hpe 3par system reporter with performance and capacity

31
HPE 3PAR System Reporter Performance and capacity management via HPE 3PAR StoreServ Management Console Technical white paper

Upload: lykhanh

Post on 04-Jan-2017

221 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: HPE 3PAR System Reporter with Performance and capacity

HPE 3PAR System Reporter Performance and capacity management via HPE 3PAR StoreServ Management Console

Technical white paper

Page 2: HPE 3PAR System Reporter with Performance and capacity

Technical white paper

Contents Executive summary ................................................................................................................................................................................................................................................................................................................................ 4 Overview of System Reporter components ..................................................................................................................................................................................................................................................................... 4

On Node System Reporter ........................................................................................................................................................................................................................................................................................................ 4

HPE 3PAR StoreServ Management Console............................................................................................................................................................................................................................................................ 4

HPE 3PAR Management Console (Legacy) ............................................................................................................................................................................................................................................................. 5

External System Reporter (Sustaining) ......................................................................................................................................................................................................................................................................... 5

System Reporter in SSMC ................................................................................................................................................................................................................................................................................................................ 5 Main Menu ............................................................................................................................................................................................................................................................................................................................................... 6

Navigation ................................................................................................................................................................................................................................................................................................................................................ 6

Create report template ................................................................................................................................................................................................................................................................................................................. 8

Customization ....................................................................................................................................................................................................................................................................................................................................... 9

Scheduler ................................................................................................................................................................................................................................................................................................................................................ 12

Schedule reports—updated in SSMC 2.2 ................................................................................................................................................................................................................................................................... 12 System Reporter Top or Bottom 10 List .................................................................................................................................................................................................................................................................... 14

Viewing chart....................................................................................................................................................................................................................................................................................................................................... 15

Performance reports ..................................................................................................................................................................................................................................................................................................................... 16

Compare with—new in SSMC 2.2 ...................................................................................................................................................................................................................................................................................... 16

Compare by—New in SSMC 2.2.......................................................................................................................................................................................................................................................................................... 18

At Time popups—Modified in SSMC 2.2 .................................................................................................................................................................................................................................................................... 19 Histograms ............................................................................................................................................................................................................................................................................................................................................ 19

Capacity ................................................................................................................................................................................................................................................................................................................................................... 20

Licensing ................................................................................................................................................................................................................................................................................................................................................ 20

Real time reporting .............................................................................................................................................................................................................................................................................................................................. 20

Threshold Alerts ..................................................................................................................................................................................................................................................................................................................................... 21

Report generation ........................................................................................................................................................................................................................................................................................................................... 21 Severity levels .................................................................................................................................................................................................................................................................................................................................... 22

Alert threshold customization............................................................................................................................................................................................................................................................................................. 22

Threshold Alert notification .................................................................................................................................................................................................................................................................................................. 22

Threshold alert—Updated in SSMC 2.2 ..................................................................................................................................................................................................................................................................... 24

Threshold Alert use case ......................................................................................................................................................................................................................................................................................................... 25

Page 3: HPE 3PAR System Reporter with Performance and capacity

Technical white paper

System Reporter database ........................................................................................................................................................................................................................................................................................................... 26

Database retention ....................................................................................................................................................................................................................................................................................................................... 26

Collection of data samples ..................................................................................................................................................................................................................................................................................................... 27

Real time data samples ............................................................................................................................................................................................................................................................................................................. 27 Grow .srdata—new in SSMC 2.2......................................................................................................................................................................................................................................................................................... 28

Backing up .srdata—New in SSMC 2.2......................................................................................................................................................................................................................................................................... 28

Exporting the .srdata—New in SSMC 2.2 .................................................................................................................................................................................................................................................................. 29

Examining data with HPE 3PAR Excel plugin—updated in SSMC 2.2 ........................................................................................................................................................................................... 30

Terminology................................................................................................................................................................................................................................................................................................................................................ 31

Page 4: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 4

Executive summary The importance of managing an array’s performance and capacity is paramount to ensuring that critical applications have the resources needed during peak demands. These resources rely on a tool or tools to report metrics in a timely and accurate fashion. Performance monitoring allows the storage administrator the flexibility to view real time data as well as performance trends by using historical data stored within a database. HPE 3PAR System Reporter enables users to do complete reporting of their HPE 3PAR StoreServ Storage systems.

HPE 3PAR System Reporter Software is a feature that can be easily enabled on all HPE 3PAR StoreServ systems. Once an HPE 3PAR StoreServ system is installed, System Reporter automatically collects data on a number of different object data points in the background without the need of any additional setup.

With HPE 3PAR StoreServ Management Console (SSMC) or the HPE 3PAR Command Line Interface (CLI), a user displays collected data from an array. Although some reports are available to the user without a license, to view all collected data a license is required. Using the SSMC, graphical reports are available for the following metrics.

• Historical data

– Performance

– Histogram

– Capacity

• Real time data

– Performance

HPE 3PAR System Reporter is a feature-rich analytical engine, which helps users interpret collected performance data. With the release of HPE 3PAR OS 3.2.1 MU2, System Reporter also incorporates reporting on both block and file HPE 3PAR StoreServ Storage systems. The availability to manage and report on both Block and File HPE 3PAR StoreServ Storage systems is available via HPE 3PAR SSMC 2.0.

