Recovering LUNs of the failed node
You must actively track whether LUNs are online or offline in a MetroCluster configuration. If the cf.takeover.change_fsid option is set to on , and there is a disaster, all LUNs in the aggregates that were mirrored at the surviving site are offline. You can’t determine if they were online prior to the disaster unless you track their state.
If you have a MetroCluster configuration, you must actively track the state of LUNs (track whether they are online or offline) on the node at each site. If there is a failure to a MetroCluster configuration that qualifies as a disaster and the node at one site is inaccessible, all LUNs in the aggregates that were mirrored at the surviving site are offline. There is no way to distinguish the LUNs that were offline before the disaster from the LUNs that were online before the disaster unless you have been tracking their status.
When you recover access to the failed node’s LUNs, it is important to bring back online only the LUNs that were online before the disaster. To avoid igroup mapping conflicts, do not bring a LUN online if it was offline before the disaster. For example, suppose you have two LUNs with IDs of 5 mapped to the same igroup, but one of these LUNs was offline before the disaster. If you bring the previously offline LUN online first, you cannot bring the second LUN online because you cannot have two LUNs with the same ID mapped to the same host.
- Identify the LUNs that were online before the disaster occurred.
Make sure that the LUNs are mapped to an igroup that contains the hosts attached to the surviving node.
For more information about mapping LUNs to igroups, see your Data ONTAP SAN Administration Guide for 7-Mode.
On the surviving node, enter the following command:lun online lun-path ...
lun-path is the path to the LUN you want to bring online. You can specify more than one path to bring multiple LUNs online.lun online /vol/vol1/lun5 lun online /vol/vol1/lun3 /vol/vol1/lun4Note: After you bring LUNs back online, you might have to perform some application or host-side recovery procedures. For example, the File System Identifiers (FSIDs) are rewritten, which can cause the LUN disk signatures to change. For more information, see the documentation for your application and for your host operating system.