Service-oriented storage management using IBM Spectrum Control
As a storage provider, you are confronted with managing the constantly changing requirements regarding capacity growth and performance optimization, which are necessary to provide appropriate resources for the business needs of your customers at minimum cost.
This chapter describes how IBM Spectrum Control (Spectrum Control) helps you to accomplish the following tasks:
5.1 Getting insights into how your storage is consumed
The new consumer-oriented dashboard design of Spectrum Control that is shown in Figure 5-1, which shows the Dashboard of Spectrum Control, with the Storage Systems pane highlighted, provides a first glance at your storage systems. Underneath each type of storage (Block, File, and Object Storage) is a horizontal bar chart that provides details about the used and free space and the condition of the monitored storage systems in your environment.
Figure 5-1 Dashboard Design of Spectrum Control
In the upper left pane of the Dashboard, you can see the number of monitored Storage systems. Below that, you can see the number of storage systems that are not in a normal state, which are grouped by status. For example, in Figure 5-1, 13 block storage systems have the status “error”, one is “unreachable,” and one has the “warning” status.
|
Reference: For more information about the statuses of monitored resources, see the Spectrum Control IBM Knowledge Center, found at:
|
Review the Spectrum Control IBM Knowledge Center in the “What’s new” section of future versions for further improvements of the Dashboard at the following website:
5.1.1 How to mitigate the risk of your pool running out of space
In addition to the performance charts, Spectrum Control offers capacity charts that show the growth of pools over time. Figure 5-2 shows an example of how you can initiate viewing historical capacity charts for pools.
Figure 5-2 Start the historical capacity trends for pools
If you select Capacity in Figure 5-2, the pools with the highest Available Pool Space are shown automatically in the chart. If you want to see specific pools, select the pools and click View Capacity in Figure 5-2. The capacity chart opens in a new browser window.
Figure 5-3 shows the historical capacity of the fb_flash_1 pool.
Figure 5-3 Example of a historical capacity chart
Spectrum Control gives you a projection of when your pools run out of space. Based on linear regression of the last 180 days, Spectrum Control calculates when a pool runs out of space (assuming that the capacity behavior does not change drastically). See the Zero Capacity metric, which is shown in Figure 5-4.
Figure 5-4 Example of the Zero Capacity metric in Spectrum Control
Spectrum Control calculates a trend line for the existing data by using the least square method to find a line that fits the existing data. This trend line’s equation is then used to calculate the Zero Capacity Metric as a forecast.
The Zero Capacity metric is available on all pool window, but it is not shown by default. Right-click the table header row and select the Zero Capacity metric. You can sort the Zero Capacity Column in an ascending order so that it shows you the pools that will run out of space at the top of the column.
5.1.2 Monitoring your Easy Tier pools
Spectrum Control shows you at a glance what kind of storage of which your pool is composed. With this information, you can easily determine how much space is left in each of the tiers of your IBM Easy Tier® pools or you can watch how Easy Tier is performing.
Easy Tier in IBM SAN Volume Controller and IBM Storwize family
Implementing the IBM Storwize V5000, SG24-8162 describes the Easy Tier operating modes and the resulting Easy Tier statuses that are available in IBM SAN Volume Controller and IBM Storwize family.
Easy Tier operating modes
Here is a list of Easy Tier operating modes in IBM SAN Volume Controller and the IBM Storwize family for MDisk groups (pools):
•Off: No statistics are recorded and no extents are moved.
•On: When the Easy Tier function is turned on, Easy Tier measures the I/O activity for all extents. When you have a multi-tiered pool, the extents are migrated dynamically by the Easy Tier processes to achieve the preferred performance.
•Measured: Easy Tier measures the I/O activity for all extents, but does not move any extents in the storage pool.
•Auto: Enables Easy Tier functions when the storage pool contains MDisks from more than one tier, and enables automatic rebalancing when the storage pool contains MDisks from only one tier.
Easy Tier status in IBM SAN Volume Controller and IBM Storwize family
Depending on the Easy Tier mode attributes, a storage pool can have one of the following Easy Tier statuses:
•Active: This status indicates that Easy Tier is actively managing the extents of the storage pool.
•Balanced: This status applies to homogeneous storage pools and indicates that Easy Tier is actively managing the extents to provide enhanced performance by rebalancing the extents among the MDisks within the tier.
•Measured: This status means that Easy Tier is constantly measuring the I/O activity for all extents to generate an I/O statistics report.
•Inactive: When the Easy Tier status is inactive, no extents are being monitored and no statistics are being recorded.
Spectrum Control shows the operation mode and status for Easy Tier pools in IBM SAN Volume Controller and IBM Storwize family, as shown in Figure 5-5.
Figure 5-5 Easy Tier mode and status
Table 5-1 shows the mapping of SAN Volume Controller mode and status to the values that are shown in Spectrum Control. Check the Spectrum Control IBM Knowledge Center for changes in future versions.
Table 5-1 Mapping of SAN Volume Controller Easy Tier mode and status to the values that are shown in Spectrum Control
|
SAN Volume Controller Easy Tier Mode
|
SAN Volume Controller Easy Tier Status for a single tier
|
SAN Volume Controller Easy Tier Status for multitier
|
Spectrum Control value for single tier
|
Spectrum Control value for multitier
|
|
Auto
|
Balanced
|
Active
|
Auto/Active
|
Auto/Active
|
|
On
|
Balanced
|
Active
|
Enabled/Inactive
|
Enabled/Active
|
|
Off
|
Inactive
|
Inactive
|
Disabled
|
Disabled
|
Easy Tier control in IBM System Storage DS8000
This section describes how to use the Easy Tier monitor control function in a DS8000 storage system.
Easy Tier Monitor Mode control in a DS8000 storage system
The Easy Tier monitor control setting can be set to automode, all, or none, which refers to the volumes to be monitored. The default is automode.
The following settings are available:
•Automode: Extent statistics for logical volumes in managed extent pools under control of Easy Tier Automatic Mode are updated to reflect the current workload activity.
•All: Extent statistics for all logical volumes in managed and non-managed extent pools are updated to reflect the current workload activity.
•None: Extent statistics collection is suspended.