Since HPE 3PAR StoreServ OS 3.1.2, HPE 3PAR System Reporter has moved from an external data structure to the use of an internal data structure designed to retain years’ worth of data created on an internal system volume (srdata). The enhancement in this architecture frees the user from maintaining a separate system and database for monitoring HPE 3PAR StoreServ Systems and allows the system administrator quick access to data.

Collected data can be viewed by creating reports via templates, which are available under the System Reporter section using SSMC, this same section also allows a user to configure real time performance reporting, which graphs current loads on the array.

Overview of System Reporter components On Node System Reporter HPE 3PAR Management Console or the SSMC both use On Node System Reporter. On Node System Reporter allows the administrator to manage the HPE 3PAR StoreServ Storage systems from a single management window using data stored on the HPE 3PAR StoreServ array. On node data storage allows for easy access to data used in reporting, with no user maintenance and long-term data storage.

HPE 3PAR StoreServ Management Console HPE 3PAR system management has taken on a bold new design with the release of the new HPE 3PAR SSMC. The new look SSMC uses the highly acclaimed design of the HPE OneView architecture for the management of the HPE 3PAR StoreServ array. Figure 1. SSMC management, highlights some of its features.

The SSMC software is designed using HPE Piano software. Piano toolkit is a user-center design process. It includes a library of best practices, design flows, and UI widgets. The HPE Piano software framework provides a common set of software interfaces to centralize and simplify access to HPE architecture.

Page 5: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 5

Figure 1. SSMC management

HPE 3PAR Management Console (Legacy) The focus of this paper and System Reporter is centered on the continuing migration of HPE 3PAR Management Console to HPE 3PAR SSMC. The older management console will not contain the next-generation reports outlined in this paper. SSMC and the legacy HPE 3PAR Management Console can co-exist on the same server. HPE recommends managing HPE 3PAR StoreServ with SSMC.

External System Reporter (Sustaining) HPE will continue to support the External System Reporter but no new functionality will be included. HPE 3PAR System Reporter 3.1 is the last version where software enhancements were added. External System Reporter has been moved to a support model. HPE recommends using SSMC for historical and capacity reporting of the HPE 3PAR StoreServ.

System Reporter in SSMC SSMC comes with a fresh way of reporting capacity and performance data. Using reports under the System Reporter Main Menu, the user can launch the integrated reporting tool within SSMC. Reports enable the user to run historical and real time reports. This new approach offers great advantages over the previous approaches of querying historical data.

System Reporter in SSMC provides the following main features: • Convenient access to configuration options for selecting systems to include in reporting, specifying sampling parameters, scheduling reports,

and generating alerts

• Extensive selection of reports for obtaining performance and storage utilization statistics on selected objects (i.e., hosts, ports, nodes, physical disks, Virtual Volumes, and more)

• Quick access to predefined reports that contain useful statistics for most common types of installations

• Customization of reports using the standard Web interface that provides specifically selected and formatted reports for specified systems

• Options for choosing the time and duration for the collection of reporting statistics, which can be initiated at a specific time, collected over a period of time, and compared between a ranges of periods

• Capability to isolate and zoom in and out of time periods

• Performance alerts that can be configured via Threshold Alerts, once criteria for alert is met, visual notification of the alert is displayed on the SSMC dashboard

• Ability to edit generated reports, change object definitions

• Report customization allows hundreds of different report generation

Page 6: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 6

Main Menu The Main Menu shown in Figure 2 is an express-driven interface to all points within SSMC and the objects monitored. From the Main Menu, a user is linked directly to any of the listed context areas.

Figure 2. Main Menu

Navigation Reports under System Reporter shown in Figure 2, is the direct link to HPE 3PAR System Reporter from the Main Menu. The reporting link gives users direct access to a feature-rich reporting mechanism in System Reporter. A key feature within System Reporter is predefined report templates. Templates offer the user quick and easy access to the following types of reports:

• Real time reporting

• Performance reports

• Histograms

• Capacity

Templates are categorized as using the field of Data Type and Category as shown in Figure 3. Template format.

Figure 3. Template format

As shown in figure 3, there are different data types and categories associated with the performance chart the user chooses.

Page 7: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 7

Data types Historical data: Charts that collect stored data for the defined object. Each definition of the charted historical object contains a default period for viewing. Each stored historical object can be modified to chart a specific time to suit the user’s needs. Retention time and sampling periods are covered completely in the section titled Database retention.

Real time data: Charts that collect data, which has been collected on an object in the last five seconds. The charted data is a continuous plotting graph of object data. Data charted can be the default object data or user-defined object data. Customization of charts is discussed under the section titled Customization.

Categories Histogram: The charts are a graphical representation of the distribution of data. The construct of a histogram is to divide the entire range of values into a small series of intervals and then count how many data collection samples fall into each interval. A graph is then drawn itemizing each of the collection points over the time data was collected. Figure 4. Host Port Histogram is an example of a histogram.

Figure 4. Host Port Histogram

Performance: The displayed charts contain the aggregated historical data collected for the performance object selected. Performance objects (IOPS, bandwidth, and service times) are monitored which includes logical and physical devices as shown in Figure 5. Performance reporting.

