Contents


Hosted VMware environments and recovery solutions in IBM Bluemix Local System, Part 2

Setting up a PureApplication Software workload environment

Comments

Content series:

This content is part # of # in the series: Hosted VMware environments and recovery solutions in IBM Bluemix Local System, Part 2

Stay tuned for additional content in this series.

This content is part of the series:Hosted VMware environments and recovery solutions in IBM Bluemix Local System, Part 2

Stay tuned for additional content in this series.

Editor's note: This tutorial applies to V2.2.3.0 of IBM® Bluemix® Local System and PureApplication® products and is no longer being updated. IBM is continuing to improve and add features that support hosted VMware environments and recovery solutions. Please check the Comments section at the bottom of this tutorial for links to updated content.

You can use the workload environment solutions in Bluemix Local System and PureApplication Software from the V2.2.3 release to achieve additional flexibility in how you manage workloads. This flexibility is possible by deploying PureApplication Software or PureApplication Software Suite into Bluemix Local System. You can use this feature, for example, to isolate groups of application workloads for purposes of disaster recovery. Within a single Bluemix Local System, you can create many PureApplication Software workload environments, up to the number of compute nodes that exist on that system.

PureApplication Software is deployed into a virtual machine (VM) in the same Virtual Manager cloud group as the workload VMs that the software will manage. Therefore, the PureApplication Software will also consume memory, CPU, and storage from the cloud environment. This tutorial explains how to install and configure the PureApplication Software workload environment. It also includes guidance to help you streamline this process.

Workload flexibility

Each PureApplication Software workload environment requires a dedicated Virtual Manager cloud group. A single PureApplication Software workload environment can manage and target deployments for multiple Virtual Manager cloud groups, if each cloud group is dedicated for use by that one PureApplication Software workload environment.

This solution has a lot of flexibility and granularity if you have enough compute nodes to host individual workload environments in separate, dedicated cloud groups and enough storage capacity to handle multiple environments and replicated volumes. Like traditional cloud groups in Bluemix Local System, each Virtual Manager cloud group requires at least a single compute node to operate. Compute nodes cannot be shared between Virtual Manager cloud groups. Separate cloud groups exist to separate and isolate workload instances from licensing, networking, compute, and similar operations. As a result, pattern-based instances that are deployed as a set of VMs in a PureApplication environment are tightly coupled with the cloud group in which they are deployed. It is not possible to move running instances and VMs between cloud groups. Instead, these patterns must be redeployed into a new cloud group.

If business application data must move between cloud groups for this redeployment, the pattern creator must use block storage volumes that are mounted to the workload VMs to contain this data. These block storage volumes can be moved between cloud groups and between PureApplication Software workload environments. When the pattern is redeployed, with new VMs, you can use an existing block storage volume, which allows movement of data between cloud groups, even when the instance and VMs are redeployed.

For disaster recovery, each PureApplication Software workload environment has separate storage volumes that can be replicated independently. Therefore, you can perform disaster recovery operations for PureApplication Software workload environments independent from any other workload environments on that system or the system that it is paired with for replication. For example, you can have two workload environments on a single system that run in separate Virtual Manager cloud groups. You can choose to configure disaster replication for these two environments separately and do not have to fail over both environments at the same time. In fact, one workload environment can stay running on the second system indefinitely. Meanwhile, the other workload environment stays on the original system. You can find more details about how the disaster recovery solution works in Part 3 of this series.

Prerequisites

Before you begin this tutorial, you must read Part 1: Getting started with hosted VMware environments. You must follow the procedures in Part 1 so that you can set up the following items in this part:

  • Create an external IP address for the VM.
  • Create every Virtual Manager cloud group that the PureApplication workload environment will manage.
  • Ensure that sufficient IP addresses are available in the MKS Console IP Group.
  • Add at least one compute node to each Virtual Manager cloud group that the PureApplication Software workload environment will manage.

To set up PureApplication Software workload environments, you must open a web browser to the Bluemix Local System web console that will host this environment. Log in to the web console as a user with administrative access privileges.

In addition, you will make the following connections and, therefore, must be familiar with these technologies to successfully complete the procedure:

  • VMware vSphere Web Client, which requires a current browser with the Adobe Flash plug-in
  • Secure Shell (SSH) to an ESXi host
  • SSH to the RedHat Enterprise Linux V6.x 64-bit virtual machines
  • Web console for the PureApplication Software workload environment

The key steps in this tutorial require switching between consoles and SSH sessions multiple times. To simplify the terminology, we use the following names to describe the access mechanisms:

  • System console. PureApplication System or Bluemix Local System web console
  • Web client. VMware vSphere Web Client
  • Software console. PureApplication Software web console
  • SSH compute node. SSH session to the VMware ESXi compute node that is hosting the PureApplication Software virtual machine (VM)
  • SSH software. SSH session to the PureApplication Software VM
  • PureApplication Software VM. The VM where PureApplication Software is installed and running

After you establish a session with these connections, keep the connections available because multiple steps use them at different times.

1

Create the external user for VMware

To create the external application user who is required to set up your PureApplication Software workload environment, follow the procedures in the "Configuring external access to VMware components" section in Part 1. The external application access user must be associated only with Virtual Manager cloud groups. This user must also have default permissions (not read-only) and compute node access. Therefore, select all of the Virtual Manager cloud groups that this single PureApplication Software workload environment will manage.

After you create the external user, access the VMware vSphere Web Client:

  1. In the system console, select System → System Settings.
  2. From the list of System Settings, expand External Application Access Settings.
  3. Under the Actions column, click the paragraph icon (circled in the following figure) to access the required information to open a browser to the web client. External application access settings
    External application access settings
  4. Use the Virtual Manager entry to open a web browser to the IP address listed in that row.

    purevc: Ensure that the purevc host is mapped to the correct IP address as explained in the "Accessing VMware Virtual Center" section in Part 1.

  5. Click the Login to vSphere Web Client link, and enter the user name and password. To view the password, click Show Passwords above the External Application Access settings table.

