Hitachi Command Suite 8 User Guide
Have a look at the manual Hitachi Command Suite 8 User Guide online for free. It’s possible to download the document as PDF or print. UserManuals.tech offer 913 Hitachi manuals and user’s guides for free. Share the user manual or guide on Facebook, Twitter or Google+.
Table 12-2 Volume metrics for USP V/VM and Hitachi USP (IdentifyPerformance Problems wizard)ItemMetricThresholdNotesCacheWrite Pending RateCaution: > 30%• Up to 30% is considered normal. • Frequent or sustained episodes of 40% deserve attention. • Frequent or sustained episodes to 50% require prompt remediation. • At 70%, inflow control is enabled which is disruptive. • High write pending is a sign of back end congestion.Side File Usage RateCaution: ≧ 10%• This is only applicable for TrueCopy Asynchronous. • Little, if any, should be reported under normal conditions. High side file usage is a sign of downstream congestion.Write Pending Rate + Side File Usage RateCaution: ≧ 70%• It is very important to look at both CWP and SF together since they share the same cache pool and cannot exceed 70%. • The closer CWP or SF, or CWP+SF is to 70%, the closer the storage system is to major performance problems. • When the write pending reaches 70%, the storage system starts to destage write data until the write pending usage reaches 30%, which will degrade performance.Channel ProcessorBusy RateCaution: ≧ 45%• 80% approximate maximum.Parity GroupBusy RateCaution: ≧ 50%• Parity Group utilization should be less than 50%.Read HitCaution: ≦ 25%• -Disk ProcessorBusy RateCaution: ≧ 40%• -Tip: If you want to view detailed performance information, start Performance Reporter. By doing so, you can perform an analysis per minute or a long- term trend analysis for more performance evaluation items. Related concepts • About analyzing storage system performance on page 403 Metrics of volumes of midrange storage systems (Identify Performance Problems wizard) For each item used to analyze performance displayed in the Identify Performance Problems wizard, there are guideline values for evaluation if volumes (basic volumes, DP volumes) of HUS 100, Hitachi AMS 2000, HitachiSMS, or Hitachi AMS/WMS are used. When analyzing DP volumes, all processors and caches that are related to DP volumes are targeted. As for parity groups, those to which DP pool volumes, Monitoring managed resources and resolving alerts411Hitachi Command Suite User Guide
composing a DP pool, belong are targeted. The target drives (physical disks) are all those that compose a DP pool. In the Identify Performance Problems wizard, the status is judged as Caution when items exceed the following thresholds. Table 12-3 Volume metrics for midrange storage (Identify Performance Problems wizard)ItemMetricThresholdNotesProcessorBusy RateCaution: ≧ 80%• Values over 66% are likely to be accompanied by response time increases due to contention. • Values over 50% will mean that the storage system will not be able to accommodate a controller failure without a reduction in workload capacity. • There are cases of resolving the problem by distributing the workload to other CPUs, or reducing the workload of such processes as TrueCopy, Copy-on- Write Snapshot, and ShadowImage.CacheWrite Pending Rate on Controller 1, Write Pending Rate on Controller 0Caution: > 25%• The maximum Write Pending level per Partition is 70%. • There are cases of resolving the problem by distributing the workload to other controllers, adding to the cache, using high-speed drives, adding drives to parity groups, adding parity groups, or switching to RAID 1+0.Parity GroupWrite HitCaution: < 100%• -DriveOperating Rate on Controller 1, Operating Rate on Controller 0Caution: ≧ 70%• Values over 66% will result in response deterioration due to contention.Tip: If you want to view detailed performance information, start Performance Reporter. By doing so, you can perform an analysis per minute or a long- term trend analysis for more performance evaluation items. Related concepts • About analyzing storage system performance on page 403 Creating periodic health check reports for storage systems performance analysis You can create storage system health check (performance) reports, and control scheduling, the data collection time period, and data interval settings. You can review the health check reports when completed, and later by reviewing health check history. 412Monitoring managed resources and resolving alertsHitachi Command Suite User Guide