Page 8: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 8

Figure 5. Performance reporting

Capacity: It represents the charts usage and space assessment of the selected object. Capacity charting provides historical usage data useful for forecasting. They are provided for activities such as capacity usage and space assessment for used space, available space, base volume space, copy space, RAID level, and drive type, as well as system space. The following capacity charts can be created:

• Physical drive capacity

• System capacity

• Virtual Volume capacity

• Common Provisioning Group (CPG) capacity

• Adaptive Optimization—space moved

Create report template By default, each report is created with the following categories as shown in Figure 6. Create report.

Figure 6. Create report

Page 9: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 9

Reports are created using the +Create report button in the left pane of the Reports Screen. The user may also use the drop down menu under Actions to create a report.

Customization Each object selected for reporting uses a default setting for the particular object that is being charted. The default values for each charted object will differ from object to object. Customization allows the user to select the variables in which each chart is generated. As there are hundreds of combinations, which can be chosen using customization, only the main selection points will be discussed. Refer to Figure 6. Create report for each of the main categories listed here.

General This area is used for system selection, selecting object template and report name. If more than one system is being monitored with SSMC, there is a drop down that contains system names of each HPE 3PAR StoreServ array. Figure 7. General customization illustrates selection choices. By default identification of the report chart being generated is populated in Report name, this value can be altered by the user and if more than one array is being monitored it is a best practice to alter the report name. The description field is not populated and is optional. Information populated in the description field is either displayed during the initial creation of the chart or when editing, it is not displayed when the chart is viewed.

Best practice—always modify Report name to include a unique identifier for each chart, this customizes the report and allows the user to select the Report template more than once.

Figure 7. General customization

Object selection There are two selections available, by default All is selected, which enables the user to chart all objects for the selected category. As an example, if the user selects to chart Real Time Exported Volume Performance, the choices for Object Selection are All or Filter by objects. Using the default of All then charts performance of all volumes created on the array. Selecting Filter by objects, allows the user to choose which object is charted, an example is illustrated in Figure 8. Add object.

As illustrated, the user can choose to monitor an object, objects include Virtual Volume, Host, or Port number. The user may highlight several objects and use the Add radial button. If the user needs to select multiple items, which are not all displayed, the user can highlight one object, press Add +, scroll down the list of menu items, and add additional objects by either using Add + for multiple objects or Add, if only one additional object is added.

The user may also search for objects to chart using the search bar that the magnifying glass identifies.

Page 10: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 10

Using Filter by objects, the user must also choose the plotting style. The two choices are:

• Individual series for each selected objects

• Aggregated series for all selected objects

Individual plot for each selected object is applicable only when filter by objects is selected. For all objects, this option is not valid.

Best practice—use search bar to identify objects to enter rather than scrolling through menu items.

Figure 8. Add object

Time Settings Time Settings are fully dependent upon selected object reporting. In general, time settings make enquiries on the following attributes, each attribute can be customized to fit the user’s needs in reporting.

• Report type

– Time interval—plots data over a period of time

– Specific time—user can compare a large number objects specific to an instance of time

• Time interval

– Default—time frame, which is charted is defined by period under settings

– Custom date range—if selected, two independent boxes appear in the tool; when the cursor is placed within the box, a calendar will pop up to assist the user for date and time selection

From—date and time when the user starts reporting

To—date and time when the user ends reporting

• Sampling—user is given three data sample choices to create a chart

– Hi-res—an average of all collected data during the last five-minute period

Page 11: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 11

– Hourly—an average of all collections during the last one hour

– Daily—an average of all collections during the last 24 hours

• Period—this setting only applies if the user selects Time interval as a report type and Default as the time interval. A drop-down menu appears and the user can select a reporting time of one hour to three months.

Chart options Chart options will vary with each selected object. In general, they will fall into four categories.

• Capacity—there are no chart selections for capacity charts

• Historical performance—here chart selection generally contains several historical types as listed below, it should be noted that not all of these selections are available for all historical performance charts:

– IOPS

– Service time

– Bandwidth

– I/O size

– Queue length

– Average busy

• Histogram—charting this option, the user has multiple choices in creating a customized chart, these include:

– I/O time—minimum and maximum values chosen through drop-down boxes

– I/O size—minimum and maximum values chosen through drop-down boxes

– I/O types—read, write, total chosen through check boxes

– Show chart—the user selects to chart I/O time, I/O size, or both

– Access count—chart displays percentage or number of entries

• Real time performance—many of the same chart options available with historical performance are available for selection with the show chart as shown in Figure 9. Real time performance options. I/O types are also selectable for charting.

Figure 9. Real time performance options

Page 12: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 12

Scheduler Creating charts in a scheduled timeframe is an important aspect in chart creation. This gives the user the ability to share chart content outside the confines of the SSMC. Charts created during a scheduled timeframe can be emailed or displayed as a .csv or .pdf on the localhost.

After all the chart selections have been made, the schedule pattern creates a chart. The scheduler is available at the bottom of the page. However, if a user choses to create a reoccurring chart, they can select one of the following schedule patterns:

• Once later

• Hourly

• Daily

• Weekly

• Advanced

In figure 10, a “Weekly” schedule pattern has been selected. In this example, using the weekly schedule the user creates a chart on weekly basis, which is updated on the SSMC at the predefined time. In the example, the user has scheduled the report to be run weekly showing the capacity usage for fast class (FC) drives. Other parameters can be associated with the chart as shown in the figure.

The result of this action is that the displayed chart is updated the next time the user logs into the SSMC after the scheduled time has been met. Included as part of the scheduler, if the user so chooses, they can email the report or view it as a .pdf or .csv file.

