Systemwalker Service Quality Coordinator User's Guide (Console Edition)

4.2.1 Planning

This section describes the reports used in planning by category.

The results of forecasts and simulations shown in the reports are calculated according to Fujitsu's own methods, based on past performance. These are offered as guides for the computer environment you are working with, but are not guaranteed to work in actual environments. Be aware that actual conditions may yield different results.

4.2.1.1 Virtual Aggregate

4.2.1.1.1 P2V (Physical to virtual)

Reports	Analysis target and condition settings	Display item	Description
Server Distribution by rsc. usage cond. (Summary)	G Specify System Group	Server Distribution - By CPU Usage Rate Server Distribution - By Memory Usage Rate Server Distribution - By Disk I/O Count Server Distribution - By Disk Throughput Server Distribution - By Count for data sent/received over network Server Distribution - By Network Throughput Hardware Resources information	Check the status of resources to see whether the physical server's resources are being used effectively across the entire system. If a resource was changed, for example because of memory being added within a specified period, only the post-change information for this server will be displayed. When considering virtual aggregation The benefits of aggregation are mostly achieved when there are many servers that have low usage. Note that resource usage is high when you choose to virtually aggregate servers that have high usage. Point The CPU performance (MHz) displayed is the performance value of one CPU core. The loaded memory displayed is the memory capacity that is available for actual use.
List of rsc. usage cond. (Detail)	G Specify System Group	CPU (usage rate/usage) Memory (usage rate/usage) Disk (I/O count/throughput) Network (Count for data sent/received over network/throughput) Hardware Resources information	Check the usage status of physical resources and the capacity of hardware resources (CPU and memory) across the entire system. If a resource was changed, for example because of memory being added within a specified period, only the period after the change will be displayed for that server. If this was saved in Excel(R) format, the information for all resources is displayed in a list on the [ALL] sheet. Point The CPU performance (MHz) displayed is the performance value of one CPU core. The loaded memory displayed is the memory capacity that is available for actual use.
P2V simulation	- Specifications for displays that include P2V simulations	CPU Usage Memory Usage Disk I/O Count Disk Throughput Count for data sent/received over network Network Throughput	The resource usage status can be simulated for a case where the selected servers are aggregated on one virtual environment. If a value becomes too great throughout a period of time or there is a bias depending on the time of day or day of the week, consider the combinations of the servers to be aggregated. If a tolerated threshold value was set, this checks that the value is not greatly exceeded. Not available with scheduled reports.

Note

The above report types are not displayed for servers being monitored by Agents from Systemwalker Service Quality Coordinator V13.5.0 or earlier or by an Agent for agentless monitoring.

4.2.1.2 Effective Use of Resources

4.2.1.2.1 VMware virtual machine relocation

Note

• For VMware ESX (ssh connection), there are items that cannot be collected, so reports from this category cannot be used.

• When performing VMware Virtual machine relocation simulations, register only one virtual host on the vCenter within a single system group.

Reports	Analysis target and condition settings	Display item	Description
VMware rsc. usage cond. (List of host)	G Specify System Group	CPU (usage rate/usage) Memory (usage rate/usage) Disk (I/O count/throughput) Network (Count for data sent/received over network/throughput) Hardware Resources information	Track resource use on the VMware virtual hosts registered in the system group. Virtual hosts with available resources will be detected as virtual hosts that might be aggregated. The overall CPU frequency (MHz) displayed is the total of all CPU cores fitted.
VMware rsc. usage cond. (Virtual machine stack)	H Specify System Group and Host	CPU Usage Memory Usage Disk I/O Count Disk Throughput Count for data sent/received over network Network Throughput	Check the virtual machines located in the virtual host and their resource usage on the cumulative graph for the virtual hosts considered as migration sources in "VMware rsc. usage cond. (List of host)". Check the resource usage in each virtual machine in the migration source, and use this information to work out which virtual machine should go to which destination candidate host.
VMware virtual machine relocation simulation	- Specifications for displays that include simulations of VMware Virtual machine relocation	CPU Usage Memory Usage Disk I/O Count Disk Throughput Count for data sent/received over network Network Throughput	Simulate the resource usage if virtual machines considered as migration candidates are relocated to virtual hosts considered as migration targets. Check results on graphs by time period, by day of the week, and by time series. Not available with scheduled reports.

