, it is ensured it sees all LUNs.
Whether this is beneficial or not, it is up to you after all ...