IBM AIX application workloads on IBM Hyperconverged Systems powered by Nutanix
An integrated solution that enables you to run enterprise transactional and analytical workloads
Hyperconverged system is a software-centric infrastructure which tightly integrates compute, networking, storage resources, and other technologies pre-integrated in a scale-out server environment. IBM® Hyperconverged Systems powered by Nutanix now supports IBM AIX®. AIX runs fully virtualized on the Nutanix Acropolis Hypervisor and its virtual machine (VM) lifecycle is managed through Nutanix Prism. This article demonstrates how the advanced architecture of IBM Power® can be blended with Nutanix web-scale technology, for quick and easy deploy of application workload on an AIX guest VM, through migration from an existing AIX environment or through fresh deployment.
It is assumed that the reader is already familiar with IBM Hyperconverged Systems, AIX, Nutanix Prism, IBM DB2®, IBM WebSphere® Application Server (WAS), and IBM WebSphere MQ. If not, it is recommended to review the documentation in the ” Related topics” section first.
An enterprise cloud infrastructure combining Nutanix software stack with IBM Power Systems™ architecture, is a good fit for transactional and analytic workloads, offering agility, scalability, security and ease of management. The IBM Hyperconverged Systems, powered by Nutanix are scale-out nodes that combines storage, computing, and networking into a single system.
The IBM Power System CS822 model, integrated with Nutanix enterprise cloud platform software, targets enterprise workloads. Nutanix’s built-in AHV hypervisor allows IBM Power customers to run a variety of virtualized applications. In addition, Nutanix offers a management solution through its Prism console for centralized IT management that streamlines and automates common workflows, thus reducing data center complexity. This joint initiative enables hyperconverged deployments for enterprise workloads such as databases, large-scale data warehouses, web infrastructures, and mainstream enterprise applications.
Figure 1. Nutanix Prism console
The primary objective of this article is to provide guidelines on how IBM Power customers can leverage IBM Hyperconverged systems using AIX on Nutanix to run their workloads, with examples in these scenarios:
- Fresh deployment of AIX guest VM containers on Nutanix
- Migration of existing AIX workloads to a Nutanix environment
IBM AIX enabled for IBM Hyperconverged Systems
IBM Hyperconverged Systems powered by Nutanix now supports AIX 7.2 TL 2 SP2, or later, on. AIX can be a guest VM on the Nutanix Acropolis Hypervisor, including storage and networking and it can be managed by the Prism console. For ease of deployment, AIX cloud ready raw disk images are available. These images can be used directly for VM deployments on CS series nodes. You can also opt for migrating an AIX logical partition (LPAR) to the Nutanix environment. The following section provides the step-by-step procedure to migrate the IBM AIX partitions to the Nutanix environment.
Setup considerations for AIX on Nutanix
We tested the following procedure in our environment by setting up a Hyperconverged environment using the following products/versions.
Table 1. Components version
|Operating system||AIX 7.2 TL2 SP2|
|Database||IBM DB2 Server 10.5 for AIX and DB2 10.5 Fix Pack 8 for AIX (64 bit)|
|Messaging Queue||IBM MQ Version 8.0 and Fix Pack 8|
|IBM WebSphere Application Server|| IBM WebSphere Application Server Version – 8.5.5 |
IBM Installation Manager V1.6.2IBM WebSphere SDK Java Technology Edition – V8.0 and WebSphere Application Server Supplements – IBM HTTP Server and Web Server plug-ins, and WebSphere.
Fixes: iFIX (IT23364) and ifpi73197
Table 2. Nutanix cluster details
|Nutanix Acropolis operating system (AOS)||188.8.131.52|
|Acropolis Hypervisor (AHV)||20170331.78|
|Management console||Prism UI|
|Nutanix Cluster Check (NCC)||184.108.40.206|
Table 3. IBM Power server details
|Model||Processor||Speed||Memory||Hard drive||CD ROM|
|CS822||22-core|| 2.92 GHz |
(3.492 GHz turbo)
|256 GB to 512 GB||Up to 15.38 TB|| Up to 5 PCIe I/O |
Table 4. Step-by-step procedure
| Step ||Description|
|Step 1: Create a VM.|| |
|Step 2: Install DB2 and create a database.|| |
|Step 3: Install and configure MQ.|| |
|Step 4: Install IBM WebSphere and IBM Java SDK.|| |
|Step 5: Install the test application.|| |
|Step 6: Migrate non-Hyperconverged AIX LPAR.|| |
|Step 7: Run the test application.|| |
New deployment process for AIX on Nutanix
IBM team has done a deployment to capture the considerations on how enterprise applications can be easily deployed and managed on IBM Hyperconverged Systems, powered by Nutanix.