4.2.1.2.2 VMware resource allocation optimization

Note

For VMware ESX (ssh connection), there are items that cannot be collected, so reports from this category cannot be used.

Reports	Analysis target and condition settings	Display item	Description
VMware rsc. usage cond. (List of virtual machine)	H Specify System Group and Host	CPU (usage rate/usage) Memory (usage rate/usage)	The resource usage status of the virtual machines on the selected virtual host are displayed in a list. Sort by CPU usage rate and memory usage rates to detect virtual machines with available resources and reduce the allocation of resources to virtual machines with low usage rates. CPU usage rate is calculated with each physical CPU having a value of 100%. This means that values over 100% may be displayed.

4.2.1.2.3 VMware tuning guidance

Note

For VMware ESX (ssh connection) and VMware ESX 3.5, there are items that cannot be collected, so reports from this category cannot be used.

Reports	Analysis target and condition settings	Display item	Description
VMware CPU tuning guidance	H Specify System Group and Host	Tuning guidance	Check if there is a problem with the CPU usage status of the host or virtual machine. Refer to the tuning guidance when displayed.
		Virtual Host, CPU Usage Rate	If the CPU usage rate of the host is constantly above 90% and the CPU wait rate on the virtual machine is constantly above 20% CPU waits are occurring due to overloading of the CPU. Consider increasing physical CPUs. If the CPU usage rate of the host is constantly above 90% and the CPU wait rate on one of the virtual machines is constantly above 90% There is a bias in usage. Review the allocation of virtual CPUs in the virtual machines. If the CPU usage rates on machines other than the host (0) are constantly above 90% and the CPU wait rate on the virtual machine is constantly above 20% CPUs are being overloaded, but CPU waits are not occurring. If batch processes are slow, consider switching to CPUs with higher CPU performance. Warning level: CPU usage rate of virtual host >= 90%
		Virtual Machine, CPU Usage Rate Virtual Machine, Average CPU Usage Rate Virtual Machine, CPU Wait Rate	CPU usage rate is calculated with each physical CPU having a value of 100%. This means that values over 100% may be displayed. If there is no problem with the CPU usage rate of the host, but the CPU usage rate (per CPU) of the virtual machines is constantly above 90% Review the allocation of virtual CPUs in the virtual machines, (for example, by allocating virtual CPUs that are not being used on other virtual machines). Also, check that there are no problems with specific applications running on the virtual machines. If there are no problems, move the virtual machine to a different host, or migrate to a different virtual machine. If there is no problem with the CPU usage rate of the virtual machine, but the CPU wait rate is constantly above 20% There may have been an over-allocation of virtual CPUs. Review the overcommit rate. Warning level: Virtual machine Average CPU usage rate (per CPU) >= 90% Virtual machine CPU wait rate >= 20%
VMware Memory (Host) tuning guidance	H Specify System Group and Host	Tuning guidance	Check if there is a problem with the memory usage status of the virtual host. Refer to the tuning guidance when displayed.
		Virtual Host, Swap Occurrence Status	When swap is occurring There is insufficient memory. Consider expanding memory. Warning level: Swap occurrence status > 0 MB
		Virtual Host, Memory Usage Status	When swapping volume is greater than 0 Swapping may have occurred previously. Check whether memory insufficiency has occurred at times other than when creating and starting the virtual machine. Warning level: Swapping volume > 0 MB/sec When the Memory compression volume is greater than 0 This may be due to insufficient physical memory. Consider expanding memory if there are problems with performance. Warning level: Memory compression volume > 0 MB/sec When balloon occurrence is greater than 0 This may be due to insufficient physical memory. Consider expanding memory if there are problems with performance. Warning level: Balloon occurrence > 0 MB/sec
		Virtual Host, Memory Usage	When the available memory is less than 6% of the total Indications of future memory insufficiency such as ballooning, memory compression, and swapping, may have already occurred. Consider expanding memory to prevent memory insufficiency occurring. Warning level: Free memory rate < 6%
VMware memory (Virtual machine) tuning guidance	H Specify System Group and Host	Tuning guidance	Check if there is a problem with the memory usage status of the virtual machine. Refer to the tuning guidance when displayed.
		Virtual Machine, Swap Occurrence Status(Reads/Writes)	When swap occurrence status (reads) or (writes) is greater than 0 Swap is occurring and there is insufficient memory. Consider expanding the memory. Warning level: Swap occurrence status (reads) > 0 MB Swap occurrence status (writes) > 0 MB
		Virtual Machine, Swap Wait Rate	When swap wait rate is greater than 0 CPU waits are occurring on the virtual machine and there is insufficient memory. Review the allocation of memory on the virtual machines. Warning level: Swap wait rate > 0%
		Virtual Machine, Swap Amount	When swap usage volume is greater than 0 Swapping may have occurred previously. Check whether memory insufficiency has occurred at times other than when creating and starting the virtual machine. Warning level: Swap usage volume > 0 MB/sec
