Hello, everyone!
The number of paths displayed after the upadm show path command is executed to map LUNs to a Linux host using FC is incorrect.
Problem Description
The number of paths displayed after the upadm show path command is executed to map LUNs to a Linux host using FC is incorrect.
System: CentOS6.8 UltraPath version: 21.01.027
Two Fibre Channel links between the host and the storage device are mapped to three LUNs. The output of the upadm show path command is as follows.
The number of paths between controllers A and B is different.
Process
1. Run the upadmin show path command. The query result is as follows:
In addition to two Fibre Channel links, there is an iSCSI link between the storage device and the host.
2. Run the iscsiadm -m session command to check whether an iSCSI session exists.
3. The upadm command output shows the path of each LUN under each controller.
Currently, three LUNs are mapped to one Fibre Channel link to controller A and controller B respectively. The upadm corresponds to L001 002 003, and L000 is the device identified by the iSCSI link.
4. Confirm that the iSCSI link is not used. Ask engineer to cancel the iSCSI link and scan for disks again. Run the upadm show path command to check whether the link is normal. The LUNs of controllers A and B are identified as the same.
Root Cause
In addition to the FC link, there is an iSCSI link on the live network. The iSCSI link has been registered but not added to the host. Therefore, the device identified by the host should be 16kLUNs. For details about how to shield 16kLUNs, see the case at http://support.huawei.com.
Solution
Release the unused iSCSI connection and scan for disks. The query result is normal.
This is my solution, how about yours? Go ahead and share it with us!
