IBM Spectrum Accelerate command-line deployment
The command-line deployment kit is a self-contained application that features all of the necessary files that are required to install IBM Spectrum Accelerate. The IBM Spectrum Accelerate deployment kit is available for Microsoft Windows and Linux systems.
This chapter describes the steps to install an IBM Spectrum Accelerate system by using the command-line version of the deployment kit for Windows and Linux hosts.
This chapter includes the following topics:
3.1 Preparing the environment for IBM Spectrum Accelerate
Before the IBM Spectrum Accelerate system can be deployed, the ESXi host servers and network environment must be properly configured. This configuration includes preparing the physical network connectivity and reserving TCP/IP addresses for the IBM Spectrum Accelerate system management. Configuration also includes reserving iSCSI connections and TCP/IP addresses for the ESXi server management connections. The ESXi servers must be properly configured to work as IBM Spectrum Accelerate modules, and have a local data store and the virtual networking infrastructure setup correctly. This section reviews the preparation of the environment.
3.1.1 Network configuration
Preparing the network for an IBM Spectrum Accelerate system requires that sufficient physical connectivity be in place, and that the virtual networking switches be associated with the physical network and the virtual network ports. The physical network connectivity must contain 10 Gbps Ethernet (GigE) connections to support the IBM Spectrum Accelerate module interconnection. In addition, it is best to provide multiple physical network device connections to each IBM Spectrum Accelerate module to ensure network redundancy in case of network device failure. For more information about the physical network configuration, see 6.2.1, “Network infrastructure” on page 125.
There must be sufficient independent subnets to properly segregate the logical networks on which the IBM Spectrum Accelerate system relies. These subnets include a separate subnet for the following tasks:
•ESXi host management
•IBM Spectrum Accelerate system management
•IBM Spectrum Accelerate module interconnect traffic
•IBM Spectrum Accelerate iSCSI host attachment
The IBM Spectrum Accelerate system management TCP/IP addresses can be within the same subnet as the ESXi host management TCP/IP addresses. Managing the host ESXi server and the XIV require little physical connectivity, and can be connected by using 10/1000 Ethernet.
The iSCSI host attachment and IBM Spectrum Accelerate module interconnect networks must be on separate subnets and optimally separate physical 10 Gbps Ethernet connections to provide sufficient bandwidth to support high workloads.
The configuration of the ESXi host must contain three virtual switches that are dedicated to IBM Spectrum Accelerate system management, iSCSI host connectivity, and module interconnection traffic. Each of these virtual switches must contain a port group that is dedicated to these functions and named appropriately. The port group names are used within the IBM Spectrum Accelerate deployment kit configuration XML file to identify the networks.
3.1.2 ESXi server configuration
The VMware ESXi servers must be correctly configured with appropriate data store and virtual network settings. IBM Spectrum Accelerate systems require a data store with at least 250 GB of available space for the system files. This data store can be local to the ESXi server to reduce the amount of latency that the system experiences during operation. For more information, see Chapter 2, “Planning for IBM Spectrum Accelerate” on page 13. A conceptual depiction of the ESXi networking is shown in Figure 1-4 on page 9.
 
