The abnormity of the HBA causes the LUN’s working controller hasn’t switched back to the owning controller
Publication Date: 2012-11-05Views: 73Downloads: 0
Product and version:
S5500T V100R001 V100R002
S5600T V100R001 V100R002
S5800T V100R001 V100R002
S6800T V100R001 V100R002
The operating system of the application server: Windows Server 2003 X86
HBA card: QLogic QLE2562
Driver file: ql2300.sys
Driver version: v184.108.40.206
The storage device map the LUN to the host, the LUN’s working controller and owning controller are both the controller A. When plugging out the fiber, the LUN’S working controller switches to the peer end controller, i.e. the controller B; and plug in the fiber, this LUN’s working controller hasn’t switched back to the controller A.
Influence: if the working controller and the owning controller are not identical, it will reduce the performance of the host dispensing the IO.
Plug back the fiber, scan the disk manually (right click the “scan and check the hardware modification” in the disk driver), make sure to create the physical disk successfully. As the figure 1 displayed:
Figure 1 start the WMI service
Plug in the fiber again, the system hasn’t created the physical LUN, which causes the multiple paths can’t execute the failback. The creation of the physical LUN is responsible by the HBA, then we can conclude that there is collaboration problem between the HBA card and the system PNP manager (Plug-and-Play manager).
Plug back the fiber and scan the disk manually, this can trigger the HBA card to make the command to scan the LUN for the storage device. When the storage device receiving the command, it will report the disk to the host, which confirms the reporting process successfully.