Once a scheduled task has been created, specific information related to the task can be viewed—either from Schedules on the Main Menu or by opening the created chart using the drop-down list that is associated with the created chart.

Opening up Schedules from the Main Menu will list all currently scheduled tasks, list the type of scheduled tasks, the state of each task whether it is active or suspended, and the next run of the task. From this menu, the user can also use the drop down on the right-hand panel to modify a selected task. Modifications include:

• Delete the task

• Suspend the task

• Resume the task, if in an suspended state

• Delete the task

Scheduled tasks viewed from the charts menu under System Reporter allows the user to modify or view the task for the chart selected. The user would select a chart from the list of already created charts, use the drop down next to the named chart, and select Schedules. A list of the charts created from the schedule will be displayed; the charts by default will only remain for seven days. There could be two entries associated with each scheduled collection, a .csv file and a .pdf file.

Schedule reports—updated in SSMC 2.2 All reports created in System Reporter now have the option to be exported via a .pdf report or a .csv file. The user can use the Actions drop down on the right-hand side of the screen and select to export the report. The exported file can be saved to a location (typically download directory on server which host SSMC) or opened for processing and examination.

Problem statement System administrator must report to management each week on space usage. Reports are intended to help management forecast budgets for future storage projects. The manager is only interested in a .pdf chart. Due to security regulations, the system administrator is also required to keep all generated reports for 30 days. The manager is only interested in usage of FC drives.

Actions System administrator uses the following actions to create a report.

1. Go to Settings in the Main Menu

2. Under Actions, choose Select to edit fields

Page 13: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 13

3. Change retention period to 30

4. Populate fields for SMTP server and SMTP port along with email of manager requesting report

5. Save and close Settings menu and use Main Menu to open Reports under System Reporter

6. Use + Create report (if not already created) to create a report for System Capacity

7. Use filter by objects and rules to select just FC drives

8. Use Figure 10. Scheduler as a reference to change time settings, chart options, and to set up a schedule

9. Save your settings

Figure 10. Scheduler

By default, all reports created with the scheduler are retained for seven days unless changed in the settings menu. Reports can be viewed with a link in System Reporter or they are also kept locally on the system, where System Reporter is installed. All settings for display and retention polices are configured using the Settings section located under the Main Menu.

Note If the user deletes the report where scheduled generated reports are available, previous generated reports will be deleted, and all future scheduled reports will be canceled. So, prior to deletion, the user should copy the data to another location.

Page 14: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 14

Figure 11 is an example of a generated .pdf file. The report is two pages in length and was generated capturing usage for FC disks.

Figure 11. Report from scheduled reports

A chart, which has been created using the scheduler, can be exported from the system in which the schedule was created. Use case: Management expects system administrator to create a report of consumed storage space each month.

System Reporter Top or Bottom 10 List Open System Reporter and select + Create report, after examining the report type of performance under the section titled Category, the user will find a number of reports which end in Performance Statistics. With the exception of reports, which report on CPU statistics and the reports identified as real time, the user can select from Time Settings the radial of Specific time.

After selecting the Specific Time, expand the chart options to include the metric for plotting. The user will have the choice to plot the Top 10 or the Bottom 10. The user also can change the setting to reflect the Top 20 and above in increments of 10. Options to select which measurements are to be captured and the order in which to present them is also included.

As an example, use the illustration in Figure 12. Top 10 host IOPS, where we have charted exported volume performance of the Top 10 hosts at the time of collection. The upper portion of the graph displays hosts with the largest amount of IOPS. The lower section of the graph displays service times of the Top 10 hosts.

Page 15: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 15

Figure 12. Top 10 host IOPS

Viewing chart Once charts are created, the chart will be displayed in the main viewing window. Dependent upon the chart created, the views will be different. Use the illustration in Figure 13. Host Port Histogram chart.

The chart can be broken down into three areas:

• Summary: Displays details of selection criteria chosen for the chart that’s currently displayed

• Chart: Is detailed view of data collected during the reporting period

• Actions drop down: This is discussed later in section

Interactive display and zoom: The user can interactively place the mouse over any portion of the graph to display the values at the time of collection as shown in Figure 13. Host Port Histogram chart. Since this is a daily collection, the data, which was collected at midnight, is shown.

The user may also zoom in on any area by clicking on a data point and then dragging the mouse either left or right of the data point. The result of this action is that the graph will be repainted and the sampling period is now shown as hourly instead of daily as first displayed. A user can zoom in farther on the graph by repeating the same action. Further action would result in a hi-res display.

Figure 13. Host Port Histogram chart

Page 16: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 16

Actions drop down: The drop down at the top right of the chart allows the user to modify parameters of the chart currently being displayed. These options will change depending on the type of report selected. In general, the following four items are available on all charts.

• Create—selection is identical to the + Create report button on the left context pane. Create will open create report template as discussed earlier in the paper

• Edit—allows user to edit current chart options to include changing name, time, chart objects, and chart options

• Delete—deletes selected chart

• Reset zoom—resets a chart, which has been zoomed in to the original chart

Performance reports Performance reports are charts of aggregated historical data collected for the selected performance object. Performance data is categorized as historical data or real time data. Historical data can be displayed using either default values or custom values. Historical performance data makes up approximately 70 percent of all reporting objects collected; it includes but is not limited to the following objects in Figure 14. Historical performance objects.