IBM DS8000 Easy Tier, REDP-4667 describes the DS8000 Easy Tier concepts, design, and implementations.
Easy Tier Automatic Mode control in a DS8000 storage system
As a prerequisite, monitor mode is set to all, which is preferable.
This control can be set to tiered, all, or none, and defaults to tiered.
The following settings are available:
•Tiered: Automatic extent migrations that are performed by Easy Tier Automatic Mode are enabled only for multitier or hybrid pools.
•All: Automatic extent migrations that are performed by Easy Tier Automatic Mode are enabled for hybrid (multitier) and homogeneous (single-tier) pools.
•None: Easy Tier Automatic Mode is disabled and all automatic extent migrations are deferred for all the extent pools.
Table 5-2 shows the mapping of DS8000 Monitor Mode and DS8000 Automatic Mode to the values that are shown in Spectrum Control.
Table 5-2 Mapping of DS8000 Easy Tier Monitor and Automatic Mode to Spectrum Control values
|
DS8000 Monitor Mode
|
DS8000 Automatic Mode
|
Spectrum Control value for single tier
|
Spectrum Control value for multiple tiers
|
|
Automatic Mode
|
Tiered
|
Tiered Pools/No
|
Tiered Pools/Yes
|
|
Automatic Mode
|
All
|
All Pools/Yes
|
All Pools/Yes
|
|
Automatic Mode
|
None
|
Disabled
|
Disabled
|
|
All
|
All
|
All Pools/Yes
|
All Pools/Yes
|
|
All
|
Tiered
|
Tiered Pools/No
|
Tiered Pools/Yes
|
|
All
|
None
|
Disabled
|
Disabled
|
|
None
|
All
|
Disabled
|
Disabled
|
|
None
|
Tiered
|
Disabled
|
Disabled
|
|
None
|
None
|
Disabled
|
Disabled
|
Table 5-3 shows a summary of possible Easy Tier values in Spectrum Control for DS8000 Easy Tier pools.
Table 5-3 Spectrum Control Easy tier values for DS8000 pools
|
Spectrum Control Easy Tier values for DS8000 pools
|
Explanation
|
|
Tiered Pools / Yes
|
Easy Tier is enabled for multitiered pools only and there is more than one tier in the pool.
|
|
Tiered Pools / No
|
Easy Tier is enabled for multitiered pools only and there is only one tier in the pool.
|
|
All Pools / Yes
|
Easy Tier is enabled for multitier and single-tier pools.
|
|
Disabled
|
Either Easy Tier mode or Monitor Mode is disabled.
|
Figure 5-6 shows an example of Easy Tier values of DS8000 pools in Spectrum Control.
Figure 5-6 Easy Tier values of DS8000 pools in Spectrum Control
Easy Tier distribution shown in Spectrum Control
Spectrum Control shows the distribution and the capacity values of the volume extents across each Easy Tier pool for volume extents on solid-state drives (SSDs) or flash modules, on enterprise hard disk drives (HDDs), and on nearline HDDs.
In Figure 5-7, Spectrum Control shows the capacity values for up to three different types of extents in a pool.
Figure 5-7 Example of the tier distribution of a pool
These metrics might not be shown by default. If not, right-click the table header row and select the metrics.
The second, third, and fourth columns show the total amount of storage space for each different type of extent that can be used by Easy Tier for retiering the volume extents in the pool.
The fifth column shows the tier distribution of the total amount of storage space in percentage (%).
The sixth, seventh, and eight columns show the amount of storage space that is available for each different type of extent that can be used by Easy Tier for retiering the volume extents in the pool.
The volumes table (Figure 5-8) shows the tier distribution. If you want to know how tiers are distributed in the volumes of an Easy Tier pool, open the pool’s Properties window and click the Volumes tab.
Figure 5-8 shows the volume’s Easy Tier distribution.
Figure 5-8 Volume’s Easy Tier distribution
5.1.3 How to mitigate the risk of overprovisioning
If you are concerned that you will run out of space because of overprovisioning, there are some ways that Spectrum Control can help you to prevent this situation:
•Computed metrics: Virtual Allocation Percentage and Shortfall Percentage
•Alerts and thresholds
•Restriction settings in Service Class Definitions
Virtual Allocation Percentage and Shortfall Percentage
This section describes how the metrics Shortfall Percentage and Virtual Allocation Percentage are calculated:
The Virtual Allocation Percentage is defined as follows:
The percentage of physical space in storage virtualizer pools that is committed to the total virtual capacity of the volumes in the pool. In thin-provisioned environments, this percentage exceeds 100% if a pool is overcommitted (over-provisioned).
Virtual Allocation Percentage = (Total Volume Capacity ÷ Pool Capacity) * 100
Shortfall Percentage is defined as follows:
The percentage of the remaining unallocated volume space in storage virtualizer pools that is not available to be allocated (= “Unallocatable Volume Space “).
For more information about these definitions, see the Storage Systems Capacity Report topic in the Spectrum Control V5.2.8 IBM Knowledge Center, found at:
Figure 5-9 shows the Unallocatable Volume Space.
Figure 5-9 Unallocatable Volume Space, Unallocated Volume Space, and Available Pool Space
Shortfall Percentage = (Unallocatable Space ÷ Unallocated Volume Space) * 100, and Unallocatable Volume Space = Total Volume Capacity - Allocated Space - Available Space
The Shortfall Percentage cannot exceed 100%.
In summary, Unallocatable Volume Space is the space that you must add to your pool so that all thin-provisioned volumes can grow to their maximum size, and Shortfall Percentage is the relationship of the Unallocatable Space to the Unallocated Volume Space.
Figure 5-10 on page 123 explains the pool metrics that are needed to calculate Shortfall Percentage.
Figure 5-10 Pool metrics
The pool metrics that are shown in Figure 5-10 are displayed in Spectrum Control.
Figure 5-11 shows an example of a pool with thin-provisioned volumes.