Prerequisites • Install Tuning Manager and link it to HCS. • Register the Tuning Manager license. • Import the report definition file in Tuning Manager before analyzing performance with the Analytics tab (ensure that you delete the report definition file for the previous version before importing the latest report definition file). • To receive the results of Health Check execution by email, you must specify the SMTP server and edit the server properties file of the Device Manager server. • For more details on these tasks, see the Hitachi Command Suite Administrator Guide . Procedure 1. On the Analytics tab, select Storage Systems . 2. Select one or more storage systems, and click Create Schedule. 3. Specify the required items, and run the task. 4. After verifying that the task is complete, check for any performance problems. If there is a problem, check the Health Check report of the corresponding storage system. In accordance with the specified threshold values, the background of the displayed analysis result chart is colored for each threshold range. The Caution threshold range is displayed in yellow. The Warning threshold range is displayed in pink. 5. Verify the problem areas from the report information, and then analyze and identify the problems.Tip: Observe the following guidelines when analyzing the report information: • If you are using the UR Performance Analysis window of the Replication tab for Universal Replicator performance analysis, and a periodic refresh schedule has been configured for data collection, make sure that storage system health checks (performance reports) run at a different time to ensure the distribution of processing loads. • If you want to analyze the performance of storage systems immediately, select the target storage systems, and then click Run Health Check . You can view the execution results in the Task & Alerts tab. • If you want to delete a schedule, from Schedule, select the target schedule from the Health Check Schedules list, and then click Delete Schedule . • The Utilization Reports (Threshold Exceeded) chart of a health check report displays the maximum value among the monitored targets that exceeded the threshold value. For other report charts, an average value of the monitored targets is displayed.Monitoring managed resources and resolving alerts413Hitachi Command Suite User Guide
• If you want to view detailed performance information, start TuningManager from the Tools menu. By displaying Performance Reporter reports from Tuning Manager, you can analyzed based on a per-minute basis or on a long-term trend basis. For details on how to use Performance Reporter, see the Hitachi Command Suite Tuning Manager Software User Guide .Tip: The following categories are part of the health check reports, but are not displayed when using the Analytics tab. These health check report categories represent storage system total front-end and back-end metrics that can be used with the Performance Reporter metrics to get a better understanding of storage system performance: • Storage system total front-end IOPS: represents the total number of IOPS (inclusive reads and writes) issued to a storage system by its connected hosts. • Storage system total front-end MB/s: represents the total number in megabytes per second that are read from or written to a storagesystem by its connected hosts. • Storage system total back-end IOPS: represents the part of the storage system total front-end IOPS that was not served by the cachememory in the storage system. Related concepts • About analyzing storage system performance on page 403 Related tasks • Editing periodic health check schedule on page 414 • Exporting health check reports on page 415 • Viewing and deleting health check reports on page 415 Editing periodic health check schedule You can change a previously created storage system health check schedule. Procedure 1. From the Analytics tab, select Schedule. 2. Select the schedule to change, and click Edit Schedule. 3. Specify the required items, and then execute the task. Result Display the window showing the target schedule in detail, and confirm that the schedule has been changed. 414Monitoring managed resources and resolving alertsHitachi Command Suite User Guide
Tip: If you are using the UR Performance Analysis window of the Replication tab, and a periodic refresh schedule has been configured for data collection, make sure that storage system health checks (performance reports) run at a different time to ensure the distribution of processing loads. Related concepts • About analyzing storage system performance on page 403 Related tasks • Creating periodic health check reports for storage systems performance analysis on page 412 Exporting health check reports You can export health check reports to a PDF or CSV file for reporting purposes. Procedure 1. From the Analytics tab, select Report History . 2. From the Reports list, select the report to export, and click Export Reports . 3. Select the required items and click Submit. 4. Select the export location for the health check report. Result A CSV file or PDF file is saved in the specified location. Related concepts • About analyzing storage system performance on page 403 Related tasks • Viewing and deleting health check reports on page 415 Viewing and deleting health check reports Previously created health check reports can be reviewed to check for andresolve performance issues. Procedure 1. From the Analytics tab, select Report History . 2. From the Reports list, select the report to review, and click View Report of the Health Status column. 3. Review the displayed health check report. In accordance with the specified threshold values, the background of the displayed analysis result chart is colored for each threshold range. The Monitoring managed resources and resolving alerts415Hitachi Command Suite User Guide