Figure 14. Historical performance objects

Compare with—new in SSMC 2.2 A new feature has been added to SSMC 2.2 whereby a user can compare two arrays on the same chart. The Compare with feature is currently only enabled with Physical Drive performance. The user creates a report using the drop down selector for Physical Drive performance as shown in Figure 15. Compare with feature.

Figure 15. Compare with feature

Page 17: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 17

If the user selects None then no comparisons will be used when creating the graph output. If the user chooses to select Systems as shown in the figure, the user can select systems to compare with by clicking on Add systems. A separate pop-up window displays systems to be added for comparison charts. Once systems are added, the chart output is displayed as shown in figure 16. A maximum of four systems can be used for the comparison.

Figure 16. Compare with—two systems

In figure 16, a comparison of two systems is displayed. An individual line represents each system. The user can mouse over any intersection of lines to display recorded values at the time of the collection. The user can left click and an “At Time” popup chart will be displayed as shown in Figure 17. Compare with—At Time pop-up.

Figure 17. Compare with—At Time pop-up

Page 18: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 18

Compare by—New in SSMC 2.2 New in SSMC 2.2 is the Compare by function. Compare by allows the user to compare top and bottom objects against a selectable metric. Selection of a metric is chosen by a drop-down menu, which is further filtered by the charted object that was chosen. Using Figure 18. Compare by—Drop-down menu, the user has the option to compare values for numerous collection points. In this example, we will create a chart titled “Exported Volumes Performance” as illustrated in Figure 19 from the values selected through the following drop down menu.

Figure 18. Compare by—Drop-down menu

The result of the selected choices is displayed in Figure 19. Exported Volumes Performance—Compare by. In Figure 19, the Top 10 volumes are being compared. As shown in figure 18, the direction of top or bottom objects can be charted and up to 15 volumes can be compared. By default, the top direction and 10 objects are always compared.

Compare by is available for Physical drives, Exported volumes, Ports, Adaptive flash cache, and Priority optimization performance reports. As shown in figure 19, the user can mouse over any collection point and view the data at the time of collection. Although not shown here, a user can use their mouse and point to a timeframe interactively on the graph. A popup box will appear with the values which were recorded at the time of the collection. The user can left click and an “At Time” report will be generated in another popup box with a bar chart of those collected samples, figure 17 illustrates an example of an “At Time” report.

Figure 19. Exported Volumes Performance—Compare by

Page 19: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 19

At Time popups—Modified in SSMC 2.2 In several examples shown throughout the paper, the user can create an At Time popup. As illustrated in, a new interactive feature exists with drop down menus. Dependent upon the object being charted the drop downs will change. In figure 17, the top ribbon contains “Top 10, Write bandwidth, Id, Hi-res,” collection time. Each one of these items can be changed to view a different set of charts. Options for this chart are as follows:

• Top 10 to Top 100

• Bottom 10 to Bottom 100

• Write bandwidth or read bandwidth

• Grouped by id, port node, port slot, port number, device type, or speed

• Current collection time, previous collection time, next collection time, or custom

Histograms As mentioned previously, a histogram charts a graphical representation of data distribution. Currently, there are four histograms available to report.

• Enclosure port histogram

• Exported volumes histogram

• Host port histogram

• Physical drive histogram

All histograms have common charting options, I/O time, I/O size, and I/O types. Object selection will differ with each histogram type. An example of a histogram plotting host port data is illustrated in Figure 13. Host Port Histogram chart. With this example, the user can select to filter by object or by rules as illustrated in Figure 20. Host port histogram—object or rules.

Figure 20. Host port histogram—object or rules

Best practice—displays only one rule at a time, combining rules does not allow user to differentiate the outcome of the object.

Page 20: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 20

Capacity Good capacity planning is paramount with any storage array. Using the SSMC, reports can be generated to assist with capacity planning by reporting free space available for the volumes to grow for each of the CPGs as one example. Another example is tracking capacity utilization trends. System Reporter reports on the following capacity objects.

• Virtual Volume

• System

• CPG space

• Physical drive

Licensing System Reporter license requirements are described in Table 1.

Table 1. License requirements

System Reporter function License required

Live performance reports No

Historical capacity and performance reports Yes

I/O density reports No

Physical drive capacity report No

Real time reporting Real time reporting offers the user reporting capabilities on select monitored objects. Currently, the following can be monitored through real time reporting, other objects will be added in future releases of System Reporter.

• Exported volumes

• Physical drive

• Port data

• Port control

This style of reporting offers the user the greatest insight on monitoring an HPE 3PAR StoreServ array. Real time reporting by default is displayed in 5-second intervals plotting the latest collection of data. It is highly recommended that users create customized charts to monitor specific data points on the array.

Plotting: Data will plot once every 5 seconds on real time collections. After the first plot, the graph will continuously update. The user can change to other graphs and return to continue monitoring the object.

If the user changes charts and then returns to a chart, which is displaying real time data, they can either choose to display only the data which has been collected since returning to the chart or they can use the drop-down menu under the Actions tab and select All real time samples. This will change the chart to reflect all samples collected during the current collection period.

Actions drop down (additional actions): With real time reporting there are two additional action items in the actions drop-down menu.