Figure 5-11 Example for the Shortfall Percentage metric - thin-provisioned volumes
The following formulas show how the Unallocated Volume Space, Shortfall Percentage, and Virtual Allocation Percentage metrics from Figure 5-11 are calculated in Spectrum Control:
•Unallocatable Volume Space = 5215.50 GiB - 982.5 GiB- 467 GiB = 3766 GiB
•Shortfall Percentage = 3766 GiB ÷ 4233 GiB * 100 = 89%
•Virtual Allocation Percentage = (5215.50 GiB ÷ 1480 GiB) * 100 = 352%
|
Note: If Virtual Allocation Percentage is less than 100% Shortfall Percentage will be 0%, and there is no risk. When Virtual Allocation Percentage is greater than 100% Shortfall Percentage increases, and there is a risk of running out of space.
|
For more information about this topic, see the Spectrum Control IBM Knowledge Center:
Set alerts to mitigate the risk of overprovisioning
For the following parameters, you can configure alerts to mitigate the risk of overprovisioning:
•On the pool level:
– Virtual Allocation %
– Pool Shortfall %
– Available Repository Space for DS8000 track space-efficient (TSE) Pools
– Unallocated Volume Space
– Unallocatable Volume Space
– Virtual Capacity Limit
•On the volume level: Unallocatable Space
Available Repository Space is defined as the unallocated storage space in the repository for TSE thin-provisioned Volumes. This value is available only for DS8000 thin-provisioned pools.
Virtual Capacity Limit is defined as the maximum amount of virtual storage space that is available to allocate to volumes in the storage pools that are associated with the storage system. You can use this alert to be notified about space-efficient volumes.
For more information, see Chapter 6, “Advanced alert management” on page 177.
Using Spectrum Control service classes to control overprovisioning
You can control the level of overprovisioning with Spectrum Control when you use the service class attribute Overallocation Limit % to control your overprovisioning, as shown in Figure 5-12.
Figure 5-12 Service class property Overallocation Limit %
If an overallocation limit is specified, Spectrum Control refines the list of candidate storage of a specified service class to include only the storage pools that have an overallocation percentage that does not exceed the overallocation limit.
The overallocation percentage for a pool is determined by using the following measurements, as shown in Figure 5-13, which shows the Pools window with the Virtual Allocation percentage and the Soft Space parameter:
•For an XIV system storage pool, the Soft Space is divided by the Capacity.
•For other storage system pools, use the Virtual Allocation.
Figure 5-13 Virtual Allocation percentage and the Soft Space parameter
This parameter helps you control, for each service class, the amount that Spectrum Control overprovisions your physical capacity.
For more information about using service classes with Spectrum Control, see IBM SmartCloud Virtual Storage Center, SG24-8239.
5.1.4 Creating input for your chargeback report
Spectrum Control offers several options that help you better control the consumption of storage and get the necessary data for charging consumers based on their usage.
Spectrum Control is not meant to provide invoicing capabilities, so it cannot provide a chargeback report. But, Spectrum Control is aware of all the relevant data:
•Capacities
•Entity attributes
•Host mappings
•Spectrum Control applications
•User-defined properties
Therefore, Spectrum Control can be used to generate input for your chargeback report. Because input for a chargeback report is not yet considered a chargeback report, call the input report a showback report.
Considerations
Customer feedback shows that each customer has his own requirements that must be included in a showback report. Customer requirements depend on whether they are a cost or a profit center, or some of them might want to show their management an analysis of the IT costs because of each consumer.
To create a showback or a chargeback report, think about how you are going to meter, rate, group, and bill the costs. This section lists considerations for each of these categories. Depending on your environment, there are additional items. Here are the considerations:
•Metering
•Rating
•Grouping
•Example of a showback report instead of billing
Metering
The following prerequisites are required because the capacity influences the charges and functions like encryption (see “Rating”) influence the charges. Some of the information is stored with the volume, and other information is stored with the pool or even with the storage system. Data must be available when the report is generated.
•Devices must be configured in Spectrum Control and probes must run regularly.
•You might want to use custom tags, for example, for each pool in Spectrum Control to define costs per capacity or to define the usage of the pool.
•You can specify Spectrum Control tiers as needed.
•You can specify and use service classes to provision volumes with Spectrum Control.
•You can configure Spectrum Control applications and departments, as described in 5.2, “Organizing capacity data around applications and business units” on page 145.
Rating
You must define which entity on which you want to base your rating:
•Volume Capacity, which includes the following functions:
– Easy Tier
– Thin provisioning
– Compression
– Encryption
– RAID level
•Spectrum Control: Tier
•Spectrum Control: Service classes
You also must determine when and how to meter the following items:
•The cutoff date, for example, the first day of a month is easier to establish, but needs definitions for a month with 28, 29, 30, and 31 days.
•Average usage over a specified time frame.
•Different rates may be applied if your data center is managed as a cost center versus being managed as a profit center.
•You might want to apply special rates for specific customers, for example, the more capacity they buy, the cheaper the costs per GB, or for customers within or external to your organization.
Grouping
Grouping can be done by using one of the following methods:
•Using the Spectrum Control Grouping function, which is described in 5.2, “Organizing capacity data around applications and business units” on page 145.
When you use the Spectrum Control Grouping function, consider the following items:
– Usage of naming conventions.
– Only application groups or other departments can be added to departments.
– Avoid double counting: Costs for shared resources can be automatically split between the groups.
•Naming conventions only.
Example of a showback report
For an example script that generates a showback report, see Appendix B, “Reporting scripts” on page 243.
|
Note: The script is also available on IBM developerWorks in the IBM Tivoli Storage Productivity Center wiki under Custom Reporting, found at:
That website is a good place to share custom reports.
|
The showback report that is shown in Appendix B, “Reporting scripts” on page 243 is provided as an example only and is highly dependent upon your environment (there is no guarantee that it will work for your environment and your requirements), but you might want to use this script to get started and modify it for your own needs. SQL skills are required.
The example is written by using native SQL and can be rendered by using the following options:
•Use native SQL in a batch report with the option of automating. This action generates a CSV file that can be used for further processing.
•IBM Cognos BI is required.
•Use Microsoft Excel. ODBC connectivity is required.
|
Note: If you have not installed the Cognos BI Reporting tool yet, you can do so at any time. For installation guidelines, see the following website:
|
Here are the limitations of the example script to create a showback report:
•Works only with block storage.
