Siemens SM2 monitoring system manual VM2000 operation on Risc architecture VM2000/SR

Special applications (VM2000) Notes on SM2 operation

In the case of the CHANNEL-IO monitoring program, the values supplied (e.g. the I/O operations per channel) always refer to the respective guest system. Correlation of these values on channel utilization is not possible, due to the above-mentioned restriction.

For CPUs with DCS, the SERVICETIME monitoring program can only be started by one guest system. Any attempt to activate this function from a second guest system is accepted without an error message, but no DCS-specific monitored data will be supplied.

The VM2000 command /SHOW-VM-RESOURCES INFORMATION=*STD/*ALL can be used to check whether SERVICETIME monitoring is already active in a guest system. If it is, message VMS2035 is output. In the case of SHARED-DEVICES under VM2000 (TYPE=*STD), DCS-specific monitoring values are generally not supplied by any guest system (see also the “VM2000” manual [9]).

VM2000 operation on RISC architecture (VM2000/SR)

The real CPUs are assigned exclusively to the guest systems under VM2000/SR.

The CPU MAX and CPU PLAN values output in the VM2000 report therefore correspond with the relationship of the assigned CPUs to the total number of physical CPUs in the VM2000/SR system (NUMBER OF CPU’S / NUMBER OF PHYS. CPU’S).

Due to the exclusive assignment of CPUs, the CPU MEAS value corresponds to the utili- zation of the guest system, apart from the active times of the hypervisor and the asynchronous firmware activities.

The active times of the hypervisor are contained in the HYPERVISOR ACTIVE value, while the times for asynchronous firmware activities are contained in the CPU MEAS value. The IDLE value corresponds to the relationship of unused CPUs to physically available CPUs.

The HYPERVISOR ACTIVE value corresponds to the average value of active times of the hypervisor over all CPUs.

In contrast to the VM2000 report, the active times of the hypervisor in the CPU/ACTIVITY report are contained in the respective interrupted processor states (TU, TPR, SIH or IDLE). The asynchronous firmware activities are likewise contained in the interrupted processor states (TU, TPR, SIH or IDLE).

Asynchronous firmware or hypervisor activities may have been caused by a remote guest system (e.g. when processing asynchronous I/O events of shared disks).

In the case of SM2 monitoring values formed by the difference between two time values, there is no direct dilation by other guest systems under VM2000/SR. However, minimal dilation by the hypervisor activities and asynchronous firmware activities is unavoidable.

350 U3585-J-Z125-8-76