You are currently on IBM Systems Media’s archival website. Click here to view our new website.

Linux on Power > Tips and Techniques > Systems Management

HMC Architecture Options


You have a variety of configuration options for the connections between your hardware management consoles (HMC) and managed servers. Some questions on this topic include: How do redundant HMCs work together? How do I configure the HMC connections with the redundant service processors on the E870/E880 servers? This article will provide answers to these questions along with an overview of HMC to server networking.

HMC Network Connections

There are four primary types of network connections on the HMC. They are:

  • The HMC to managed system connection that is used for communications between the HMC and the Flexible Service Processor (FSP) ports on each managed server.
  • The HMC to logical partition connection, which is a network connection to each LPAR on the managed servers. This network connection is used to communicate errors and performance data between the operating system and the HMC. It also used to report hardware inventory data. If you’re performing any Dynamic LPAR (DLPAR) changes to a running LPAR, these changes will be communicated over this connection.
  • The HMC to remote users connection, which provides access for remote users and administrators who need to run the HMC GUI on their workstations.
  • The HMC to server and support connection, which is used for electronic service reporting to IBM. If you have multiple HMCs, one HMC can serve as the gateway for sending reports to IBM.

This article will focus on the HMC to managed system connections and how they can are configured with one or two HMCs. Servers that have redundant FSPs will also be addressed.

Flexible Service Processor

All IBM Power Systems servers have at least one FSP. Every FSP has two, 1GbE RJ45 HMC connection ports identified as HMC-1 and HMC-2. HMC-1 is always connected to an HMC. HMC-2 will only used if you’re deploying redundant HMCs. A private Ethernet network is configured for connection of these ports to HMC(s). The POWER8 scale-out servers (e.g., S822, S814, S824 and Linux-only variants) have a single FSP. Note that the scale-out servers can be operated without an HMC if they’re going to be configured with just a single OS image. The POWER8 Enterprise servers (E870/E880) have two (redundant) FSPs, which are located in the System Control Unit. The HMC connection ports for the scale-out and E870/E880 POWER8 servers are shown in Figure 1. For E870/E880 servers, the HMC(s) should always be connected to both FSPs. In the event of an E870/E880 FSP failure, HMC communications will automatically switch to the redundant FSP. This functionality is designed into the system and you don’t need to do any configuration work to enable the automatic FSP failover. For all servers, the HMC-1 ports should be connected to the first HMC (HMC-1). If a second HMC is configured, all HMC-2 ports should be connected to the second HMC (HMC-2). HMC-1 and HMC-2 ports on the same server should never be connected to the same HMC. On the HMCs themselves, port en0 should always be for the HMC to FSP private network connection.

One or Two HMCs?

You can choose to implement one or two HMCs to mange your Power Systems servers. With just one HMC configured, should the HMC fail, the attached servers will continue to run. However, you won’t be able to perform LPAR operations like starting and stopping LPARs or make DLPAR configuration changes. Having a second (redundant) HMC permits full management capabilities for all attached servers in the event of an HMC failure. Figures 2 and 3 show the HMC connections for single and dual HMC implementations. When two HMCs are configured, server management operations can be conducted from either or both HMCs (i.e., the HMCs operate in an active-active mode). Any change made on an HMC is automatically replicated to the other HMC. During the execution of a command that modifies the configuration of a server, the other HMC will be briefly blocked out during the change. One exception to active-active, dual HMC operation is if you have implemented Power Enterprise Pools. With Power Enterprise Pools, a two-HMC configuration requires the HMCs to operate in a master and backup relationship for activity related to mobile activations. All changes to mobile processor or memory activations must be done using the master HMC. If the master HMC fails, the backup HMC can be changed to master designation.

Distributed HMC Configuration

When implementing two HMCs, they’re usually located within the same data center. For dual-site installations (production and DR) some businesses have been placing an HMC in each data center using a flat network to establish HMCs to server connections at both locations as shown in Figure 4. The distributed HMC configuration reduces the number of required HMCs from three (two at production and one at DR) to two. Use of a distributed HMC configuration is a favorable design for use with Power Enterprise Pools. With a distributed HMC configuration, you’ll be able to create a single server pool that includes servers located in both the production and DR data centers. This provides the capability to move mobile processor and memory activations between servers located in both the production and DR data centers for both DR tests and active DR failovers. In the event of a complete production data center failure, you can change the status of the DR HMC (HMC-2) from backup to master. Then all of the mobile activations can be assigned to the DR servers.

HMC Configuration Recommendations

It’s highly recommended that you deploy a redundant (dual) HMC design for use with your Power Systems servers. When deploying Enterprise E870/E880 servers, be sure to connect each HMC to both FSPs to take advantage of the automatic FSP failover capability. And, remember that all POWER8 servers require HMC Version 8 software.

Charlie Cler supports customers in a solutions-architect role at Forsythe Technology Inc. He can be reached at ccler@forsythe.com.





Advertisement

Advertisement

2019 Solutions Edition

A Comprehensive Online Buyer's Guide to Solutions, Services and Education.

LINUX ON POWER > TIPS AND TECHNIQUES > SYSTEMS MANAGEMENT

Planned Active Use of Elastic CoD Into Power Systems Solutions

LINUX ON POWER > TIPS AND TECHNIQUES > SYSTEMS MANAGEMENT

Introducing IBM FlashSystem 840

LINUX ON POWER > ADMINISTRATOR > SYSTEMS MANAGEMENT

HMC Architecture Options

IBM Systems Magazine Subscribe Box Read Now Link Subscribe Now Link iPad App Google Play Store