		Virtual Machine, Memory Compression Amount	When the memory compression status is greater than 0 This may be due to insufficient physical memory. Consider expanding the memory if there are problems with performance. Warning level: Memory compression status > 0 MB/sec
		Virtual Machine, Memory Extension Amount	When the memory extension status is greater than 0 This may be due to insufficient physical memory. Consider expanding the memory if there are problems with performance. Warning level: Memory extension status > 0 MB/sec
		Virtual Machine, Balloon Occurrence Amount	When the balloon usage volume is greater than 0 This may be due to insufficient physical memory. Consider expanding the memory if there are problems with performance. Warning level: Balloon usage volume > 0 MB/sec
VMware Physical Disk	H Specify System Group and Host	Virtual Host, Physical Disk Throughput(Reads/Writes) Virtual Host, Physical Disk I/O Count	Track disk throughput and the I/O count on the virtual host. When the CM (Controller Module) exceeds the maximum performance of each disk device Consider changing the deployment of RAID groups with high loads to distribute the load. When there are RAID groups that exceed the maximum performance of the disk devices Consider adding RAID groups to distribute the load. If RAID5 is being used, consider changing to RAID0+1 or RAID1.
		Virtual Host, Physical Disk Driver Access Performance(Reads/Writes)	Check the disk driver access performance of the virtual host. When disk access performance is constantly 4ms or more When multiple machines are fitted in one datastore, reduce the number of virtual machines fitted in each single datastore (to distribute the load). Review the RAID configuration of the RAID group specified in the datastore. (Increase disks and review RAID level). Check the makeup of the virtual disks within the same RAID group (datastore). Warning level: Disk access performance (kernel) (reads) >=4 ms *constantly Disk access performance (kernel) (writes) >=4 ms *constantly
		Virtual Host, Physical Disk Kernel Access Performance(Reads/Writes)	Check the disk kernel access performance of the virtual host. When disk access performance is constantly 20ms or more When multiple machines are fitted in one datastore, reduce the number of virtual machines fitted in each single datastore (to distribute the load). Review the RAID configuration of the RAID group specified in the datastore. (Increase disks and review RAID level). Check the makeup of the virtual disks within the same RAID group (datastore). Warning level: Disk access performance (driver) (reads) >= 20 ms *constantly Disk access performance (driver) (writes) >= 20 ms *constantly
VMware Virtual Disk	H Specify System Group and Host	Virtual Disk Throughput Virtual Disk I/O Count	Track disk throughput and the I/O count on the virtual machine. When the CM (Controller Module) exceeds the maximum performance of each disk device Consider changing the deployment of RAID groups with high loads to distribute the load. When there are RAID groups that exceed the maximum performance of the disk devices Consider adding RAID groups to distribute the load. If RAID5 is being used, consider changing to RAID0+1 or RAID1.
		Virtual Disk, Access Performance(Reads/Writes)	Check the disk access performance of the virtual machine.
VMware Physical NIC	H Specify System Group and Host	Physical NIC, Network Usage Rate(Transmission/Reception)	Track the network usage rate of the physical NIC. When the network usage rate (transmission) is exceeding a threshold The load on a physical NIC is high. Try reviewing the makeup of the physical network and dividing segments. (Refer to the VMware Virtual NIC and see which virtual machines are tying up the bandwidth of the physical NIC). When the network usage rate (reception) is exceeding a threshold The load on a physical NIC is high. Try reviewing the makeup of the physical network and dividing segments. (Refer to the VMware Virtual NIC and see which virtual machines are tying up the bandwidth of the physical NIC). Add physical NICs to the vSwitch connected to the physical NIC where load is high. (Teaming). Warning level: Network usage rate >= threshold *constantly Circuit type: Threshold When a dedicated circuit is being used: 80% When a switching hub is being used: 60% When a repeater hub is being used: 30%
		Physical NIC, Packet Drop Rate(Transmission/Reception)	Track the packet drop ratio of the physical NIC. When packet drop rate (reception) is greater than 0 Check the CPU usage rate of the virtual machine and add virtual CPUs if the CPU usage rate is high. Also review the configuration of the virtual machine's network. When packet drop rate (transmission) is greater than 0 Check the network usage rate of the physical NIC, and move virtual machines connected to physical NIC with high values to vSwitches connected to different physical NICs. Alternatively, add physical NICs to the vSwitch connected to the physical NIC where packet drop is occurring (Teaming). Warning level: Packet drop rate (transmission) > 0% Packet drop rate (reception) > 0%
VMware Virtual NIC	H Specify System Group and Host	Virtual NIC, Network Usage (Transmission/Reception)	Check the network usage of virtual NICs.
		Virtual NIC, Packet Drop Rate (Transmission/Reception)	Check the packet drop ratio of the virtual NIC. Note The following are not displayed for VMware ESX 3.5, VMware ESX 4.0, or VMware ESX/ESXi 4.1: Virtual NIC packet drop rate (transmission) Virtual NIC packet drop rate (reception)