• Start or stop real time—this selection offers the user the ability to either start or stop real time data displays. If user moves to another report and does not visit this report for more than 30 minutes; then the report is stopped. The user can restart the report from the drop-down menu.

• All real time samples—if a user creates a report and then moves to another report, the real time report display will be suspended. If the user returns to the real time report within 30 minutes, the user can invoke All real time samples and this will return the last thousand samples retained in the backend cache.

Once the user disconnects from System Reporter or removes a system, all live reports running on the system will be terminated.

Page 21: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 21

Threshold Alerts Alerting on predefined threshold set against performance metrics is essential to ensuring the HPE 3PAR StoreServ array meets demands of everyday computing. HPE 3PAR System Reporter enables the storage administrator to set thresholds and alert for performance on system components.

Report generation Starting from SSMC version 2.1, the user can set threshold alerts from within the SSMC. Threshold alerts selected from the SSMC are predefined templates, which may have predefined values assigned to each alert. Some threshold alert definitions may be shell alerts. A shell alert has a defined template but no default metric associated to it; it is the user’s responsibility to set the threshold.

Threshold alerts are designed in categories, with each category containing a number of alert templates. The major categories include:

• Controller node

• Exported volumes

• Physical drives

• Port thresholds

• Other (node link alert, QoS, RC, and RC volume alert)

An example of setting a controller alert is shown in figure 21. This example displays predefined values as being set. There may be other values, which have not been defined, and can be altered by selecting “Add metrics” and modifying the assigned values.

Figure 21. Exported Volume Threshold Alert

Page 22: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 22

Severity levels Each of the reports as shown in Figure 21, contains a severity level. These levels are indicators of the threshold impact on the array performance. Levels are categorized as:

• Info—reference to a threshold attained with no impact to performance

• Minor—a threshold, which signals a low-level impact on array performance i.e., delayed I/O to host

• Major—a threshold, which characterizes an increased impact on array performance; as an example refer to Figure 21, if CPU idle time were to drop below 10 percent, this signifies CPU usage is above 90 percent, which would affect host I/O

Alert threshold customization As the case with creating reports within System Reporter, each of the threshold alert templates can be customized. Each of the templates contains the following categories:

• General—customization in the type of resolution and severity level

• Objects—customization in selection of objects to be monitored for alerting

– All

– Filter by objects

– Filter by rules

• Metrics—object threshold witch is monitored and the values of the monitored object

If an alert is not part of the predefined alert templates contained within the SSMC, the user can create a custom event alert using the HPE 3PAR StoreServ CLI. Once the custom alert is created, the alert can be viewed in the SSMC as part of the Threshold Alerts. In Figure 22. CLI Threshold Alert creation a custom alert “cpuMinorAlertIdle20pct” was created. It is now part of the Threshold Alerts and the SSMC can view it.

Figure 22. CLI Threshold Alert creation

Threshold Alert notification A Threshold Alert notification can be viewed from several sources. The SSMC dashboard will display a performance donut, which can be viewed in one of two colors:

• Green—no alerts have been recorded

• Yellow—an alert criteria has been met

If an alert criterion has been met, System Reporter displays the performance alert on the dashboard as shown in Figure 27. SSMC dashboard. Clicking on the Performance donut links the user to threshold alerts page within the SSMC.

To check if an alert has been triggered, as shown in figure 23, monitor the threshold alert page linked from the Main Menu. The user can further identify the alert area by clicking on the alert.

Page 23: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 23

Figure 23. Triggered alert an informational banner that displays the alert criterion. The user can further identify the alert area by clicking on the alert.

Figure 23. Triggered alert

Clicking on the alert displays the values at the time the threshold alert was exceeded.

Figure 24. Detailed Threshold Alert is an example, where the port that has exceeded the threshold is displayed along with the measured values that tripped the alert. In this example, we are displaying the thresholds reached on port 1:6:4 or controller (node) 1, slot 6, and port 4. Subsequently, if the user highlights the next alert up from the present location (on the right side), this would change the controller (node), slot, and port number, which registered the triggered alert.

Note The displayed values were taken from an HPE 3PAR StoreServ Array 10400, the port numbers will differ from array to array. The user should reference the node, slot, and port layout specific to their array.

Figure 24. Detailed Threshold Alert

The user can also display the information as shown in figure 25. In this display, we have changed from an “Overview” as seen in the other displays to an “Activity”. Each of the triggered ports and the recorded values for the triggers are now displayed on one page.

Page 24: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 24

Figure 25. Threshold Alert—Activity

Note The alerts are not cumulative, meaning if the alert is triggered during the next sample period, only the current alert is displayed. Therefore, many alerts may have occurred prior to the sampling period but only the most recent trigger will be displayed. The user can refer to prior collection periods by customizing SR reports for the metric being monitored.

Threshold alert—Updated in SSMC 2.2 Email notification SSMC 2.2 now supports direct email alerts from SSMC. The user must enter the required email criteria under “Settings” from the Main Menu. Once the setup is ready, tested, and after alert creation, the email notification is enabled.

Note Email notification is a selection, it can be adjusted by editing the threshold alert.

Recurrence pattern A recurrence pattern is self-defined by the user to meet specific criteria prior to a Threshold Alert sending a notification. Figure 26. Recurrence pattern illustrates the use of a recurrence pattern. In this figure, the thresholds, which are set, must be exceeded twice in the last two samples (in this example, samples are collected hourly).