•Only tested on IBM systems.
•Does not include IBM FlashCopy®, Metro Mirror, Global Mirror, or Global Mirror with Change Volume relationships.
•Shows capacities (no historical data) so it can be used on a fixed date, for example, to run on the first day of every month. (See the information about the cutoff date in “Rating” on page 127).
•User-defined properties, such as custom tags, are string fields. To use them for calculations, they must be converted into a numerical value.
•This report is based on volumes and lists host connections (and not servers or hypervisors). Volumes that are assigned to the same hypervisor or server but to different HBAs on the same server or hypervisor are assigned to different host connections.
|
Note: If you use this report as an import to create other reports, make sure that you avoid double counting. As a preferred practice, create verification reports.
|
This Accounting report (one row per volume) provides the following data:
•VOLUMENAME
•VOLID
•HOSTLIST
•CAPACITY
•ALLOCATED_SPACE
•USED_SPACE
•POOLNAME
•POOL_CUSTOM_TAG1
•POOL_CUSTOM_TAG2
•POOL_CUSTOM_TAG3
•STORAGE_SYSTEM
•THINPROVISIONED
•ENCRYPTABLE
•ENCRYPTED
•MIRROR_COUNT
•STATUS
•RAID_LEVEL
•IS_COMPRESSED
•SPECTRUMCONTROL_TIER
•POOL_EASYTIER
•POOL_EASYTIER_STATUS
•STORAGESUBSYSTEM_CUSTOM_TAG1
•STORAGESUBSYSTEM_CUSTOM_TAG2
•STORAGESUBSYSTEM_CUSTOM_TAG3
•APPLIST
•VOL_TIER_CAPACITY
•CURRENT_DATE
Figure 5-14 shows some of the fields that you receive from the script.
Figure 5-14 Example of a showback report output
Figure 5-15 also shows some additional columns and fields that are received from the example showback report output.
Figure 5-15 Additional columns of the example showback report output
The script’s result shows volumes that are mapped to a host (assuming that only volumes with host connection are used for an import to a charge back report). If you want to see volumes with no host connections, replace the “inner join” text that is shown in Figure 5-16, which shows how to change the script to also get volumes with no host connections, with the “left outer join” in “t1”.
Figure 5-16 Code snippet of the SQL query
The script shows the volumes and their VDisk mirrors. To get only the primary volumes, remove the two dashes in front of AL6.IS_PRIMARY=1 at the end of the temporary table t4 (Figure 5-17, which shows how to change the script to get only primary volumes).
Figure 5-17 Another SQL code snippet
Then, you might want to see the Mirror_Count. Therefore, remove the dashes, as shown in Figure 5-18 on page 131.
Figure 5-18 A third SQL code snippet
Using native SQL in a batch report with the option of automating
The following section describes the steps to use native SQL in a batch report to automate the script to run on a regular basis.
Complete the following steps:
1. Obtain the showbackexport.sql and runshowbackexport.bat scripts, which can be found in Appendix B, “Reporting scripts” on page 243.
2. If you want to use these scripts outside of your Spectrum Control Server, then you must configure your database in the IBM DB2 client first, as described in “Instructions for creating a rollup report” on page 138.
3. Store the files in a location without a blank in the path, for example, to C:showback.
4. Update the connection, path details, and the credentials in runshowbackexport.bat.
5. Update the path in runshowbackexport.bat, if necessary.
6. Now you can run or schedule runshowbackexport.bat to obtain a CSV file as an import for your chargeback on a regular basis.
|
Reference: The showbackeport.sql script uses the DB2 EXPORT command. For more information about this command, go to the following website, where you find information about how to specify the delimiter and to format data:
|
Using reports in IBM Cognos Business Intelligence
For basic instructions about how to work with Cognos Business Intelligence (BI) Reports within Spectrum Control, see IBM Tivoli Storage Productivity Center V5.2 Release Guide, SG24-8204.
To use reports IBM Cognos BI, complete the following steps:
1. Obtain the showback_Cognos_xml.txt file, which is found in “Sample script for Cognos” on page 253.
2. Copy the whole script into your clipboard.
3. Start Cognos Report Studio.
4. Choose any package, for example, Spectrum Control Capacity and Relationship Package.
5. Create a blank report.
Figure 5-19 Open Report from Clipboard function
7. Save, run, or schedule your report.
|
Note: Using IBM Cognos Report Studio gives you the opportunity to schedule the report. You also can have the report output automatically sent to specified recipients or saved to the file system.
|
If you are using Spectrum Control, or a Tivoli Storage Productivity Center version earlier than Version 5.2.8. you might need to import the showback.sql file, which can be found in “Sample script with native SQL statement” on page 244, into an SQL Query. To do so, see Figure 5-31 on page 142 and steps 3a on page 142 to 3l on page 145 in “Instructions for creating a rollup report” on page 138.
Using Microsoft Excel
Databases are excellent for storing information, but sometimes you want to use or analyze the data in Excel. You can use the Data Connection wizard to create a dynamic connection between a DB2 database and your Excel workbook. To accomplish this task, complete the following steps.
|
Note: The following description is based on Microsoft Excel 2013 and might be slightly different for other versions of Microsoft Excel.
|
1. Obtain the showback.sql script, which is found in “Sample script with native SQL statement” on page 244.
2. Configure the Spectrum Control database TPCDB in the DB2 client, as described in “Instructions for creating a rollup report” on page 138
3. Set up your ODBC connection to the Spectrum Control database TPCDB. For Microsoft 64-bit systems, see the following website:
4. Open a blank workbook in Microsoft Excel.
5. Click the Data tab, and then click From Other Sources → From Data Connection Wizard, as shown in Figure 5-20.
Figure 5-20 How to configure a DB2 database in Microsoft Excel
6. Click Other/Advanced and then Next.
7. In the next window, select IBM OLE DB Provider for DB2 - DB2COPY1.
8. Click the Connection tab of the Data Link Properties window. Select Existing data source and select TPCDB171 from the drop-down menu. Enter the correct credentials, and click Test Connection, as shown in Figure 5-21.