Networking

The PureApplication Software workload environment will access at least two networks—the data network and the system management network—in your environment. In most cases, the PureApplication Software VM will access the data network directly and access the system management network through the existing routing from the data network. If it does not, you must define two interfaces on the PureApplication Software VM. One interface will connect to each network.

The following figure illustrates the network connectivity of the VM, including the two networking options:

  • Option 1: One network interface with connectivity to customer data and management networks
  • Option 2: Two network interfaces where customer data and customer management networks are separated

The preferred configuration is option 1 because it is a simpler configuration.

VM network connectivity
VM network connectivity

To ensure proper network connectivity to the components in the rack:

  1. In the system console, click System → Network Configuration.
  2. Expand VLANs and Links.
  3. Verify that the management VLAN shows the two network types. If only System Management is specified, under Actions, click the pencil icon to edit the entry.
  4. Select both the IP Group and System Management types.
  5. Save your changes.
Network types for the management VLAN
Network types for the management VLAN

Storage

The storage requirements for a PureApplication Software workload environment installation can be different across configurations. Two environment types are described here.

To install the PureApplication Software workload environment, you must create a minimum of two block VMware VM File Systems (VMFS):

  • Image repository datastore: This datastore contains a cache of images that are used for deploying workloads. Each software installation has only one image repository that is shared across compute nodes and cloud groups.
  • Deployment datastore: This datastore contains deployed VMs. The first VM that is deployed is the PureApplication Software VM. You need one deployment datastore to get started, but you can add more deployment datastores later. The initial deployment datastore must be large enough to host the PureApplication Software VM along with additional capacity that is needed for the operating system disk and other miscellaneous files.

To simplify storage configuration and to ensure that you properly size the PureApplication Software VM, choose one of the following two environment types for which the storage requirements most closely match your needs:

  • Development/Test: You will use the installation as a development or test environment to create, validate, and test patterns.
  • Production: You will use the installation in a production environment, including environments that use disaster recovery and block storage replication of the volumes in the installation. VMware cannot replicate or clone volumes. Therefore, you need larger initial sizes because it might not be possible to expand them later.

Based on the environment type that best matches your storage requirements, you then use the corresponding storage sizes from the following table.

Table 1. Storage sizes for environment type
Environment typeImage repository datastore sizeInitial deployment datastore sizePureApplication Software VM hard disk size
Development or test 300 GB 800 GB 580 GB
Production 2 TB 3 TB 2.1 TB

Image repository: The image repository can be associated with multiple cloud groups, but for other volumes that you create, it should be associated with only one cloud group. If you create additional Virtual Manager cloud groups for this PureApplication Software instance to manage, you must update the image repository volume to list an association with each Virtual Manager cloud group.

2

Create the virtual machine for PureApplication Software

To create the PureApplication Software VM:

  1. Log in to the compute node in your Virtual Manager cloud group that you want to host the PureApplication Software VM. In the system console, you should still see the details for your external application accesses user ID.
    1. From the External Application Access Settings section that matches the compute node for the initial deploy of the PureApplication Software VM, get the compute node user ID, password, and IP address.
    2. Log in to the compute node by using the SSH command or an application, such as Putty, that supports the SSH protocol. The following example shows how to log in by using the SSH command:
      		ssh app_560aab7f@172.25.120.101
  2. Change directory to the datastore (block VMFS) where you want to create this VM. You cannot use the datastore that you created for the image repository.
    1. Find the Datastore name. Click Cloud-> Volumes in the system console. Then, select the block VMFS volume that you created for virtual machine deployments. The datastore name is the Datastore name property in the details for the data volume that is associated with the cloud group. In this example, the datastore name is p_9baf11f6-6d8c-4faa-b631-5458655a8ecb.
    2. Change directory to this datastore name. In this example, we use SSH as shown in the following example:
      cd /vmfs/volumes/p_9baf11f6-6d8c-4faa-b631-5458655a8ecb
  3. Create a temporary directory and print its location:
    mkdir temp
     cd temp
     pwd

    The pwd command has the following output path, which you will need in the following sections:

    /vmfs/volumes/p_9baf11f6-6d8c-4faa-b631-5458655a8ecb/temp
3

Export the RHEL6.x 64-bit base OS image

For PureApplication Software V2.2.3, you must run Red Hat Enterprise Linux (RHEL) Version 6.X (minimum of RHEL 6.6) on the VM where you will install the software. One option is to use a Core OS image for RHEL 6 from IBM that comes with a Bluemix Local System license. You can also purchase this license separately. The PureApplication Software installation includes sample patterns and other pattern components, but it does not include a base OS image.

Attention: Before you begin the following steps, make sure that Bluemix Local System contains at least one conventional cloud group with at least one compute node.

To use a Core OS image based on RHEL 6 from IBM as the PureApplication Software VM operating system:

  1. In the system console, click Catalog -> Virtual Images.
  2. Click the IBM OS Image for Red Hat Linux Systems link for version 2.1.8.0.
  3. In the right pane, click Export. Export option
    Export option
  4. Specify the IP address, user ID, and password for the compute node from the External Users table. The path is the one from Step 1. Entering the login details
    Entering the login details

The export takes about 10 minutes. When it is done, the IBM_OS_Image_for_Red_Hat_Linux_Systems.ova file is placed in the temporary directory on the compute node.

