A fiber channel switch (hereinafter, FC-SW) connected to a plurality of communications devices is provided with a buffer for temporarily storing data received from servers via a first communications port, a trace data storage memory for storing trace data, and a microprocessor (MP) for sending data accumulated in the buffer to a storage control system. The MP stores the information about the data about the received data into the trace data storage memory as trace data, issues a write command to the storage control system if it is detected that trace data sending condition is satisfied, and reads out the trace data stored in the trace data storage memory and sends it to the storage control system.
A method, system, and computer-usable medium for supporting debugging of host channel adapters in a logical partitioning environment. In a preferred embodiment of the present invention, a hypervisor acquires control of a trace facility and sets trace parameters for the host channel adapter. In response to determining a trace event that matches said trace parameter has been triggered, the hypervisor retrieves trace information from a buffer. In response to determining the buffer does not include any more trace information, the hypervisor determines if modification of the trace parameters is required. If the modification of the trace parameters is required, the hypervisor alters the trace parameters in anticipation of another trace event.
A disk array device has: a CM separating part that, when abnormality occurs in a CM, separates the CM in which the abnormality has occurred from a storage system after write of CM operation information data to a hard disk is completed; and communication path reconfiguring parts that reconfigure communication paths between the CM in which the abnormality has occurred and DAs. When abnormality occurs in the CM, the communication paths between the CM and the DAs are reconfigured, and after the CM operation information data of the CM is written to the hard disk, the CM is separated. Therefore, the write processing is not interrupted halfway, so that the useful operation information data regarding the cause of the problem occurrence can be written to the hard disk having a sufficient storage area.