Figure 5-21 Specify the Spectrum Control TPCDB in Microsoft Excel
9. Clear Connect to a specific table in the Data Connection wizard, as shown in Figure 5-22 on page 135. Click Next.
Figure 5-22 Select Database and Table in Microsoft Excel
Figure 5-23 An example for specifying the Data Connection file name
11. If prompted, enter the credentials again.
12. Without selecting a table, click OK in the Select Table window.
Figure 5-24 Import Data window
14. In the Connection Properties window (Figure 5-25 on page 137), click the Definition tab, and type SQL in to the Command type field. Copy and paste the content of the showback.sql file in to the Comment field.
Figure 5-25 Specify the SQL statement in Microsoft Excel
15. Click OK twice. If prompted, add your credentials again.
The showback report is now shown in Microsoft Excel. You might want to save this report. The next time that you open the report, click the Data tab, and click Refresh.
5.1.5 Rollup reports across different Spectrum Control servers
This section describes the scenario where you manage more than one Spectrum Control server and want to get one report with data from multiple Spectrum Control servers.
This task is possible by using the Report Studio of the Cognos Business Intelligence (Cognos BI) Reporting tool. The Cognos BI reporting tool is available as an optional component of Spectrum Control. If you have not installed the Cognos BI Reporting tool yet, you can do so at any time. For installation guidelines, see the Spectrum Control IBM Knowledge Center, found at:
In Spectrum Control, Cognos BI uses the local database (TPCDB) as a data source for creating reports. The configuration of the local TPCDB database (and in DB2 and Cognos) is done during installation. The basic idea is to use the Spectrum Control server that has Cognos installed to use the local TPCDB and TPCDBs from other Spectrum Control servers as data sources. TPCDBs from other Spectrum Control servers must be configured manually to be accessed by DB2 and Cognos BI.
Depending on how you configure remote databases (TPCDBs) to Spectrum Control Cognos BI, you can use them differently:
•Configure remote databases so that users can choose which database they want to run the report against. In this case, they are prompted to choose the database when running the report. With this configuration, predefined reports or custom reports that use the Spectrum Control Packages can be used to run against local and remote databases. With this configuration, users can select only one database per report.
•Configure remote databases so that you can define a database per query. You can use this configuration to create and use a union query to combine multiple queries that can run against different databases in the same report. The advantage is that one report can contain data from multiple databases, so this method can be used to create a rollup report. The disadvantage is that you cannot use the predefined reports or the Spectrum Control Packages. Therefore, you must use native SQL for creating queries against multiple databases.
For more information about these setups, see the following website:
Instructions for creating a rollup report
This section describes how to configure a second Spectrum Control database and run a report across two Spectrum Control databases. You must repeat these steps to configure more Spectrum Control databases.
1. Configure your remote database.
On the Spectrum Control server where you want to run the report, you must add the remote TPCDB database to the system database catalog and assign an alias to the database. For example, you can catalog the remote TPCDB database as tpcdb02.
For more information about how to configure your database, see the following websites:
Example 5-1 shows the commands that are used to configure your database.
Example 5-1 Configure your database by using the db2cmd command-line interface
db2 catalog tcpip node tpc02 remote myserver server 50000
db2 catalog database tpcdb as tpcdb02 at node tpc02
If you run db2 list node directory, you should see your newly configured database tpcdb02,
2. Next, you must configure your database in Cognos BI:
a. In the Spectrum Control web-based GUI, click Cognos → View predefined reports.
b. Click Launch → IBM Cognos Administration, click the Configuration tab, and click Data Source Connections (make sure that the path shows Directory → Cognos, as shown in Figure 5-26 on page 139).
Figure 5-26 Add a data source in Cognos administration
d. Enter the alias that you created in step 1 on page 138 in to the Name field, and then click Next.
Figure 5-27 New Data Source wizard to specify a DB2 connection
g. Enter the alias that you created in step 1 on page 138 in to the DB2 database name field.
h. Select Signons, select the Password check box, and then enter a User ID and Password for the remote database, as shown in Figure 5-28.
Figure 5-28 Provide credentials for a new data source connection
i. Click Test the connection, and then click Test. Verify that the status of the connection is Succeeded, as shown in Figure 5-29 on page 141.
Figure 5-29 Result of the connection test
j. Click Close. Click Next, and then click Next again. Click Finish.
k. Check that TPCDB02, shown in Figure 5-30, now shows in the window that opens when you click IBM Cognos Administration, → Configuration tab → Data Source Connections. (Ensure that the path shows Directory → Cognos.)
Figure 5-30 shows the newly configured remote TPCDB02 (compare to Figure 5-26 on page 139).
Figure 5-30 Data Source Connections
3. Create your Cognos BI Report from multiple Spectrum Control instances by using Report Studio.
For basic instructions about how to work with Cognos BI Reports within Spectrum Control see IBM Tivoli Storage Productivity Center V5.2 Release Guide, SG24-8204.
Complete the following steps:
a. Start Cognos Report Studio.
b. Choose any package, for example, Spectrum Control Capacity and Relationship Package, and create a blank report.
c. Open Query Explorer, as shown in Figure 5-31.
d. Drag the SQL Item from the Toolbox to the right of your Query1.
e. Select the SQL item and set the SQL Syntax Property to Pass-Through. Click the three dots next to the Data source field ad select the local TPCDB as the data source, as shown in Figure 5-31.
Figure 5-31 SQL properties
f. Click OK.
g. Double-click the SQL item. Insert the query statement that is shown in Example 5-2 on page 143, which gives you a list of storage virtualizer pools and their capacity, in which “SC1” refers to your local Spectrum Control server.
Example 5-2 SQL query
SELECT
'SC1' as SCserver
, AL2.DISPLAY_NAME as StorageSubsystemName
, AL1.DISPLAY_NAME as StoragePoolName
, AL1.STORAGE_POOL_SPACE/ (1024*1024*1024) as Capacity
, AL1.REAL_SPACE/ (1024*1024*1024) as Allocated_Space
, AL1.STORAGE_POOL_AVAILABLE_SPACE/ (1024*1024*1024) as Available_Space
, AL1.STORAGE_POOL_CONSUMED_SPACE/ (1024*1024*1024) as Used_Space
FROM TPCREPORT.STORAGEPOOL AL1, TPCREPORT.STORAGESUBSYSTEM AL2
WHERE (AL2.SUBSYSTEM_ID=AL1.SUBSYSTEM_ID) AND (AL2.DISPLAY_NAME LIKE '%SVC%')
with ur for read only
|
Note: Always specify with ur for read only in your SQL queries. Otherwise, your tables might get locked during the read operation, which might slow down the performance of your TPCDB.