4

Create a new virtual machine

If you are not already logged in to the Virtual Center web client, log in to it now as explained in Step 2. To create a new VM:

  1. In the Navigator, select the Hosts and Clusters entry.
  2. Expand the navigation tree until you see the cluster that matches the Virtual Manager cloud group name. Select the cluster.
  3. On the Cluster pane, click Actions -> New Virtual Machine. Creating a new virtual machine
    Creating a new virtual machine
  4. In the New Virtual Machine pane, select Create a new Virtual Machine, and click Next.
  5. In the Select a name and folder pane, enter a name for the VM. This VM will be your PureApplication Software VM. Under Select a location for the virtual machine, select the folder with the cloud group name from the tree view of the environment. Click Next. Select a name and folder pane
    Select a name and folder pane
  6. Select the compute resource (host) that the virtual machine will be registered on. Click Next.
  7. In the Select storage pane, in the table, select the datastore that you created to hold data (not the datastore for the image repository). The datastore that you select must be the same one that you exported the IBM OS Image for RedHat Linux Systems to. Click Next.
  8. In the Select compatibility pane, click Next to accept the defaults.
  9. In the Select a guest OS pane, for the Guest OS Family, select Linux. For the Guest OS Version, select Red Hat Enterprise Linux6(64-bit). Click Next.
  10. Ensure that the amount of CPU and memory that are needed for the PureApplication Software VM is accounted for in PureApplication Software's placement calculations. In the Customize hardware pane, on the Virtual Hardware tab, expand CPU and make the following changes:
    1. Set the virtual sockets to 8.
    2. Set Cores per Socket to 1.
    3. Change Reservation to 20792 MHz.
    4. Change Limit to Unlimited MHz.
    5. Change Shares to Custom, and set the number of shares to 256000. CPU settings in the Customize hardware pane
      CPU settings in the Customize hardware pane
  11. Collapse the settings for CPU.
  12. Still on the Virtual Hardware tab, expand Memory, make the following changes:
    1. Change RAM to 64 and the unit to GB.
    2. Change Reservation to 65536 MB.
    3. Change Limit to 65536 MB.
    4. Change Shares to Custom and set the value to 32400. Memory settings in the Customize hardware pane
      Memory settings in the Customize hardware pane
  13. Collapse the settings for Memory.
  14. Expand New Network, and select the network that corresponds to the Shared Data VLAN. To locate the network, click System -> Network Configuration -> VLAN and Links. New Network settings in the Customize hardware                     pane
    New Network settings in the Customize hardware pane
  15. Expand New SCSI controller, and for Change Type, select LSI Logic Parallel. Change Type setting
    Change Type setting
  16. Expand New Hard disk. On the far right side of that row, click the X to delete that entry.
  17. Click Next.
  18. In the Ready to complete pane, click Finish. Deleting entry for New Hard disk setting
    Deleting entry for New Hard disk setting
5

Set up the activation engine

  1. On the compute node that you accessed by using SSH in Step 2, create the ovf-env.xml file with the following initial contents. After you create them, you will update them (next step). Also, before you paste this file into the vi editor on the compute node, turn off autoindent by using the vi:set noautoindent command.
    <?xml version="1.0" encoding="UTF-8"?>
    <Environment xmlns="http://schemas.dmtf.org/ovf/environment/1" xmlns:ovf="http://schemas.dmtf.org/ovf/envelope/1" xmlns:ovfenv="http://schemas.dmtf.org/ovf/environment/1" xmlns:rasd="http://schemas.dmtf.org/wbem/wscim/1/cim-schema/2/CIM_ResourceAllocationSettingData" xmlns:vssd="http://schemas.dmtf.org/wbem/wscim/1/cim-schema/2/CIM_VirtualSystemSettingData" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" ovf:id="IBM OS Image for Red Hat Linux Systems/Virtual_System">
        <PropertySection>
    
    <!-- Update these network settings for this virtual machine -->
            <Property ovfenv:key="ConfigNET.netmask.1" ovfenv:value="255.255.248.0"/>
            <Property ovfenv:key="ConfigNET.gateway.1" ovfenv:value="172.20.80.1"/>
            <Property ovfenv:key="ConfigNET.pri_dns.1" ovfenv:value="172.15.248.101"/>
            <Property ovfenv:key="ConfigNET.sec_dns.1" ovfenv:value="172.15.248.102"/>
            <Property ovfenv:key="ConfigNET.ipaddr.1" ovfenv:value="172.20.84.1"/>
            <Property ovfenv:key="ConfigNET.hostname.1" ovfenv:value="mysoftwarevm"/>
            <Property ovfenv:key="ConfigNET.domain.1" ovfenv:value="domain.com"/>
            <Property ovfenv:key="ConfigNTP.ntp_0" ovfenv:value="172.15.248.101"/>
    
    <!-- Set your passwords -->
            <Property ovfenv:key="ConfigPWD_ROOT.password" ovfenv:value="password"/>
            <Property ovfenv:key="ConfigPWD_USER.password" ovfenv:value="password"/>
    
    <!-- set your ssh key – optional -->
            <Property ovfenv:key="ConfigSSH.ssh_key" ovfenv:value="ssh-rsa [public key]"/>
    
    <!-- These do not need to be changed -->
            <Property ovfenv:key="License.accept" ovfenv:value="yes"/>
            <Property ovfenv:key="ConfigLocale.language" ovfenv:value="en"/>
            <Property ovfenv:key="ConfigLocale.country" ovfenv:value="US"/>
            <Property ovfenv:key="ConfigLocale.encoding" ovfenv:value="UTF-8"/>
            <Property ovfenv:key="ConfigPWD_ROOT.username" ovfenv:value="root"/>
            <Property ovfenv:key="ConfigPWD_USER.username" ovfenv:value="virtuser"/>
            <Property ovfenv:key="ConfigSSH.username" ovfenv:value="virtuser"/>
            <Property ovfenv:key="ConfigTZ.timezone" ovfenv:value="UTC"/>
            <Property ovfenv:key="vm.cpushpri" ovfenv:value="8"/>
            <Property ovfenv:key="ConfigNET.bootproto.1" ovfenv:value="static"/>
            <Property ovfenv:key="ConfigNET.iface_name.1" ovfenv:value="eth0"/>
            <Property ovfenv:key="ConfigNET.dnsmasq.1" ovfenv:value="true"/>
        </PropertySection>
    </Environment>
  2. Edit that file and update the password entries and network settings so that they match your data VLAN environment. You will use this file to establish initial network connectivity for the virtual machine. Therefore, you must ensure that the settings are correct.
  3. Generate an ISO image file that contains the ovf-env.xml file:
    genisoimage -J -r -q -o activation.iso ovf-env.xml

    The activation.iso file is created in the temp directory that you created on the datastore that contains the VM.