Figure 26. Recurrence pattern

Page 25: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 25

Threshold Alert use case Problem statement: A storage administrator configures Threshold Alert for a number of key elements on the HPE 3PAR StoreServ array. The admin uses the dashboard to monitor operations of the storage system as shown in Figure 27. SSMC dashboard .

Figure 27. SSMC dashboard

During one of the routine dashboard checks, the administrator discovers that a performance alert has been triggered. To isolate the trigger, the administrator performs the following actions:

Actions: 1. Clicks on the Performance donut as seen in figure 27

2. This action redirects the screen to the user can further identify the alert area by clicking on the alert

3. Figure 23. Triggered alert

4. System administrator clicks on drop-down list next to ports display and a report of the values, which triggered the alert, are displayed (Figure 25)

5. The user clicks on “Reports” under the System Reporter directory listed in the Main Menu (figure 2); the user creates a report using “Exported Volumes Performance” from the selection of reports and observes high service time as illustrated in figure 5

6. To refine a root cause further, the user can create a histogram as shown in figure 4, which shows the values and the number of instances those values were reported that triggered the alert. In the depicted example, 77 samples were observed to exceed 64 ms to service an I/O write; our threshold was set for 30 ms

Solution 1: Using this example, the ports, which are exceeding the Threshold Alert, are disk ports. The disk ports service I/O to the backed on the array. Poor performance in destaging data from the controllers indicates a need to increase the amount of drives in the array.

It could also be an indication that the incorrect CPG was chosen for the I/O load. A write-intensive load to nearline drives would increase overhead on the array. A better solution would be to use FC drives or SSDs.

Another solution would be to use AO to increase performance during peak periods of I/O or isolate the workload, which has overdriven the array, and move to another storage array.

Solution 2: Another possible solution is to look through previous reports. A best practice could be to monitor important processes such as Exchange, Oracle, SQL, or some other high performing, real time applications. Create the following reports and compare values.

Page 26: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 26

Create the following charts:

• Export Volumes Performance—customize for Exchange as an example

• Export Volumes Performance—customize for a database

• Host Port Performance statistics—customize to include disk ports

• Host Port Performance statistics—customize to include only host ports

• Port data real time—select objects and then plot an individual series for each of the selected objects

There are various other reports, which can be generated. Using SR, the user can isolate a time when the alerts are triggered and then it is correlated to a specific action on the array through charts.

System Reporter database With the introduction of HPE 3PAR OS version 3.1.2, an instance of the System Reporter Software is by default installed on the HPE 3PAR StoreServ system, and automatically starts collecting performance data from the local system. System Reporter data resides on a local volume created on the array. The virtual volume “.srdata” is created during array initialization in the admithw phase. Table 2 illustrates the defined database size for currently sold arrays. The .srdata VV is mounted on the non-master node. The size of the .srdata is 60 GB for 2-node systems, 80 GB for 4-node systems, and 100 GB for 8-node systems. This is true across all current HPE 3PAR StoreServ platforms.

Table 2. Array database size

Storage array Database size

HPE 3PAR StoreServ 72xx or 82xx 60 GB

HPE 3PAR StoreServ 74xx or 84xx 80 GB

HPE 3PAR StoreServ 10400 or 20400 80 GB

HPE 3PAR StoreServ 10800 or 20800 100 GB

Database retention Database retention is dependent upon array size, number of objects monitored, and the size of each object. A user can view the .srdata base retention either from the HPE 3PAR cli using the showsr command or on the SSMC. An example of retention timeframe is shown in figure 28.

Data retention estimates as viewed in figure 28 are projections based on amount of space currently consumed with a known start time. Each category has a predefined allocation of the SR data volume. Hi-res, hourly, and daily data consume 14 percent of the volume. AO consumes 30 percent of the volume due to the complex algorithm to age data out. Each category shown has multiple files in the category, as one file ages we move that data out.

Figure 28. System Reporter data retention

Page 27: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 27

Looking at customer requirements, most customers keep their data retention as 14 days (hi-res), six months of hourly data, and one year of daily data. Based on telemetry data, the out-of-the-box .srdata size will satisfy the majority of use cases. Future releases will enable the user to extend the size of the .srdata volume to extended retention times as needed. Table 3 has examples of retention analysis.

Table 3. Data retention samples

Customer VLUNs VVs Hi-res Hourly Daily

Example 1 3000 5080 14 days 6 months 11 years

Example 2 1337 5606 16 days 6 months 12 years

Collection of data samples Data sampling runs continuously but reports only reported at different intervals. Time intervals used in charting data are as follows:

• Hi-res: Data is displayed once every 5 minutes

• Hourly: Data is charted hourly

• Daily: Data is charted once a day

Charted data displayed for each of the SR data samples are output because of counter deltas. Using the variables of v = collected data and t = time collection, we calculate the displayed values for each of the time intervals as follows. This is illustrated in Figure 29. SR data point calculations.

• Hi-res data is (v2-v1)/(t2-t1)

• Hourly is calculated from point-in-time values (v12-v1)/(t12-t1)

• Daily is calculated from point-in-time values (v288-v1)/(t288-t1)

Figure 29. SR data point calculations

Real time data samples The display of real time data is shown and updated at a 5-second interval. SR retains 30 minutes of real time data and is available on the array’s local cache. Real time data can display the last 1000 collection points. When the user views real time data and navigates off the real time screen for more than 30 minutes, the user will be required to start the collection again.