For more information about this topic, see the following websites:
|
h. Copy and paste Query1 so that you now have two queries. Query2 queries your second Spectrum Control server that you configured to be used remotely.
i. Change the properties of the second SQL item to TPCDB02. Also, change “SC1” as SCserver to “SC2” as SCserver, as shown in Figure 5-32. (“SC2” is your remote Spectrum Control server.)
Figure 5-32 Rollup report query
j. Drag a Union item from the Toolbox to the Query Explorer pane and add both of your Queries, as shown Figure 5-33.
Figure 5-33 Union statement
k. Open Query3 and drag all data items from the Source tab to the Data items tab of Query3, as shown in Figure 5-34.
Figure 5-34 Add all data items to Query3
l. Open Page1 and drag Query3 to your page.
m. Group the SCServer and the StorageSubsystemName columns.
n. Run your Report.
You should get a report that shows your storage virtualizers in both Spectrum Control instances with their pools and the capacity metrics, as shown in Figure 5-35.
Figure 5-35 Rollup report across two Spectrum Control servers
For more information about the Spectrum Control exposed views, see the following website:
|
Note: If you use Local OS authentication for Cognos, as described in IBM Tivoli Storage Productivity Center V5.2 Release Guide, SG24-8204, you must add the users to the db2users group to create reports with native SQL.
|
5.2 Organizing capacity data around applications and business units
When you justify costs for existing or new storage systems, have answers to the following questions:
•Which applications use how much and what kind of capacity?
•Who owns which application?
Spectrum Control can help you get answers to these questions through the Grouping function that is part of Spectrum Control. With the Grouping function, you can build a business-oriented hierarchy in Spectrum Control. This hierarchy can be as simple or as complex as you require, as shown in Figure 5-36.
Figure 5-36 An example of a business-oriented hierarchy in Spectrum Control
|
Note: A subcomponent is an application that is a child of another application.
|
In the example shown in Figure 5-36, resources such as data stores, exports, file systems, file sets, servers, VMs, volumes, and volume groups were added to the subcomponents Node, Testapp, and Web, and to applications Apps, Epic, Sharepoint, and DB2. The applications Node, TestApp, and Web were then added to the application called Connections. The applications Connections, DB2, Sharepoint, Epic, and Apps were then added to the Department Test. The departments Test, Development, and Research were added to the Main-Department Research and Development.
There are three ways to add resources to applications:
1. Use filters to assign resources to the applications that are used in step 6 on page 150. For more information, see the Spectrum Control IBM Knowledge Center, found at:
2. Add resources manually to the applications that are used in step 14 on page 154. For more information, see the Spectrum Control IBM Knowledge Center, found at:
3. Add resources by using the command-line interface (CLI). You can do so in two ways:
a. Add resources individually.
b. Add resources by using bulk assignment.
For more information, see the Spectrum Control IBM Knowledge Center, found at:
|
Note: Not all resources can be added by using all methods. For example, virtual machines (VMs) that are not listed as a server can be added only by using the CLI.
When you use the CLI, you can add all resource types and automate your bulk assignment of eligible storage resources to applications.
There are more functions that are available than are shown in this section. For more information, see the Spectrum Control IBM Knowledge Center.
|
The Application and Department function can be accessed on the Groups tab in the navigation bar.
Figure 5-37 shows the Application and Department function.
Figure 5-37 Application and Department function
In the example that is shown in Figure 5-37 on page 147, you can start creating your first application Connection by using the subcomponents Node and Web. Complete the following steps:
1. From the Groups tab, click Application.
2. Click Create Application, enter Connection as the application name, and click Create.
Type and Subtype are not mandatory. To keep this example simple, we do not specify Type and Subtype.
Figure 5-38 shows how to create an application.
Figure 5-38 Create Application wizard
3. Now, you have several options to either assign resources to the Connection application or specify subcomponents. In this example, we choose to create the subcomponent Node. Click Create subcomponents that each have their own set of storage resources and then click Close, as shown in Figure 5-39 on page 149, which shows the methods that are available after creating the Connection application.
Figure 5-39 Methods to either add storage resources or add subcomponents to the application
4. In the Create Subcomponent window, enter Node in to the Name field, and click Create, as shown in Figure 5-40. This action creates the Node subcomponent.
Figure 5-40 Subcomponent window
When you look at the Connections application, you see the Node subcomponent, but all the Capacity values are still at 0 GiB because no storage resources are added yet.
Figure 5-41 shows the Connection application with the Node subcomponent.
Figure 5-41 All capacity values are at 0 GiB because no storage resources are added yet
5. Double-click Node to add storage resources.
Figure 5-42 Example of creating a filter
8. Define the Name Patterns for the Volumes. You can use the wildcards “*” for multiple characters or “?” for one character. Because of naming conventions in our example, company, all volumes that belong the Node subcomponent start with TPC. Therefore, enter TPC*, as shown in Figure 5-43, which shows the method of adding storage resources to an application or subcomponent by using the filter method.
Figure 5-43 Add storage resources to an application or subcomponent by using the filter method
9. Click Preview to see the volumes that match this pattern. Figure 5-44 shows all volumes that match the naming pattern.
Figure 5-44 Shows all volumes that match the naming pattern.
|
Tip: If we had specified “TPC_*” in Figure 5-43 on page 151, we would not see the “tprcr*” volumes in Figure 5-44.
|
10. Because only volumes on SVC-2145-svc_27-28-IBM belong to the Node subcomponent, click Close to redefine the filter.
11. Select the Belonging to check box, select Storage System, and enter SVC-2145-svc_* in to the Storage System field to get only volumes with names that start with TPC that are on SVC-2145-svc_27-28-IBM, as shown in Figure 5-45 on page 153, which shows an example for a filter on volumes.