  4. Change directories to return to the root of the datastore that contains the directories:
    cd ..
  5. Run the ls command to verify that you see the temp directory and a directory named after the VM that you created in Step 4.
  6. Move the activation.iso file to the virtual machine directory:
    mv temp/activation.iso <VM_DIRECTORY_NAME>

    Here, <VM_DIRECTORY_NAME> is the virtual machine directory name as in the following example:

    mv temp/activation.iso TestCloud-PSM
6

Set up the disk images

  1. From your SSH session to the compute node, change directories to your new VM directory.
  2. If you did not yet disconnect your SSH session from the previous section or chang directories, run a cd command similar to the following example, where TestCloud-PSM is the name of the VM that you created:
    cd TestCloud-PSM
  3. Extract the two VMDK files from the IBM_OS_Image_for_Red_Hat_Linux_Systems.ova file that you imported in Step 3.
    tar -xvf ../temp/*.ova RHEL65-64.vmdk RHEL65-64-flat.vmdk
  4. Rename and decompress each file that was extracted from the ova file:
    mv RHEL65-64-flat.vmdk RHEL65-64-flat.vmdk.gz
    gzip -d RHEL65-64-flat.vmdk.gz
    mv RHEL65-64.vmdk RHEL65-64.vmdk.gz
    gzip -d RHEL65-64.vmdk.gz
7

Configure disk images for the virtual machine

Use the VMware vSphere Web Client to add the ISO and disk images to the VM configuration.

  1. In the Navigator, find the VM that you created as explained in Step 4. Right-click the VM, and select Edit Settings.
  2. Modify the virtual machine to attach the activation.iso file to the CD/ROM drive:
    1. In the Edit Settings pane, click the Virtual Hardware tab.
    2. Expand CD/DVD drive 1, and change the default value of Client Device to Datastore ISO file.
    3. In the Select File window, under Datastores, select the datastore where the PureApplication Software VM was installed. The contents of the selected datastore are displayed in the Contents pane. Select the directory that you named after the PureApplication Software VM that you created. The Contents pane is updated with a list of the files in the home directory of the VM. Select the activation.iso file, and click OK. Selecting the datastore
      Selecting the datastore

      Tip: If the activation.iso file is not displayed right away, browse to a new directory and then go back to cause a refresh.

      activation.iso file in the Contents                                     pane
      activation.iso file in the Contents pane
    4. Back in the CD/DVD drive 1 pane, select Connect to connect this drive at power on.
  3. Add the extracted RHEL VMDK file as the first hard drive:
    1. At the bottom of the Edit Settings window, from the New device list, select Existing Hard Disk, and click Add.
    2. In the Select File window, under the Datastores pane, select the datastore where the VM was created. In the Contents pane, select the directory that is named after the VM. Select the RHEL65-64.vmdk file, and click OK. RHEL65-64.vmdk file in the Contents                                     pane
      RHEL65-64.vmdk file in the Contents pane
  4. Add a new hard disk large enough for PureApplication Software:
    1. At the bottom of the Edit Settings window, from the New device list, select New Hard Disk, and click the Add button.
    2. In the New Hard disk row that is displayed at the bottom of the Virtual Hardware list for the VM, change the size from 16 GB to 580 GB. Expand the details for the hard disk, and make sure that Disk Provisioning is set to Thick provision lazy zeroed. New Hard disk settings
      New Hard disk settings
  5. Click OK to save the settings that you made to the VM.
8

Configure the VM runtime

Now that you have completed the settings for the VM, power it on. Then, you can do additional configuration inside the running VM.

  1. In the VMware vSphere Web Client, right-click the VM, and select Power On to power on the VM. Powering on the VM
    Powering on the VM
  2. Wait a minute for the virtual machine to completely start the operating system.
  3. Log in to the PureApplication Software VM by using SSH as explained in Step 2, with the values that you specified for the password and IP address. In the sample ovf-env.xml file, the password is passw0rd, and the IP address is 172.20.84.1. The following example shows the SSH login for the ovf-env.xml file:
    ssh root@172.20.84.1
  4. Prepare the RedHat Enterprise Linux operating system for the PureApplication Software installation.
    1. Verify that you are logged in as the root user.
    2. Validate that you are logged in to the correct VM.
    3. Run the lsblk command, and confirm that the output shows the hard disks that are configured in this example:
      -bash-4.1# lsblk
      NAME                        MAJ:MIN RM SIZE  RO TYPE MOUNTPOINT
      sda                           8:0    0   12G  0 disk ├─sda1                        8:1    0  250M 0 part /boot
      └─sda2                        8:2    0 11.8G  0 part 
      ├─vg_root-LogVol01 (dm-0)   253:0    0    2G  0 lvm  [SWAP] 
      └─vg_root-LogVol00 (dm-1)   253:1    0  9.8G  0 lvm  /sdb                          8:16   0  580G  0 disk sr0                          11:0    1  366K  0 rom
    4. Partition the sdb disk as a root user. The following example shows both the commands and their output.
      -bash-4.1# vgcreate PureApp /dev/sdb
       Physical volume "/dev/sdb" successfully created Volume group "PureApp" successfully created
      -bash-4.1# lvcreate --size 250G --name Data PureApp 
      Logical volume "Data" created.
      -bash-4.1# lvcreate --size 250G --name Runtime PureApp 
      Logical volume "Runtime" created.
      -bash-4.1# lvcreate --size 32G --name Temp PureApp 
      Logical volume "Temp" created.
      -bash-4.1# lvcreate --size 32G --name Swap PureApp 
      Logical volume "Swap" created.
    5. Set up the temporary partition:
      -bash-4.1# mkfs.ext4 /dev/PureApp/Temp

      Then, format the swap partition:

      -bash-4.1# mkswap /dev/PureApp/Swap 
      mkswap: /dev/PureApp/Swap: warning: don't erase bootbits sectors
       on whole disk. Use -f to force.
      Setting up swapspace version 1, size = 33554428 KiB
      no label, UUID=63becdbb-4454-48c8-9347-64e5cd42a1d8
    6. Append the following two lines to the /etc/fstab file:
      /dev/PureApp/Swap    swap             swap     defaults    0 0
      /dev/PureApp/Temp    /tmp             ext4     defaults    1 2
  5. Follow the steps in the "PureSystems Manager Operating System setup" topic in IBM Knowledge Center. Begin at step 1.e. (Activate the LUKS module), and continue through to step 3. Note the following differences between the procedure in this tutorial and the procedure IBM Knowledge Center:
    • This tutorial uses the volume group name of PureApp. The procedure in IBM Knowledge Center does not specify the volume group name, but refers to the name as <volume_group_name>.
    • This tutorial uses the logical volume names Runtime and Data. IBM Knowledge Center does not specify the logical volume name. Instead, it uses the value <logical_volume_name> and expects you to know the name to use.
  6. After you reboot the VM, log in and validate that you successfully configured the virtual machine.
    1. Run the free command as shown in the following example. This command shows the memory and swap information that is available to the operating system. The total column must show the correct amount of memory and swap. PureApplication Software requires the total of those two values to be at least 80 GB.
      -bash-4.1# free -h
                  total       used       free     shared    buffers cached
      Mem:         63G        667M        62G       192K     14M     154M
      -/+ buffers/cache:      498M        62G
      Swap:        33G          0B        33G
    2. Run the df command to see which file system information is configured on the operating system. Validate the Size and Mounted on columns for the two luks partitions and the temp partition.
      -bash-4.1# df -hP
      Filesystem                    Size  Used  Avail Use% Mounted on
      /dev/mapper/vg_root-LogVol00  9.5G  2.9G   6.2G  32% /
      tmpfs                          32G     0    32G   0% /dev/shm
      /dev/sda1                     239M   52M   174M  23% /boot
      /dev/mapper/luks-runtime      246G   60M   234G   1% /pureapp-sw-runtime
      /dev/mapper/luks-data         246G   60M   234G   1% /pureapp-sw-data
      /dev/mapper/PureApp-Temp       32G   48M    30G   1% /tmp

You have now set up the VM and file systems to install PureApplication Software.

9

Install PureApplication Software

  1. Obtain and install the prerequisite software and then PureApplication Software.
    -bash-4.1# cd /pureapp-sw-runtime/
     -bash-4.1# mkdir install
     -bash-4.1# cd install/
  2. Shut down the firewall temporarily to allow the downloads to install on the VM. The PureApplication installation will replace the firewall rules.
    -bash-4.1# service iptables stop
    iptables: Setting chains to policy ACCEPT: filter          [  OK ]
    iptables: Flushing firewall rules:                         [  OK ]
    iptables: Unloading modules:                               [  OK ]
    
    -bash-4.1# service iptables status
    iptables: Firewall is not running.
  3. Install the following prerequisite RPMs from the Extra Packages for Enterprise Linux (EPEL) site:
    • mod_security-2.7.3-3.el6.x86_64.rpm
    • mod_security_crs-2.2.6-3.el6.noarch.rpm
    • mod_security_crs-extras-2.2.6-3.el6.noarch.rpm
    1. Download the three RPM files to the current directory.
    2. Use yum to install these RPMs as shown in the following example because it has prerequisites that are available in a local package repository:
      -bash-4.1# yum install -y ./mod*.rpm

      The following snippet of the command output shows the packages that were installed:

      Installed:
        mod_security.x86_64 0:2.7.3-3.el6                      		mod_security_crs.noarch 0:2.2.6-3.el6 
        mod_security_crs-extras.noarch 0:2.2.6-3.el6 
      
      Dependency Installed:
        apr.x86_64 0:1.3.9-5.el6_2                           			apr-util.x86_64 0:1.3.9-3.el6_0.1 
        apr-util-ldap.x86_64 0:1.3.9-3.el6_0.1               			httpd.x86_64 0:2.2.15-54.el6_8 
        httpd-tools.x86_64 0:2.2.15-54.el6_8 
      
      Complete!
  4. Complete the steps in the "Installing PureApplication Software" topic in IBM Knowledge Center. Note the following changes:
    • Use the /pureapp-sw-runtime/install directory as the temporary directory in step 2.
    • Use software_system as the installation type in step 5c.

The installer can run for several hours to upload the default content. After it is done running, continue with the next step.

10

Configure PureApplication Software

You are now ready to configure PureApplication Software to deploy and manage workloads. You complete these steps in the Software web console.

  1. Open a web browser, and go to the URL http://<IP-address>, where <IP-address> is the IP address of the VM that you created and logged in to. In this example, the IP address is 172.20.84.1. Therefore, you go to the URL http://172.20.84.1.
  2. Log in to the web console with the administrative user and password that were specified when you installed PureApplication Software.
  3. Click System, and under Settings, click System Settings to go to the PureApplication Software web console. Accessing the system settings
    Accessing the system settings
  4. Expand the Cloud Management IP section, and enter the following information for your configuration:
    1. Management IP. Enter the IP address of eth0 on the PSM VM. In this example, the IP address is 172.20.84.1. You cannot change this address after you set it because deployed VMs in pattern instances will use it to communicate to the pattern engine of the PureApplication Software VM. Therefore, you must ensure that this IP address is going to be used for communication back to the PSM.
    2. Hostname. Enter the host name that is returned by the DNS server for the Management IP address. In this example, the host name that is returned by the DNS server for the IP address 172.20.84.1 is ipas-pvm-084-001.
    3. Domain Name. Enter the domain name that is returned by the DNS server for the Management IP address. In this example, the domain name that is returned by the name server is purescale.raleigh.ibm.com.
    4. Netmask. Enter the netmask that is used by the Management IP address on eth0 of the PSM Virtual Machine. In this example, the netmask is 255.255.248.0.
    5. Default Gateway. Enter the default gateway on the PSM Virtual Machine. In this example, the default gateway is set to 172.20.80.1.
    6. VLANS. Enter the comma-separated list of VLANs that can be configured in your IP groups. This list must contain the VLAN that you used for eth0 on the PSM VM. In this example, VLAN 1864 is used for creating the PSM VM.
    7. VLANS properly. If you use this solution for disaster recovery, enter a data VLAN that is accessible on both racks. When you specify this VLAN, it must be available on both Bluemix Local System in the current environment and on the backup Bluemix Local System rack. Also, you must deploy all of your deployments for failover to this VLAN (an IP group that is defined with this VLAN).
    8. After you validate that all of your settings are correct, click Save Changes.
  5. Expand the Virtual Center Access section, and configure the settings.