Note: The IBM Spectrum Accelerate support for converged infrastructures that was introduced in version 11.5.1 allows multiple virtual machines to coexist with IBM Spectrum Accelerate virtual machine. There must be sufficient memory, processing cores, and networking connectivity within the server to dedicate to the IBM Spectrum Accelerate virtual machines.
The IBM Spectrum Accelerate virtual machine must allocate a minimum of 24 GB of memory, four CPU cores, and six disk drives that can be used in RDM mode.
A VMware datastore must be defined with enough extra space. The amount of required extra space is described in 2.4, “Deployment host” on page 25. You must use a local disk for that purpose, as shown in Figure 3-1.
Figure 3-1 ESXi data stores content before IBM Spectrum Accelerate deployment
The ESXi host network configuration requires virtual switches be configured to support the subnets that are described in 3.1.1, “Network configuration” on page 38. Figure 3-2 shows an example of these virtual switches and how they are mapped to the physical connections. Each vSwitch contains a port group with an identifying network name. When deploying the IBM Spectrum Accelerate system, these port group names are what is used to determine the network identity.
Figure 3-2 ESXi network configuration in preparation of IBM Spectrum Accelerate deployment
For more information about ESXi servers and environments, see Chapter 2, “Planning for IBM Spectrum Accelerate” on page 13.
3.1.3 VMware vSphere configuration for converged infrastructure
IBM Spectrum Accelerate version 11.5.1 introduced support for converged infrastructures. This feature allows the installation of an IBM Spectrum Accelerate system onto an ESXi cluster that features virtual machines. To complete the installation of an IBM Spectrum Accelerate system with converged infrastructure support, each ESXi server within the cluster must meet the minimum resource requirements to host the IBM Spectrum Accelerate system virtual machines.
The following minimum resources are required to install and operate an IBM Spectrum Accelerate system virtual machine:
•CPU Cores
The ESXi server must dedicate at least four cores to the IBM Spectrum Accelerate virtual machine. For converged infrastructure support, the ESXi server requires enough CPU cores to operate the IBM Spectrum Accelerate virtual machine, co-existing virtual machines, and the base ESXi server software.
•Memory
The ESXi server must dedicate at least 24 GB of memory to the IBM Spectrum Accelerate virtual machine. For converged infrastructure support, the ESXi server requires enough memory to operate the IBM Spectrum Accelerate virtual machine, co-existing virtual machines, and the base ESXi server software.
•Raw Disk Devices
The ESXi server must dedicate at least six raw disk drives directly to the IBM Spectrum Accelerate virtual machine. For converged infrastructure support, the ESXi server requires enough storage devices so that the following requirements can be met:
 – The IBM Spectrum Accelerate virtual machine can be assigned at least six local disk drives.
 – The co-existing virtual machines have access to other storage.
The base ESXi server software also requires local storage resources for its system software.
•Network Resources
The ESXi server must dedicate at least one 10 Gbps Ethernet connections to the IBM Spectrum Accelerate virtual machine to support the interconnect and iSCSI networks. For converged infrastructure support, the ESXi server requires that enough networking resources be available so that the IBM Spectrum Accelerate virtual machine can be assigned the network resources. At the same time, co-existing virtual machines retain access to their network resources. This requirement includes ensuring that there are sufficient unique TCP/IP addresses and addressable subnets available for the IBM Spectrum Accelerate system, co-existing virtual machines, and the base ESXi server.
 
Note: A minimum of two 10-GigE ports is needed for fault tolerance and load balancing.
•SSD s (optional)
IBM Spectrum Accelerate systems allows the use of an optional solid-state drive (SSD) to accelerate read performance. If the SSD Caching feature is used, the ESXi server must dedicate at least one 500 GB SSD device directly to the IBM Spectrum Accelerate virtual machine. For converged infrastructure support, the ESXi server requires enough SSDs so that the IBM Spectrum Accelerate virtual machine can be assigned a disk drive, while co-existing virtual machines have access to any SSDs they might need.
For more information about the requirements for an IBM Spectrum Accelerate with converged infrastructure support, see Chapter 2, “Planning for IBM Spectrum Accelerate” on page 13.
Converged infrastructure and vSphere high availability feature
VMware ESXi servers can be assigned to high-availability clusters to reduce the effect of hardware or connectivity failures. Because of the inherent redundancy of the IBM Spectrum Accelerate system’s grid architecture, there is no need for the IBM Spectrum Accelerate virtual machines to be added to a high-availability cluster, and doing so causes performance and stability problems. If the vSphere high availability clustering feature for ESXi is used on co-existing virtual machines, ensure that the IBM Spectrum Accelerate system is not monitored and managed.
Complete the following steps to ensure that the IBM Spectrum Accelerate virtual machines are excluded from high availability monitoring:
1. Connect as an administrative user to the VMware vCenter that manages the ESXi servers IBM Spectrum Accelerate that is deployed against.
2. Right-click the cluster that contains the individual ESXi servers and select Edit Settings to open the cluster virtual machine settings.
3. Select the VM Monitoring option that is under the VMware HA heading in the left window pane. The high-availability cluster monitoring settings are completed in the main window.
4. Ensure that the high availability monitoring for the IBM Spectrum Accelerate virtual machines option is Disabled on all ESXi servers within the cluster, as shown in Figure 3-3.
Figure 3-3 Disabling vSphere HA monitoring for the IBM Spectrum Accelerate virtual machines
3.2 Preparing the IBM Spectrum Accelerate deployment XML
Before deploying an IBM Spectrum Accelerate system, ensure that you completed the pre-deployment tasks that are described in Chapter 2, “Planning for IBM Spectrum Accelerate” on page 13.
Also, ensure that the network and the ESXi infrastructure are correctly configured. This configuration includes verifying the correct number of iSCSI ports and IP addresses, hard disk drives (HDD), and the optional SSDs are used as the base hardware for your Spectrum Accelerate system.
Deploying an IBM Spectrum Accelerate system requires a deployment configuration XML file. This configuration XML file contains the settings for the IBM Spectrum Accelerate system and the ESXi servers on which it is deployed. The IBM Spectrum Accelerate deployment kit includes an XML template file that can be edited to conform to the target environment. The XML file includes the following main sections:
•The first section of the deployment XML file contains all of the system-wide configuration settings for the IBM Spectrum Accelerate system. This section includes components, such as the system’s name, the amount of memory to dedicate to the system, the number of drives, and other system-wide settings. All configuration within this section pertains to the overall layout of the system.
•The second section of the deployment XML file is used to define a vCenter host. This section is optional. If vCenter is not being used, this section can be removed or commented out from the file.
•The third section of the deployment XML file is used to define the individual ESXi servers settings. After deployment, each server is an individual IBM Spectrum Accelerate module. This section includes defining components, such as the ESXi data store name on each server, the network port group names, and access controls.
 