Figure 5-45 Example of how a filter can be specified
|
Tip: Filters can be restricted, for example, all volumes that are mapped to a given server or contained in a given pool.
|
12. Click Preview again and check whether the filter is correct now.
Figure 5-46 shows the volume selection after redefining the filter.
Figure 5-46 Show only volumes with names that start with TPC and are on SVC-2145-svc_27_28-IBM
13. Click Save. When you provision a volume with a name that starts with “TPC on SVC-2145-svc_27_28”, it is added to the Node subcomponent automatically.
|
Note: The filter function allows resources of a given type matching a certain naming pattern to be associated. Filters are dynamic and are continuously evaluated. Using the filter function keeps the application self-maintaining.
|
14. If the naming conventions were not implemented everywhere, you must add a server explicitly through the web-based GUI. Agentless servers can also be added to an application. Right-click the agentless server 9.155.90.222_Demo and select Add to application, as shown in Figure 5-47 on page 155, which shows how to add storage resources explicitly through the web-based GUI.
Figure 5-47 Add a data store to an application
15. A list of existing applications and subcomponents opens. In our example, we must create the web application. To do so, click Create Application, enter Web in to the Name field, and click Create.
Figure 5-48 shows how to create an application while explicitly adding storage resources.
Figure 5-48 Create the Web application
16. The Web application must be added to the Connection application. To do so, click Groups, click Applications, right-click the Web application, select Add to Application, select the Connections application check box, and click Save.
17. When you view the Subcomponents application under the Connections application, you see the two subcomponents with their capacity values, as shown in Figure 5-49 on page 157.
Figure 5-49 Connection application showing all the subcomponents and their resources
Related resources are explained in 5.3, “Getting an overview of the resources that are associated with your applications” on page 164.
18. Looking at the application’s level view in Figure 5-50, you see the aggregated capacity values of the two subcomponents of the Connection application.
Figure 5-50 Aggregated capacity values of the Connection application
To add a VM to an application, you must use the tpctool command.
|
References: For more information about how to use the tpctool command, see the Spectrum Control IBM Knowledge Center at the following websites:
•For information about command modes, go to:
•For information about tpctool commands, go to:
|
19. Start a command window (CMD) and from the <install>/cli directory, run the following command to get the tpctool prompt:
tpctool -user tpcadmin -pwd password
20. Because you must use tpctool to add the VM to the application, you also must create the TestApp application by using tpctool by running the following command:
mkappgroup -name TestApp
Figure 5-51 shows how to create the TestApp application.
Figure 5-51 The mkappgroup command and the associated ID
21. To get the ID of the VM, run the following command:
lsvm -l
Figure 5-52 shows the lsvm - l command.
Figure 5-52 VMs and their parameters
22. To confirm the ID of your TestApp application, run the following command:
lsappgroup - l
The lsappgroup command lists all applications and their parameters, as shown in Figure 5-53.
Figure 5-53 The lsappgroup command shows the application IDs
23. To add the ws-vcenter VM to the TestApp application, run the following command:
modifyappgroup -add -id /Application/9352510 -type vm -key ws-vcenter+9.155.113.62
Figure 5-54 shows how to add a VM to an application.
Figure 5-54 Add the ws-vcenter VM to the TestApp application
24. To check the application members of the TestApp application, run the following command:
lsappgroupmembers -id /Application/9352510
Figure 5-55 shows the members of the TestApp application.
Figure 5-55 The ws-vcenter VM as a member of the TestApp application
25. Now, you must add the TestApp application as a subcomponent to the Connection application. You also want to add a few volumes that do not follow any naming conventions to the Connection application. In this scenario, a colleague informed you that he does not need the WIN-2I0P2DE19BJ server in his Web application anymore, and that he marked this server as “notused”. This task can be done in one step by running the modifyappgroupvia command.
|
Reference: For more information about the usage of custom tags, see IBM Tivoli Storage Productivity Center V5.2 Release Guide, SG24-8204.
|
26. To get the volume IDs, you must determine the device GUID on which the volumes are by running the following command:
lsdev -l
Figure 5-56 shows the output of the lsdev -l command.
Figure 5-56 Devices and their parameters
27. Now, you can create a list with the volumes of this DS8000 storage system. To pipe the output to the file location, open another command line, go to the <install>/cli directory, and run the following command:
tpctool -user tpcadmin -pwd password lsvol -l -dev 2107.75ZA571+0 >> C: empvol.txt
28. Open the vol.txt file and find the labels for the volumes that you want to add to the Connection application.
Figure 5-57 shows an example for the lsvol -l output.
Figure 5-57 List of volumes for the specified device
29. Using the inputfile.txt input file and the modifyappgroupviafile command, you can add and remove multiple devices to or from multiple applications in one step. To do so, run the following command:
modifyappgroupviafile -inputfile c: empinputfile.txt
Figure 5-58 shows the modifyappgroupviafile command.
Figure 5-58 Example of the modifyappgroupviafile command
Figure 5-59 shows the content of the inputfile.txt file.
Figure 5-59 Example of an input file for the modifyappgroupviafile command
Now, the Connection application lists three subcomponents and their capacity values. Figure 5-60 shows the subcomponents of the Connection application.
Figure 5-60 Subcomponents and the related resources for the Connection application
The capacity values of Connection application changed. Figure 5-61 on page 161 shows the capacity values for the Connection application.
Figure 5-61 The parameters of the Connection
30. The DB2, Sharepoint, Epic, and Apps applications can be created in a similar fashion. To create the Test department, select the Connection and Apps applications and click Add To → Department, as shown in Figure 5-62.
Figure 5-62 Add multiple applications to a department
31. Click Create Department, enter Test into the Name field, and click Create, as shown in Figure 5-63.
Figure 5-63 Create a department
Similarly, you can add the DB2 and Epic applications to the Research department and the Sharepoint application to the Development department.
32. Now, the Development, Research, and Test departments must be added to the main Research & Development department. To do so, click Groups → Departments, select the Development, Research and Test departments, click Add to → Department → Create Department, enter Name Research & Development, and click Create.
