During the S6800T (version: V100R001C01SPC013) kicked off on site, it was connected with the midrange computer via the Brocade fiber channel switch, and the corresponding multiple paths software was the DMP of Symantec. After the configuration has been complete, the host can identify the storage, but it’s not steady, sometimes it can identify 3 paths (all were “enable”), and sometimes it identified there are 4 paths (one was available, the other three were “disable”).
Login in the Brocade fiber channel switch via the web, the fiber channel switch didn’t report error. However, the number of the link error packages connected to the HP host increased gradually, and there were few (less than three) error packages at the port connected to the storage.
Checked the mode of the S6800T storage host, we found the default mode when it was produced was “loop”, and then we modified it as to “Fabric” mode, finally the system recovered to be normal.
Because there were many link error packages at the port of the switch which connected to the host, we doubted it may be the link fault of the host, checked the hot link again, but it was normal.
We checked the link between the storage and the switch, and found it was normal too, plugged the fiber into another port of the storage, the host’s status didn’t change, and during the replacing process the error packages at the switch port increased swiftly(to millions), analyzed the outcome and we found it was normal during the link changed.
Login in the fiber channel switch via the CLI, we found the status of the storage port was changed from “sync” and “no sync”, the matching of the port mode was “Loop”. We considered it was the wrong configuration of the storage port mode.
The default mode of the S6800T storage port is “Loop”, it may appear abnormal configuration when connecting with some fiber channel switch, so we need modify it as to “Fabric” mode manually.
The Web management mode of the fiber channel switch is not sensitive to the status of the physical port, therefore, when there is system abnormity, we suggest login in the CLI mode, use the command and display the port status immediately, and then we can find the abnormal status.