Note: The deployment template XML file can be obtained by running the deployment kit with the -e argument. See Example 3-1 for an example on a Linux system.
Example 3-1 Obtaining the template XML file to prepare for command line deployment
itsouser: ~$ ./xiv_sds_deployment_kit-11.5.1.c-git-8f8231f5a544.bash -e
Searching embedded archive for example files... Please wait.
SDS_Deploy/deploy/sds_machine_template.xml
Examples available in ./XIV_Deployment_Examples/
itsouser: ~$ ls ~/XIV_Deployment_Examples
sds_machine_template.xml
Example 3-2 shows the XML template file with explanations of what each value represents. This template XML lists a single ESXi server stanza only. IBM Spectrum Accelerate systems require a minimum of three ESXi servers to successfully deploy. Required fields that must be defined for successful deployment are shown in bold in the example.
Example 3-2 Sample xml file
<?XIVSDS xml_version="1.4.0"?>
<sds_machine
name="UNIQUE_NAME_FOR_THE_MACHINE"
interconnect_mtu="A_NUMBER_1500_TO_9000"
vm_gateway="GATEWAY_FOR_MGMNT_NETWORK"
vm_netmask="NETMASK_FOR_MGMENT_NETWORK"
data_disks="NUMBER_OF_RDM_DATA_DISKS_6_TO_12"
memory_gb="MEMORY_FOR_VM_IN_GB_24_TO_128"
icn="SEVEN_DIGIT_IBM_CUSTOMER_NUMBER"
ssd_disks="NUMBER_OF_RDM_SSD_DISKS_0_OR_1"
off_premise="YES_OR_NO"
num_cores="NUM_OF_CORES_4_TO_20"
enable_diagnostic_mode="YES_OR_NO"
data_disks_size_gb="WANTED_DISK_SIZE_IN_GB_IN_1TB_INCREMENTS"
ssd_disks_size_gb="WANTED_SSD_SIZE_IN_GB">
 
Important: The IBM Spectrum Accelerate system <sds_machine> stanza includes the following components:
•name: This value sets the IBM Spectrum Accelerate system name.
•interconnect_mtu: This value defines the packet size that is used over the interconnect network.
•vm_gateway: This value defines the management gateway IP address.
•vm_netmask: This value defines the netmask for the vm_gateway IP address.
•data_disks: This value defines the number of RDM disks that is allocated in each module.
•memory_gb: This value defines the amount of memory in GiB that is dedicated to each VM.
•icn: This value defines the IBM Customer Number that is used to access support.
•ssd_disks: This value defines the number of SSDs in each module.
•off_premise: This value indicates whether the deployment is on-premises or in a cloud environment.
•num_cores: This value defines the number of cores each VM uses.
•enable_diagnostic_mode: This value defines whether the system performs a post test of the system following deployment.
•data_disks_size_gb: This value defines the size of disks that is used across the system.
•ssd_disks_size_gb: This value defines the size of SSDs that are used across the system.
<vcenter_host
name="VCENTER_HOSTNAME_OR_IP"
datacenter="NAME_OF_EXISTING_DATACENTER_ON_VCENTER"
username="VCENTER_ADMIN_USERNAME"
password="VCENTER_ADMIN_PASSWORD">
</vcenter_host>
 