Caution threshold range is displayed in yellow. The Warning threshold range is displayed in pink. Result Based on the health check report, resolve performance issues if necessary.Tip: To delete reports that are no longer needed, select the reports to be deleted, and click Delete Reports button. Related concepts • About analyzing storage system performance on page 403 Related tasks • Creating periodic health check reports for storage systems performance analysis on page 412 • Exporting health check reports on page 415 Metrics of resources of Virtual Storage Platform G1000, Virtual Storage Platform and HUS VM (Health Check Report) Volume metrics for enterprise storage systems, as shown in the health checkreport, can help identify performance issues. The following provides the values for evaluating items that are used byHealth Check to verify whether there is a performance problem for Virtual Storage Platform G1000, Virtual Storage Platform or HUS VM. During Health Check, if items exceed the following thresholds, the status isjudged as Caution or Warning. Table 12-4 Resource metrics for Virtual Storage Platform G1000, Virtual Storage Platform and HUS VM (Health Check Report) ItemMetricThresholdNotesCache (by CLPRs), Cache Total (entire storage system)Write Pending RateCaution: > 30% Warning: > 60%• Up to 30% is considered normal. • Frequent or sustained episodes of 40% deserve attention. • Frequent or sustained episodes to 50% require prompt remediation. • At 70%, inflow control is enabled which is disruptive. • High write pending is a sign of back end congestion.MP Blade (Virtual Storage Platform G1000, Virtual Storage Platform or MP Unit (HUS VM)Busy RateCaution: ≧ 40% Warning: ≧ 80% • - 416Monitoring managed resources and resolving alertsHitachi Command Suite User Guide
ItemMetricThresholdNotesParity GroupMax Busy RateCaution: ≧ 50%, Warning: ≧ 80% • -Tip: If you want to view detailed performance information, start Performance Reporter. By doing so, you can perform an analysis per minute or a long- term trend analysis for more performance evaluation items. Related concepts • About analyzing storage system performance on page 403 Metrics of resources of Universal Storage Platform V/VM and Hitachi USP (Health Check Report) The following provides the values for evaluating items that are used byHealth Check to verify whether there is a performance problem for the Universal Storage Platform V/VM or Hitachi USP platforms. During Health Check, if items exceed the following thresholds, the status is judged as Caution or Warning. Table 12-5 Resource metrics of USP V/VM and Hitachi USP (Health Check Report) ItemMetricThresholdNotesCache (by CLPRs), Cache Total (entire storage system)Write Pending RateCaution: > 30% Warning: > 60%• Up to 30% is considered normal. • Frequent or sustained episodes of 40% deserve attention. • Frequent or sustained episodes to 50% require prompt remediation. • At 70%, inflow control is enabled which is disruptive. • High write pending is a sign of back end congestion.Side File Usage RateCaution: ≧ 10% Warning: ≧ 20%• This is only applicable for TrueCopy Asynchronous. • Little, if any, should be reported under normal conditions. High side file usage is a sign of downstream congestion.Channel ProcessorBusy RateCaution: ≧ 45% Warning: ≧ 80% • 80% approximate maximum. Parity GroupMax Busy RateCaution: ≧ 50%, Warning: ≧ 80% • - Disk ProcessorBusy RateCaution: ≧ 40%• -Monitoring managed resources and resolving alerts417Hitachi Command Suite User Guide