Page 28: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 28

Grow .srdata—new in SSMC 2.2 HPE 3PAR OS 3.2.2 allows the user to grow the .srdata base to retain a larger data collection. Data growth is accomplished by using an HPE 3PAR CLI command. The format of the command is as follows:

cli% controlsr grow <size> or controlsr grow <percentage>

Grow .srdata is an example of the .srdata base prior to the issuance of controlsr grow command and the results after the command is issued. Examine the difference in the database size and estimation of data retention between the two captures.

Prior to growth

After growth

Figure 30. Grow .srdata

Backing up .srdata—New in SSMC 2.2 The ability to back up the .srdata database is a new option as part of HPE 3PAR Host OS 3.2.2. The backup is accomplished by using a createsv command from the cli command prompt. The creation of the snapshot can only be accomplished from the cli and not from the SSMC GUI. If the user tries to create the snapshot from the GUI, a message will be displayed indicating the volume is part of the system and cannot be created.

Figure 31. Backup of .srdata

The example in figure 30, creates a read-only snapshot of the .srdata database labeled srdata_backup. Once the snapshot is created the user can, if needed, promote the virtual copy and restart the database using the promoted snapshot. This would be done only if the current .srdata is corrupted in some manner.

Page 29: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 29

Exporting the .srdata—New in SSMC 2.2 The user can export data from the .srdata database through the %cli interface starting with HPE 3PAR StoreServ OS 3.2.2. The command structure to export this data can be found by either using the help command in the HPE 3PAR OS CLI or referencing the HPE 3PAR Command Line Interface Administrator’s Manual HPE 3PAR OS 3.2.2.

The export command allows the user the ability to download all or part of the .srdata database to either the service processor or a local machine, which has logged in via the HPE 3PAR CLI tool. An example of a download to a local account is captured in Figure 32. Export .srdata. The example in the format uses the following switches and options.

Required switch—one of the following switches must be included:

• -highres, -hourly, -daily, -ldrg (LD region stats used by AO) any of these switches can stand by themselves

• -ldrgsum (concise summary of LD region stats)

• -all (includes hires, hourly, daily, ldrg)

Options—options to be included in export of data

• -csv (option to export to .csv format, can be opened with MS Excel) file is SQLite .db file

• -save –file <location> allows the output to be copied to the remote cli client machine (if not used SP is required to send file to HPE 3PAR for processing)

• -btsecs and –etsecs to specify time range

Figure 32. Export .srdata

Figure 32. Export .srdata is an example of the command used to export all categories of the database for the last day. It is in a .csv format at a file location on the c: directory of the local machine. In this example, a local SSMC server was connected to the StoreServ array. On the local server, a directory titled “demo” was created to capture the file contents of the download. The directory demo on the local SSMC server contains the file titled “exportsr_all_1414091_20150709_112225_CDT.tbz”, which was exported from the StoreServ.

The downloaded file can then be processed on the local SSMC server or copied to another system to process the .csv files. Figure 33. Uncompressed export data—all shows the information, which was collected with the export command. Each of the file folders contains .csv files of various commands; for example, statvlun.csv is available within the daily directory.

Figure 33. Uncompressed export data—all

Page 30: HPE 3PAR System Reporter with Performance and capacity

Technical white paper Page 30

Examining data with HPE 3PAR Excel plugin—updated in SSMC 2.2 Included with SSMC 2.1, as a separate plugin, is an Excel spreadsheet plugin, which can report on data from the HPE 3PAR StoreServ .srdata database. This separate installable option is available via HPE Software depot. The plugin extracts data using the SSMC server and contains a number of predefined Excel spreadsheets. The plugin is divided into two separate sections of performance data and configuration data. An example of an output is described in Figure 34. Excel plugin for SSMC. In this illustration, a report was generated on physical drive activity as collected over the last hour and reported as hi-res data.

Figure 34. Excel plugin for SSMC

The following categories are reportable using the Excel spreadsheet. Along with these categories, data can also be reported using hi-res, hourly, and daily resolutions.

• Physical drive performance

• Port performance

• Exported volume performance

• Physical drive capacity

• CPG capacity

• VV capacity

• Physical drive histogram

• Exported volume histogram

• Port histogram

The plugin tool extracts and displays data on configuration data. The following configuration data can be displayed using the plugin tool. This data can be saved for future reference on configuration data of the array.

• Virtual Volumes

• Hosts

• Common Provisioning Groups

Page 31: HPE 3PAR System Reporter with Performance and capacity

Technical white paper

Sign up for updates

Rate this document © Copyright 2013, 2015–2016 Hewlett Packard Enterprise Development LP. The information contained herein is subject to change without notice. The only warranties for Hewlett Packard Enterprise products and services are set forth in the express warranty statements accompanying such products and services. Nothing herein should be construed as constituting an additional warranty. Hewlett Packard Enterprise shall not be liable for technical or editorial errors or omissions contained herein.

Oracle is a registered trademark of Oracle and/or its affiliates.

4AA5-0003ENW, January 2016, Rev. 3

Terminology QoS: Quality of Service

MC: Management Console

PD: Physical Disk

SR: System Reporter

VV: Virtual Volumes

ESR: External System Reporter

Hi-res: High-resolution Collection

RC: Remote Copy

Learn more at hp.com/go/3PAR