To showcase how the IBM-Nutanix integrated solution enables you to run enterprise transactional and analytics workloads, the team created a test bed that includes Nutanix cluster, AIX workloads (such as IBM Websphere Application Server and DB2), and test applications. Here’s a brief description:
- Nutanix cluster: A group of nodes are clustered together in a scale-out topology and managed dynamically by the Prism console with one-click simplicity.
- Nodes: The node is the IBM Power System CS822 model server. This model is built on high-performance IBM Power technology and integrated with Nutanix Enterprise Cloud Platform software.
- Management console: From the Prism console, created 10 VMs, where each VM runs on AIX 7.2 and has applications Installed.
Applications: IBM Websphere Application Server, Messaging Queue, DB2, and test applications are installed on each VM that runs AIX 7.2. [ See Table 1: Components version ]
Figure 2. Overview of VM summary from Nutanix Prism console
Applications on IBM Hyperconverged Systems
The combination of Nutanix and IBM Power eases the deployment of enterprise applications by removing the infrastructure complexity and in turn provides superior performance gains and VM density in cloud environments, with IBM POWER8® chip architecture enabling more threads per core, memory bandwidth, and cache. Furthermore, a variety of workloads can be installed avoiding the need for a separate silo of infrastructure. Here is a brief description of the workloads that were installed on the Nutanix cluster for our testing.
IBM WebSphere Application Server (WAS) – IBM WebSphere Application Sever is a middleware offering that allows you to build, run, integrate, secure, and manage Java-based applications ranging from simple websites to cloud solutions. The quick way to install IBM WebSphere Application Server is to use IBM Installation Manager that comes with the installation kit and then configure an IBM repository. In addition, you can install supplements such as IBM HTTP Server, Web Server Plug-ins for WebSphere Application Server, the WebSphere Customization Toolbox, or the WebSphere Application Client, and add them to the composite repository.
IBM WebSphere Messaging Queue – WebSphere MQ provides a universal messaging backbone with robust connectivity for flexible and reliable messaging for applications, and integration of existing IT assets that use a service-oriented architecture.
DB2 – IBM DB2 is a multi-workload database management solution built to handle a massive volume of data and deliver high performance to support real-time analytics. It provides data availability for demanding applications, scalability, massive parallel processing, and flexibility for enterprise demands.
Test application – It is an application developed by IBM Internal team and has the online transaction processing (OLTP) functions to simulate customers browsing and buying books. Its database contains a catalog of all the books and its cost.
In summary, this testing configuration and stack is intended to represent a realistic configuration of the IBM hardware and software components for an infrastructure stack to support AIX on Nutanix Hyperconverged environment. A customer-like application using WebSphere Application Server, WebSphere MQ and DB2 is used to apply and measure transactional loading so this environment can always be running while tests, operations, and migrations occur. Finally, transactions generated from this application server (or grouping of like application servers) can be used to stress/test back-end database servers.
Middleware and application setup details
This section lists a few considerations and technical issues to take care while installing IBM WebSphere applications on AIX clients in a Nutanix environment.
- For WAS 8.5.5 installation – Java SDK 6, which is out of support, is embedded in WAS traditional V8.5 as the default Java SDK. Hence you can either install WebSphere Application Server (Base) 8.5.5 with Java SDK 8.0 as default, or if you install Java 8 as an optional on top of WAS 8.5.5, then use the
manageSdkcommand to switch among Java 6, 7, or 8. Here is an example of enabling Java 8:
./managesdk.sh ‑listAvailable ‑verbose ./managesdk.sh ‑setCommandDefault ‑sdkName 1.8_64 ./managesdk.sh ‑setNewProfileDefault ‑sdkName 1.8_64 ./managesdk.sh ‑enableProfile ‑profileName AppSrv01 ‑sdkName 1.8_64
- For WAS ND 9.0.0 installation – Install IBM SDK, Java Technology Edition Version 8 together with IBM WebSphere Application Server Network Deployment. To function correctly, these packages must be installed together. Here is a sample command for reference:
./imcl ‑acceptLicense install com.ibm.websphere.ND.v90_220.127.116.1160526_1854 com.ibm.java.jdk.v8_8.0.3000.20160526_1317 ‑installationDirectory <install_directory> ‑repositories /<directory_path>/repository.config,/ <directory_path>/sdk.repo.8030.java8.aix/repository.config
- To create profiles in IBM WebSphere Application Server:
For WAS 8.5.5 For WAS ND 9.0.0 ./manageprofiles.sh -create -templatePath /AppServer/profileTemplates/default -cellName -nodeName ./manageprofiles.sh -create
/AppServer/profileTemplates/default -cellName -nodeName -hostName -serverName server1 -isDeveloperServer false
Migrating an existing AIX LPAR to Hyperconverged Nutanix environment
Following is the procedure that we followed and tested to migrate an existing AIX LPAR to Hyperconverged Nutanix environment.