The view of the department’s condition and capacity consumptions are rolled up to the department level. This table can be exported to CSV, PDF, or HTML formats. Figure 5-64 shows the accumulated capacity data of the main department.
Figure 5-64 Accumulated capacity values of the Research & Development department
33. Click the Research & Development department to get to the condition and capacity consumptions of each subdepartment. Figure 5-65 shows the subdepartments of the Research & Development department.
Figure 5-65 Capacity breakdown per subdepartment
5.3 Getting an overview of the resources that are associated with your applications
In the overview window of each of the departments or applications, you also can determine the application’s or department’s historical capacity consumption and the capacity usage per Tier.
Figure 5-66 shows an overview of the Research & Development department.
Figure 5-66 Overview with historical capacity usage and capacity usage by tier
The Spectrum Control Grouping function provides application and department insights, such as:
•Volume activity over time
•Network activity over time
You can also quickly identify resources that are related to the application or department and their condition.
5.4 Drilling into resource details for problem determination and performance analysis
In high-priority situations, for example, when there is a poor response time for client application servers, Spectrum Control assists with performance troubleshooting. This section describes how Spectrum Control can help you with performance troubleshooting.
Complete the following steps:
1. Open the server’s Overview window, as shown in Figure 5-67.
Figure 5-67 Overview window
|
Note: Regardless whether you deployed a storage resource agent to see the server’s Overview window, you also can see the server’s Overview window from agentless servers. For more information about agentless servers, see IBM Tivoli Storage Productivity Center V5.2 Release Guide, SG24-8204.
To get help on how to create the CSV file that can be used as an import file for creating agentless servers, download a script from the ones found at the following website:
These scripts use SQL queries to extract the host connection information from the Spectrum Control database to a CSV file, which then can be manually imported by using the web-based GUI Agentless Server definition process.
|
2. You should be able to see the Most Active Storage System Volumes entry in one quadrant (you can use the drop-down menu to select this chart), as shown in Figure 5-67 on page 165. Determine the volumes that have high response times.
In our example, the cet-hacmp2_vol_tcrdblog volume shows high response times, as shown Figure 5-68.
Figure 5-68 The cet-hacmp2_vol_tcrdblog volume shows high response times
You can use the arrows to view other parameters, such as Volume Utilization, I/O Rate, Data Rate, and Read Cache Hits.
3. In the left navigation pane under Related Resources, click Volumes and then click the Performance tab, as shown in Figure 5-69 on page 167, which shows the Volumes Performance for a given server.
.Figure 5-69 Volume Performance for a given server
5. In Figure 5-70, which shows volume performance for a given volume and a given time frame, if you cannot see the pool column, click the header row and select the Pool check box.
Figure 5-70 Volume performance data for a given volume and a given time frame
6. Select or clear the volumes that you want see in the chart.
7. Adjust the time frame for which the performance data is shown, either relatively or absolutely, to check whether the problem occurs on a regular basis.
8. Select the Data Resolution: Sample, Hourly, or Daily
|
Note: If you choose a long time frame, data resolution switches automatically to a higher aggregation level because only a certain number of dots can be displayed in the graph.
|
9. Select or clear the performance metrics in which you are interested. For example, select Read Response Time and Write Response Time to distinguish if the reads or the writes are causing the problem.
Figure 5-71 shows how to select Volume Performance metrics.
Figure 5-71 Select Volume Performance metrics
10. Click Pools, as shown in Figure 5-72, which shows how to select the Pool Performance chart for a given pool. Select the pool in which the volume is in and click View Performance to check whether the whole pool is showing high response times.
Figure 5-72 Open the Pool performance chart for a given pool
11. To show the same time frame that is shown in the Volume performance chart, click the Synchronize Time icon in the Volume performance chart, as shown in Figure 5-73, which shows the Synchronize Time icon.
Figure 5-73 Synchronize Time icon
|
Note: Using the Synchronize Time icon synchronizes the time range across all the performance views that are displayed in separate browser windows.
Use this action when you change the time range in a performance view and want to apply the same time range to the other performance views.
|
12. Depending on the storage subsystem on which your pool is located, you can select Disk and Volume Performance metrics so that you can determine whether the problem is caused by the back end, as shown in Figure 5-74, which shows the performance metrics for Pool performance for a given pool on a storage virtualizer.
Figure 5-74 Performance metrics for pool performance for a pool on a storage virtualizer
13. From here, you can drill down, for example, to see the performance of all volumes in that pool to see whether they have the same issue. Right-click the pool and select Volume Performance, as shown in Figure 5-75, which shows the drill-down possibilities from a storage virtualizer pool.
|
Note: For information about which resources to which you can drill down, see the Spectrum Control IBM Knowledge Center, found at:
|
Figure 5-75 Drill-down possibilities from a storage virtualizer pool
|
Note: By default, the five most used volumes regarding total I/O are shown in the performance graph. But, you can select or clear the volumes for which you need the performance graph. For more information, see IBM Tivoli Storage Productivity Center V5.2 Release Guide, SG24-8204.
|
14. Select the Resources volumes for which you want to see the performance data and the Metrics that are you are interested in, as shown in Figure 5-75, which shows the volume performance for a given pool.
Figure 5-76 Volume performance for a given pool
15. You can easily correlate the graph to the volume by hovering your cursor over the datapoints, as shown in Figure 5-76.
16. Drill down to the pool’s MDisk Performance, as shown in Figure 5-75 on page 172, to see whether one of the disks is providing too long response times.
17. Drill down to Disk Performance Metrics to check whether the problem is caused by any disk as shown in Figure 5-77, which shows how to drill down from MDisk performance to Disk performance.
Figure 5-77 Drill down from MDisk performance to Disk performance
18. If you are troubleshooting a volume of a storage virtualizer pool, be sure to check the node balance of your storage system to make sure that everything is balanced. To do so, go to the Pool Performance chart and drill down to the Node Performance chart, as shown in Figure 5-75 on page 172, which shows that not all nodes of the storage virtualizer are used equally.
Figure 5-78 Not all nodes of the storage virtualizer are used equally
|
Note: For more information about available performance metrics in Spectrum Control, see the Spectrum Control IBM Knowledge Center, found at:
|
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.