    To complete the configuration settings in this section, you can get the values from the system console. Click SystemSystem Settings, and expand External Application Access Settings. Click the details icon (highlighted in the following figure) for the user that you created to access vCenter from this PureApplication Software environment.

    Accessing the user details
    Accessing the user details

    The External Users pane that is displayed contains the access information for the compute nodes and vCenter VM. The following settings come from the Virtual Manager row of the table.

    1. For the vCenter IP field, enter the information from the IP Address column.
    2. For the User name field, enter the information from the User Name column.
    3. For the Password field, enter the information from the Password column.
    4. Click Test Connection to validate your settings. After a moment, a message is displayed that indicates that the connection test was successful.


      For the next two configuration settings in this section, use the VMware vCenter web console.

    5. For the Datacenter field, from the VMware vCenter navigator view, click Hosts and Clusters. The object under the purevc node is the name of the datacenter, which is Rack-8283_Rack32 in this example. Datacenter name under the purevc                                     node
    6. For the Distributed Virtual Switch field, again in the navigator, on the Related Objects tab, select Distributed Switches. The object that appears in the table is the name of the distributed virtual switch. (In vSphere Client, go to Inventory > Networking.) The switch name is similar to the name and format of DVS-Rack-8283_Rack32 as used in this example. Distributed virtual switch
      Distributed virtual switch
  6. Click Save Changes to save the configuration settings in this section.
  7. Expand Date and Time to access the settings for the NTP servers.
    1. In the NTP server address field, enter the NTP server IP address, and click Add. Repeat this step if you have multiple NTP servers to specify.
    2. After you add the NTP server addresses, click Save.
  8. Expand Domain Name Service (DNS) to access settings for the DNS servers.
    1. In the DNS service search order field, enter the DNS server IP address, and click Add. Repeat this step if you have multiple DNS servers to specify.
    2. After you add the DNS server addresses, click Save.
11

Discover compute resources

  1. In the upper-left corner of the software console, click IBM PureApplication Software. On the Welcome page, you see green check marks next to each step that is associated with step 1.
  2. Under step 2, validate that the first item (Connect to external virtual center server) has a green check mark, and click Discover compute resources.
  3. Click the eye or discover icon to begin discovery of the compute nodes. Discovery of compute nodes
    Discovery of compute nodes
  4. Wait a few minutes, and then click the refresh icon.
  5. Enter the credentials of the compute node so that it changes from the Discovered state to the Available state.
    1. Select the compute node, and then click Update. Updating the compute node stat
      Updating the compute node stat
    2. Enter the user, password, and IP of the compute node.

      The compute node information is in the "External Application Access Settings" section of the System Settings page on the system console. Click the details icon for the user that you created to access vCenter from the PureApplication Software workload environment.

      Accessing the vCenter
      Accessing the vCenter

      The External Users pane that is displayed contains access information for compute nodes and the Virtual Manager (vCenter). The following four settings come from the Compute Node row in this pane and match the compute node that you deployed the PureApplication Software VM to.

    3. For the User name field, enter the information from the User Name column.
    4. For the Password field, enter the information from the Password column.
    5. Optional: For the Location field, enter the information from the Name column.

      Although this field is optional, it is helpful in identifying which compute nodes are in use by PureApplication Software. As a best practice, enter the serial number of the compute node here.

    6. For the IP address field, enter the information from the IP Address column.
    7. Click OK to save the changes.
    8. Wait a minute, and refresh the software console. The state of the compute node changes from Discovered to Available.
12

Discover storage

  1. Back on the Welcome page of IBM PureApplication Software, click Discover storage resources. You see a green check mark because discovery found the storage when it discovered the compute node. The storage volumes created for this environment are shown on the LUNs tab of the storage table.
  2. Click the Data Stores tab. The capacity columns in the table show that some space is used for the datastore that contains the PureApplication Software VM. The other datastore in the table was created as the image repository.
  3. In the Is image repository? column, select the check boxes for the datastore that is not used, but that was created for the image repository.
  4. Verify that you selected the correct datastore.
  5. Click OK in the confirmation pane.
13

Create IP groups

  1. Back on the Welcome page of IBM PureApplication, click Create IP Groups.
    Click the plusicon to add a new IP group.
  2. Define the settings for this IP group. The networking settings for this IP group must match your configuration settings in Bluemix Local System that use the same VLAN. The In cloud group setting must be the Virtual Manager cloud group that was created for this VM.

    Note: The VLAN that you select for your IP groups must be shared.

  3. Click OK to create the IP group. Adding a new IP group
    Adding a new IP group
  4. After you create the IP group, under IP addresses, click the plus sign to add IP addresses to the IP.
  5. Specify a range of IP addresses, and then click OK.
    Notes:
    • VMs that are deployed directly to vCenter (such as the PureApplication Software VM) must have IP addresses that do not conflict with IP addresses in the IP groups that you create in this step.
    • IP addresses in IP groups on Bluemix Local System must not overlap with the IP addresses that you add to the IP groups in PureApplication Software.

The IP group status changes to "Available" after the IP addresses are associated with the IP group.