Prepare for the migration
- Unmirror rootvg: If rootvg is mirrored on multiple disks on the source LPAR, we must unmirror rootvg before carrying out the migration under discussion. To unmirror the rootvg, use the
unmirrorvgcommand. For example, run the following command to unmirror the root volume group on hdisk11.
unmirrorvg rootvg hdisk11
- If multiple copies are available on different disks, unmirror the rootvg from all the mirrored disks.
- Run the
reducevgcommand to reduce the disk out of the root volume group.
reducevg rootvg hdisk11
- Use the
bootlistcommands to re-initialize the boot record of the remaining disk and modify the boot list in order to remove the unmirrored disk from the list. If the source LPAR has rootvg in hdisk1, run the following commands:
bosboot ‑a ‑d /dev/hdisk1 bootlist ‑m normal hdisk1
- Before migration, make sure that the following packages are available in the system at the OS level.
Figure 3. List of packages and levels to be installed on AIX LPAR before the migration
Review the supported levels of software products on AIX 7.2 and perform the necessary migration activities.
Create images of the disks using the
dd command creates image with disk contents to a file. The dd images can later be used for building volume groups on the virtual machine on Nutanix.
- Create a list of rootvg, and other volume group names, and the current disk location. This would help restoring all volume groups on the Nutanix environment.
- Create images of all rootvg and datavg disks to files
dd if=/dev/<rootvg disk name> of=/<desired NFS location>/<rootvg disk name>.dd.rtvg bs=1m dd if=/dev/<datavg disk name> of=/<desired NFS location>/<datavg disk name>.dd.dtvg1 bs=1m <use the same procedure for all the data disks>
Create an image on Nutanix and build a new VM: Upload the rooting image file to the Nutanix environment using the Prism console and build a new VM using the same image. You need to add additional data disks with the same size in the existing LPAR while building a new VM and finally activate the VM. You can transfer the datavg image files to a new VM or a NFS location. Refer to the Prism documentation for more details.
Re-create data volume groups on the new VM
To clean up the new VM and re-create all the data volume groups:
- Export all the data volume groups by using the
exportvgcommand. This command exports the definition of a volume group. The
exportvgcommand removes the definitions of all the data volume groups from the new VM.
exportvg <datavg name> <use the same procedure for all the datavg>
- After removing the definitions of all the data volume groups, re-create all the data volume groups by using
importvgcommand re-creates the data volume group and later mounts all the file systems manually.
importvg ‑y <datavg name> <one of the datavg disk name> <use the same procedure for all the datavg>
- You may set the new host name or the new IP address after activating the migrated virtual LPAR. You may need to change the application configuration file manually to reflect the new host name or the new IP address.
IBM and Nutanix bring the best of private and public clouds to its enterprise customers by offering agility, efficiency, and security with their infrastructure. This combination eases the deployment and management of enterprise workloads such as Enterprise DB, MongoDB, WebSphere Application Server, Business Applications, Database as a Service (DBaaS), using the Power performance and server virtualization without worrying about storage or network bottlenecks.
The solution in this article is deployed as a cluster of nodes. This Nutanix cluster can grow by adding more nodes for demanding workload, thus giving you a linear predictable performance and capacity increase as the cluster grows. In addition, IBM Power customers can migrate an existing AIX LPAR to a hyperconverged Nutanix environment and benefit from Nutanix web-scale technology that simplifies infrastructure management and reduces operational cost.
The goal of this integrated solution is to provide an infrastructure that is easy to manage and simple to scale for demanding transactional and analytics workloads.