4.2.1.3 Demand Forecast

4.2.1.3.1 VMware resource pool

Reports	Analysis target and condition settings	Display item	Description
Resource pool (CPU) (demand forecast)	H Specify System Group, Host, and Resource ID	Resource Pool(CPU)(demand forecast)	Perform regression analysis based on past CPU usage rates in the resource pool managed by VMware to anticipate the resources that will be used.
Resource pool (Memory) (demand forecas	H Specify System Group, Host, and Resource ID	Resource Pool(Memory)(demand forecast)	Perform regression analysis based on past memory usage rates in the resource pool managed by VMware to anticipate the resources that will be used.

4.2.1.3.2 ServerView Resource Orchestrator resource pool

Reports	Analysis target and condition settings	Display item	Description
VM pool (CPU) (demand forecast)	H Specify System Group, Host, and Resource ID	VM pool (CPU) (demand forecast)	Perform regression analysis based on past CPU usage rates in the VM pool managed by ServerView Resource Orchestrator to anticipate the resources that will be used.
VM pool (Memory) (demand forecast)	H Specify System Group, Host, and Resource ID	VM pool (Memory) (demand forecast)	Perform regression analysis based on past memory usage rates in the VM pool managed by ServerView Resource Orchestrator to anticipate the resources that will be used.
Storage pool (demand forecast)	H Specify System Group, Host, and Resource ID	Storage pool (demand forecast)	Perform regression analysis based on past usage rates in the storage pool managed by ServerView Resource Orchestrator to anticipate the resources that will be used.
Network pool (demand forecast)	H Specify System Group, Host, and Resource ID	Network pool (demand forecast)	Perform regression analysis based on past usage rates in the network pool managed by ServerView Resource Orchestrator to anticipate the resources that will be used.
Server pool (demand forecast)	H Specify System Group, Host, and Resource ID	Server pool (demand forecast)	Perform regression analysis based on past usage rates in the server pool managed by ServerView Resource Orchestrator to anticipate the resources that will be used.
Address pool (demand forecast)	H Specify System Group, Host, and Resource ID	Address pool (demand forecast)	Perform regression analysis based on past usage rates in the address pool managed by ServerView Resource Orchestrator to anticipate the resources that will be used.