14

Create environment profiles

  1. Again, on the Welcome page of IBM PureApplication Software, click Create environment profiles.
  2. On the Environment Profiles page, click the plus sign to add a new environment profile.
    1. Specify a name for the environment profile.
    2. Optional: In the Description field, enter a description for this environment profile. You can leave the remaining fields at their default values.
    3. Click OK.
      The status of the environment profile has a warning message, indicating that the cloud group is not associated yet.
  3. To associate the cloud group, select the new environment profile.
    1. For Deploy to cloud groups, click the Add more field. From the list of cloud groups that is displayed, select the cloud group. An entry for the cloud group is now listed in a table.
    2. Click the plus sign to expand the settings for the cloud group. The IP group that you created is listed in the settings.
    3. Select the check box next to the IP Group. The environment profile now changes from the Warning state to the Available state.

Additional configuration tasks

The following sections explain the additional tasks that you might need to finish setting up your PureApplication Software workload environment. Whether you need to follow a particular task depends on your environment and requirements. Therefore, review each of the tasks to see which ones apply to the needs of your environment.

Allocate more storage

You might need to allocate more storage to deploy VMs within the PureApplication Software installation. Create volumes that you know you will need by using this two-part procedure:

  1. Create the storage volumes on the web console in Bluemix Local System.
  2. Discover the storage volumes on the web console in PureApplication Software.

Create storage volumes

  1. On the Bluemix Local System web console, click Cloud -> Volumes.
  2. Click Create New to create a new storage volume.
  3. In the Create a new volume pane:
    1. Enter a name and description for the volume.
    2. For Type, select one of the following values:
      • Block VMFS. Use this type to create a datastore that attaches to the compute nodes in the cloud group to contain the VMs, add-on disks, and snapshots.
      • Block. Use this type to create raw storage that will be available to a VM.
    3. For In Cloud Group, select the Virtual Manager cloud group that the storage will be associated with. If you plan to replicate the software workload environment to another Bluemix Local System for disaster recovery, assign all of the volumes that are associated with the software environment to a volume group. In this case, select the correct volume group.
    4. Under Volume configuration, for Size, enter an appropriate size for the purpose of your volume.
    5. Click OK to save the changes and create this volume. A job starts that creates the volume in the background.

Repeat this task for additional storage as needed for your requirements.

Image repository name: PureApplication Software requires an image repository that, before V2.2.3, required the name to be cache_2TB-1. Do not use this name in this environment because it will conflict with the image repository of PureApplication System.

Storage discovery in PureApplication Software

After you assign storage to a Virtual Manager cloud group in Bluemix Local System, switch to the web console for PureApplication Software.

  1. Click Hardware -> Storage Resource.
  2. Click the eye icon to discover new storage.
  3. After a few moments, click the refresh icon to see the new storage that was discovered.
  4. If the storage that was discovered is a Block volume, rename the storage volume so that it matches the name that is used on the Bluemix Local System web console. The names do not need to be identical, but the name must be usable to identify the volume if this environment is part of a disaster recovery solution. To rename the volume:
    1. Select the name in the list of volumes.
    2. Click Save.

Compute node high availability

When a compute node fails, you want to ensure that your workloads do not experience an outage. Therefore, you can designate a compute node as the high availability (HA) dedicated node. This way, you can reserve a compute node per cloud group as a standby node that can take over workloads when a compute node in the cloud group fails.

Turning on vSphere HA: After you complete these steps, you must turn on vSphere HA to ensure that the VM recovery works correctly. Turning on vSphere HA is required even you if you do not want to configure a spare node as a high availability dedicated node because you do not have enough capacity or you are working in a development or test environment.

To configure a compute node as the HA node for a cloud group:

  1. On the VMware vSphere Web Client, right-click the cluster and select Settings. Selecting the vSphere Web Client                             settings
    Selecting the vSphere Web Client settings
  2. On the Manage tab, under Settings, select vSphere HA. Then, in the right pane, click Edit. Editing vSphere HA
    Editing vSphere HA
  3. In the Edit Cluster Settings pane, select Turn on vSphere HA.
  4. Expand Admission Control. Expanding Admission Control
    Expanding Admission Control
  5. From the list of Admission control policies, select the Use dedicated failover hosts. Use dedicated failover hosts option
    Use dedicated failover hosts option
  6. Click the plus icon.
  7. Select the host to use as the dedicated failover host. If you have multiple hosts to choose from, select the one that has the most capacity and is empty. Click OK to save your host selection.
  8. Click OK to save the cluster settings.

For more information, see the "Enabling compute node high availability" topic in IBM Knowledge Center.

Create a new virtual image

Installing the PureApplication Software workload environment does not come with a base operating system virtual image. You can import the RedHat Enterprise Linux V6 64-bit virtual image (OVA image) from your Bluemix Local System as explained in Step 3. To make that image available in this installation:

  1. Using SSH, log in to the compute node that you used to set up the VM in Step 3.
  2. Change the current working directory to the same temporary directory that you created on the datastore where you created the VM:
    cd /vmfs/volumes/p_9baf11f6-6d8c-4faa-b631-5458655a8ecb/temp
  3. Print the current working directory by running the pwd command. Use this path in step 6 of this task.
  4. Use SSH to log in to the PureApplication Software VM.
  5. Change to the installation directory:
    cd /pureapp-sw-runtime/install
  6. Download the operating system image from the ESXi. Substitute the user ID and IP address with the values for the compute node that you logged in to. For the path, substitute the path that is shown with the value in step 2 in this task.
    scp app_77a6b935@172.26.120.232:/vmfs/volumes/p_1000edda-a501-4fef-b243-4f7396915ea4/temp/*.ova .

    When you are prompted for a password here, use the password for this compute node in the external access settings that is associated with the user ID and IP address you used.

