Hitachi Ops Center​

 View Only

Validating the Hitachi Ops Center Protector v7.3 Global Replication Report including its impact on the Ops Center CM-REST Server

By Tanmoy Panja posted 12-20-2021 00:17

  

Objective

  • Validate the Global Replication Report (GRR) in Hitachi Ops Center Protector v7.3 in a large-scale environment.
  • Validate the impact of running the reports on an Ops Center V10.8.0 API Configuration Manager (CM-REST) server by registering ten VSP 5600 storage systems (approximately 9K replication pairs)


Introduction

The Global Replication Reports provide a complete list of all replications on the storage systems including replications that are not managed by Protector. This feature works by interfacing with the Ops Center CM-REST server, which provides an inventory of all replications on each storage system. Protector caches the result from the CM-REST server and aggregates the results.


Test Methodology

  1. Ops Center Protector was installed on a virtual machine (VM) server using the standalone installer. The Ops Center CM-REST server, which we need for reporting, was installed on a separate VM server using the Ops Center OVA (Open Virtualization Appliance) v10.8.0.
  2. Ten VSP 5600 storage systems were registered in the Ops Center CM-REST server instance.
  3. The ten VSP 5600 storage systems were discovered as block storage nodes in the Protector UI by specifying the address and port of the Ops Center CM-REST interface for the storage system being configured (as described below).
  4. Different replication pairs (Snapshots/SI/TC/HUR/GAD) were created on these storage systems outside of Protector.
  5. Periodically, the “All Reports” version of the Global Replication Report was run from the Protector Reports section of the UI and the CM-REST server resource utilization was checked.


Environment Configuration Layout

The overall environment layout for the testing is shown in the following diagrams:

Figure 1 – Environment Layout (GRR)

As shown in the previous diagram, the test environment consisted of the following:

Configuring using the Protector UI

To configure a Block Node Device to generate the Global Replication Report from the Report wizard, complete the following steps:

  1. From the Protector UI, open the Edit Node – Hitachi Block Device

    When setting up the Protector Block Node, there is now an additional section where we can specify the connection and credentials details for the REST interface on the CM-REST server.
  2. Proceed through the wizard until you reach the window titled Specify configuration for Global Replication Reports as shown in the following screenshot:

Figure 2 – Edit Node wizard

 Specify the address and port of the Ops Center CM-REST server for the storage system being configured.

  1. Specify whether to use HTTPS, and if so, whether to ignore certificate errors.
  2. Specify the credentials for the storage system user account to use.
  3. Specify the refresh interval.
  4. Click Next and continue until you complete the wizard.

 

Test Implementation

After the storage systems are configured in CM-REST and Protector, we completed the following procedure periodically to validate the CM-REST server operation under a heavy load.

Generating the Global Replication Report

When generating the Global Replication report, there are multiple choices that enable you to specify how to group the replications. By default, the report includes all replications on all storage systems.

  1. Navigate to the Reports screen.
  2. Determine which Global Replication Report to generate from the following:
    • All Pairs (ungrouped pairs)
    • Pairs by Storage Record (pairs managed by Protector)
    • Pairs by Consistency Group (consistency grouped pairs)
    • Pairs by Copy Group (copy grouped pairs)
    • Pairs by Snapshot Group(snapshot grouped pairs)

In our lab test, we navigated through All Pairs (ungrouped pairs) only. This report provides a list of all the replication pairs with no grouping.

Figure 3 – Protector Reports wizard

3. Click All Pairs to display the following:

Figure 4 – Protector All Pairs wizard

4. Click on Search button to display the following screen:

Figure5 – Protector All Pairs wizard 


Ops Center CM-REST Server utilization observation

The CPU and memory usage of the CM-REST server was observed during the Global Replication report  refresh intervals as follows:

Figure6 – Ops Center CM-REST VM (Ops Center OVA) 

The following graphs show the CPU/Memory utilization:

Figure7 – Ops Center CM-REST server CPU/Memory Utilization


Limitations

  • Only the Ops Center version of CM-REST is supported (using PF-REST or other CM-REST versions is not supported).
  • Restrictions exist on sharing resources between Ops Center CM-REST and Protector as follows:
  • Fibre command devices cannot be shared between products.
  • Storage system user accounts cannot be shared between products if they are running on the same host.
  • Short report refresh intervals should not be used in busy environments.
  • A limited number of results is allowed per request (maximum of 1000 records); however, filters can be used to target requests more accurately. It is extremely difficult to visually manage 1000 items, so filters are commonly used to reduce the displayed records to something more manageable.


Test Conclusion

 Based on the test results, we came to the following conclusions:

  • The CM-REST server responded normally to periodic requests from Protector for all replication pairs for ten VSP 5600 storage systems.
  • CPU and Memory usage on the CM-REST server showed no unusual increases during the testing.
  • All replication pairs were listed in the report and could be exported from the wizard.
1 comment
28 views

Permalink

Comments

05-04-2022 11:44

Great post