4.2.1.4 Increment Simulation

4.2.1.4.1 Response simulation

Reports	Analysis target and condition settings	Display item	Description
Request count (future prediction)	H Specifications for displays that include request count future predictions	Request count (future prediction)	Perform regression analysis based on past performance information for Web transactions to anticipate future request volume. Track the potential future increases in requests compared to the current number of requests.
Response simulation (Request count)	- Specifications for displays that include response simulations (requests or adding servers)	Response Time(Request Increase)	The correlation coefficient between Web transaction performance information and the performance information of the operating systems of the servers making up the system is analyzed based on past performance, and a simulation can be performed of the trend in daily response times when the number of responses increases (expressed as a multiple of the current number of requests). Determine whether improvements are required in response time. Not available with scheduled reports. Point Response simulation is a simulation of response times, achieved by analyzing the relationship between the number of service requests in the past and the performance information (OS) of the various servers. The precision of the simulation can be increased by excluding time periods when batch processes are performed (after hours, for example) that have no bearing on actual request processes. Confirm the precision of the simulation by looking at the '(High)', '(Medium)', and '(Low)' reliability displayed after the response time in the table. '(High)' indicates that the simulation was performed with little interference and the precision of the simulation is therefore high. 'N/A' is displayed in the table if the simulation could not be performed because no performance information was found or no correlation was achieved between the number of requests and the performance information of the OS. The reliability can be improved by increasing the period in Analysis Conditions, and taking actions to decrease interference by excluding periods when batch work is performed, such as after hours. Refer to "Operating the Analysis/Planning Window" for information on how to configure analysis conditions.
Response simulation (Adding servers)	- Specifications for displays that include response simulations (requests or adding servers)	Request Times(Adding servers)	The correlation coefficient between Web transaction performance information and the performance information of the operating systems of the servers making up the system is analyzed based on past performance, and a simulation can be performed of the trend in response times when a server is added. Determine where to add servers (in the Web layer, application layer, or database layer) and how many to add to improve response times. Not available with scheduled reports. Point Response simulation performs the simulation by analyzing the correlation coefficient between the number of requests and the performance information of the OS, so the precision of the simulated response times improves when there is sufficient correlation. The reliability of the above correlation coefficient is indicated by the three levels displayed in the table; '(High)', '(Medium)', and '(Low)'. '(High)' indicates that the response time has been simulated to a relatively high degree of precision due to sufficient correlations being achieved. '(Medium)' and '(Low)' indicate that the precision of the response time has been reduced due to insufficient correlations being achieved. If there is insufficient performance information for calculation, there may be interference from processes not directly related to requests, such as after-hours batch processing. If this is the case, reliability may be improved by changing the Analysis Conditions settings as shown below: - Make the analysis period longer. - Set the times other than service time as times when services are not running, such as nights or days off. Refer to "4.3 Operating the Analysis/Planning Window" for information on how to configure analysis conditions. Note If there is no performance information or there are many processes running that are not directly related to the number of requests, a correlation may not be established between the number of requests and the performance information of the OS, making it impossible to generate an analysis, in which case 'N/A' is displayed instead of a value. If this is the case, analysis may be made possible by changing the Analysis Conditions settings as shown below: - Make the analysis period longer. - Set the times other than service time as times when services are not running, such as nights or days off. Refer to "4.3 Operating the Analysis/Planning Window" for information on how to configure analysis conditions.

4.2.1.5 Generic Report

4.2.1.5.1 Generic report

Reports	Analysis target and condition settings	Display item	Description
Future forecast display	H Specifications for displays that include future predictions	-	Displays specified field values as regression line graphs.