  7. In the software web console, click Catalog → Virtual Images.
  8. At the top of the page, click Create New.
  9. In the Create virtual image pane:
    1. Verify that Create by importing an OVA image is selected.
    2. For the OVA file location text box, enter file:/pureapp-sw-runtime/install/IBM_OS_Image_for_Red_Hat_Linux_Systems.ova. Create virtual image pane
      Create virtual image pane
    3. Click OK. The import of the virtual image begins and takes about 10 minutes to complete.
  10. After the image is imported, click the new virtual image. You must accept the licenses that are associated with this virtual image before you can use the virtual image.
  11. In the details view of the virtual image in the right pane, for License agreement, click accept. Virtual image license agreement
    Virtual image license agreement
  12. In the pane that shows the licenses that are associated with the virtual image, accept each license.

    The files in the temporary directory on the compute node are no longer needed and must be cleaned up to allow PureApplication Software workload environments to make full use of the datastore.

  13. In the SSH session for the compute node, remove only the temporary directory that was created and its contents:
    rm -rf /vmfs/volumes/p_9baf11f6-6d8c-4faa-b631-5458655a8ecb/temp

Clean up the production workload environment

You can remove an entire workload environment that is in production. If you set up the environment previously in disaster recovery, before you complete this task, you must complete the steps as explained in Part 3.

  1. Stop all the workload instances from the software web console.
    1. Click Patterns -> View All. Viewing all patterns
      Viewing all patterns
    2. From the four tables that are shown that contain pattern instances for each pattern type, select an instance from one of the tables. The pattern types that you can choose from are Virtual Applications, Virtual Systems, Virtual Systems (Classic), and Shared Services. Selecting a pattern instance
      Selecting a pattern instance
    3. Click Stop in the details view to stop the instance. Stop the other instances of that same type.
    4. Repeat these steps for each pattern type. Wait for all of the instances to finish stopping.
  2. Store all of the VMs by using the software web console:
    1. Click Patterns->View All. All virtual machines should be in the Stopped state.
    2. One at a time, select each stopped VM, and click Store next to the Status field. Storing each stopped VM
      Storing each stopped VM
    3. In the confirmation pane, click Store to proceed with storing the instance. Notice that Virtual System Classic instances show the Store button in the list of actions for the instance instead of next to the Current status.
    4. Verify that all the VMs, except the PureApplication Software VM, are marked as Stored in the Virtual Machines view in PureApplication Software for the environment. (In the vSphere Web Client, the only registered VM is the PureApplication Software VM.)
  3. Run SSH to log in to the PureApplication Software VM as root (if the root password is known) or as admin_shell with the PureApplication Software admin password. Then, run psm shutdownto turn off the software VM in a clean manner.
    $ ssh admin_shell@172.25.127.1admin_shell@172.25.127.1's password: 
    Welcome to the IBM PureApplication System Administrative Interface
    PureApplication>psm shutdown
    Starting to shutdown the local PureSystems Manager...
    Broadcast message from admin_shell@ipas-vm25-127-001(unknown) at 17:50 ...
    The system is going down for power off NOW!
    Connection to 172.25.127.1 closed by remote host.
    Connection to 172.25.127.1 closed.
  4. In the vSphere Web Client, select the PureApplication Software VM that is now stopped.
    1. Remove this VM from inventory. Right-click the VM, and select Remove from Inventory. Removing the VM from inventory
      Removing the VM from inventory
    2. In the Confirm Remove pane, click Yes.

      Other VMs in the cluster: If any other VMs are still listed under this cluster, stop and remove them from inventory, too.

  5. In the vSphere Web Client, refresh the view, and validate that no VMs are registered. Refreshing the vSphere Web Client view
    Refreshing the vSphere Web Client view
  6. On the Bluemix Local System web console that is hosting the environment:
    1. Click Cloud ->Volumes.
    2. Find all of the volumes that were used in the environment. Click the check box next to their name, and then click the trash can icon above the table to delete them. Deleting volumes from the environment
      Deleting volumes from the environment
    3. In the Batch Action pane, click Delete to validate the deletion of multiple volumes.

    Cleaning up from a disaster recovery practice failover: If the practice environment is no longer needed, make sure that the volumes that are chosen are the clones. They cannot be the replicating volumes that are still being synchronized with the production environment.

  7. On vSphere Web Client, on each of the compute nodes in the Virtual Manager cloud groups, verify that all the storage is removed, except for the 5.2 GB boot LUNs for the ESXi. To find the storage devices:
    1. Select the compute node.
    2. On the Manage tab, select Storage to see the list of storage devices. Verifying that storage is removed
      Verifying that storage is removed
  8. In the vSphere Web Client, select each cluster, and then turn off vSphere HA.
    Select the cluster. On the Manage tab, under Settings, select vSphere HA from the list of Services. vSphere HA host monitoring setting
    vSphere HA host monitoring setting

    If you see the "vSphere HA is Turned ON" message at the top of the details view, click Edit, and in the Edit Cluster Settings pane, expand Admission Control. Ensure that the Admission control policy is set to Do not reserve failover capacity.

  9. Optional: If you no longer need the cloud groups, delete the compute nodes from the Virtual Manager Cloud Groups and the Virtual Manager cloud groups.

Conclusion

By completing the steps in Part 2 of this series, you have installed a working PureApplication Software environment on to Bluemix Local System. This PureApplication Software environment is fully configured and is ready to be used to develop, deploy, and manage patterns. For more information about using and administering PureApplication Software, see the IBM PureApplication Software for x86 V2.2.3 documentation in IBM Knowledge Center.

You can design disaster recovery and backup solutions around using this software environment on Bluemix Local System. Part 3 of this series explains a disaster recovery solution for this environment.

Acknowledgements

The authors thank Kevin Cormier, Larry Hamann, Jessica Stevens, and Joe Wigglesworth for reviewing this article.


Downloadable resources


Related topics


Comments

Sign in or register to add and subscribe to comments.

static.content.url=http://www.ibm.com/developerworks/js/artrating/
SITE_ID=1
Zone=Middleware, Cloud computing
ArticleID=1048044
ArticleTitle=Hosted VMware environments and recovery solutions in IBM Bluemix Local System, Part 2: Setting up a PureApplication Software workload environment
publish-date=10302017