ItemMetricThresholdNotesWarning: ≧ 80%Tip: If you want to view detailed performance information, start Performance Reporter. By doing so, you can perform an analysis per minute or a long- term trend analysis for more performance evaluation items. Related concepts • About analyzing storage system performance on page 403 Metrics of resources of midrange storage systems (Health Check Report) The following provides the values for evaluating items that are used byHealth Check to verify whether there is a performance problem for the HUS 100, Hitachi AMS 2000, Hitachi SMS, and Hitachi AMS/WMS platforms. During Health Check, if items exceed the following thresholds, the status is judged as Caution or Warning. Table 12-6 Resource metrics of midrange storage (Health Check Report) ItemMetricThresholdNotesCacheWrite Pending Rate on Controller 1, Write Pending Rate on Controller 0Caution: > 25% Warning: > 50%• The maximum Write Pending level per Partition is 70%. • There are cases of resolving the problem by distributing the workload to other controllers, adding to the cache, using high-speed drives, adding drives to parity groups, adding parity groups, or switching to RAID 1+0.ProcessorBusy RateCaution: ≧ 80% Warning: ≧ 90% • - Tip: If you want to view detailed performance information, start Performance Reporter. By doing so, you can perform an analysis per minute or a long- term trend analysis for more performance evaluation items. Related concepts • About analyzing storage system performance on page 403 418Monitoring managed resources and resolving alertsHitachi Command Suite User Guide
Changing peformance threshold settings to correctly analyze storage system performance The recommended threshold values for analyzing storage system performance are set in advance. If the recommended values are not suitable, change the threshold values. Use the metrics of volumes of each storage system as a reference to change the threshold. There are four types of thresholds: • Thresholds that are used only in the Identify Performance Problems wizard. • Thresholds that are used only in the Health Check window. • Thresholds that are used only in the Analyze MP Blades/Units window. • Thresholds that are common to the Identify Performance Problems wizard, the Analyze MP Blades/Units window and the Health Check window.Note: When you change a threshold common to the Identify Performance Problems, Analyze MP Blades/Units and Health Check, the change is applied to all analysis tools in their respective windows.Tip: In accordance with the specified threshold values, the background of the displayed analysis result chart is colored for each threshold range. The Caution threshold range is displayed in yellow. The Warning threshold range is displayed in pink. Procedure 1. On the Analytics tab, in General Tasks , select Performance Threshold Settings . 2. Select the series of the target storage system, and then use the metrics of volumes as a reference to change the threshold value for each metricof interest. 3. To change threshold settings, do the following: a. Select the target storage system. b. Select the analysis tool whose metrics you want to change (for example Analyze MP Blades/Units) or select All to display and change all available metrics. Tip: Metrics might be deactivated based on your analysis tool selection. c. Change the caution and warning (Health Check only) thresholds to meet your requirements. d. Click OK. Monitoring managed resources and resolving alerts419Hitachi Command Suite User Guide
Result Metrics are changed for one or more analysis tools. Related concepts • About analyzing storage system performance on page 403 Related tasks • Identifying storage system performance problems on page 405 • Analyzing and load balancing MP Blades or MP Units on page 407 • Creating periodic health check reports for storage systems performance analysis on page 412 Related references • Metrics of volumes of Virtual Storage Platform G1000, Virtual Storage Platform, and HUS VM (Identify Performance Problems wizard) on page 409 • Metrics of volumes of Universal Storage Platform V/VM and Hitachi USP (Identify Performance Problems wizard) on page 410 • Metrics of volumes of midrange storage systems (Identify Performance Problems wizard) on page 411 • Metrics of resources of Virtual Storage Platform G1000, Virtual Storage Platform and HUS VM (Health Check Report) on page 416 • Metrics of resources of Universal Storage Platform V/VM and Hitachi USP (Health Check Report) on page 417 • Metrics of resources of midrange storage systems (Health Check Report) on page 418 Configuring the Performance Reporter display After activating Performance Reporter, you can analyze detailed storagesystem information based on Identify Performance Problems results or Health Check report results. The following table lists the configurable resources for displaying performance and status information based on storage system, selected resources, and the available Identify Performance Problems or Health Check information. Table 12-7 Performance Reporter configurable resources for display Storage SystemResourceValueVirtual Storage Platform G1000 Virtual Storage Platform Unified Storage VMLDEV NumberValue of the Volume columnCLPR NumberValue of the CLPR columnAdaptor IDValue of the MP Blade column (or value of the MP Unit column)Processor ID (for MP Blade or MP Unit)_Total (default value)420Monitoring managed resources and resolving alertsHitachi Command Suite User Guide