Optional: The IBM Spectrum Accelerate vCenter <vcenter_host> stanza includes the following components:
•name: This value defines the host name or IP address of the vCenter host.
•datacenter: This value defines the datacenter that contains the ESXi servers that are used as IBM Spectrum Accelerate modules.
•username: This value defines the user ID of the Administrative user of the vCenter host.
•password: This value defines the password of the Administrator for the vCenter host.
<esx_servers>
<server
hostname="NAME_OR_IP_OF_ESX_SERVER"
username="root"
password="ESX_ROOT_PASSWORD"
Data store="DATASTRORE_NAME_ON_SERVER"
mgmt_network="NAME_OF_VSWITCH_TO_BE_USED_AS_MGMT_NETWORK"
iscsi_network="NAME_OF_VSWITCH_TO_BE_USED_AS_ISCSI_NETWORK"
vm_mgmt_ip_address="IP_ADDRESS_TO_BE_GIVEN_TO_THE_MGMT_NETWORK"
interconnect_ip_address="IP_ADDRESS_TO_BE_USED_FOR_THIS_VM"
interconnect_ip_netmask="NETMASK_FOR_THE_INTERCONNECT_ADDRESS"
interconnect_network="VSWITCH_TO_BE_USED_AS_INTERCONNECT_NETWORK">
</server>
 
Important: The IBM Spectrum Accelerate ESXi server <esx_servers> stanza includes the following components:
•hostname: This value defines the hostname or IP address of the ESXi server.
•username: This value defines the user ID of the Administrative user of the ESXi server.
•password: This value defines the password for the Administrator for the ESXi server.
•Data store: This value defines the name of the ESXi server data store where the Spectrum Accelerate system software is installed.
•mgmt_network: This value defines the port group name of the management network on the ESXi server.
•iscsi_network: This value defines the port group name of the iSCSI network on the ESXi server.
•vm_mgmt_ip_address: This value defines the IP address that is used to manage the IBM Spectrum Accelerate system with the IBM XIV Management tools.
•interconnect_ip_address: This value defines the IP address that is used for interconnection traffic.
•interconnect_ip_netmask: This value defines the netmask for the subnet of which the interconnect IP is a member.
•interconnect_network: This value defines the port group name of the interconnect network on the ESXi server.
</esx_servers>
 
Tip: Add extra ESXi servers to the <esx_servers> stanza as more <server> </server> sub-stanzas.
</sds_machine>
A prepared deployment XML can also be generated by using the IBM XIV Management GUI, IBM Spectrum Accelerate deployment tool, which is described in Chapter 4, “IBM Spectrum Accelerate GUI-based deployment” on page 55. You can export a completed deployment XML when all of the deployment values are entered in the GUI, which simplifies creating the XML file.
A customized XML file is shown in Example 3-5 on page 47.
3.3 Windows command-line deployment procedure
The IBM Spectrum Accelerate system deployment kit is delivered as a compressed archive that is available from IBM Passport Advantage. For more information about acquiring the deployment kit, see 2.3.2, “Software download” on page 24.
3.3.1 Obtaining and extracting the Windows deployment kit
Obtain the IBM Spectrum Accelerate deployment kit from IBM Passport Advantage. The package name for the Windows version is named something similar to xiv_sds_deployment_kit.zip. Save the deployment kit archive in a working directory.
Example 3-3 uses a directory on the C drive called C:SpectrumAccelerateSDS_Deploy_Win.
 
Tip: To avoid potential issues, keep the path length for the deployment kit fewer than 255 characters and ensure that the path does not contain any folders with spaces in the name.
When the deployment kit is placed into a directory, extract the file. Example 3-3 displays a deployment kit archive after it is extracted. Future versions might differ, but the structure is similar.
Example 3-3 Contents of the deployment utility
c:SpectrumAccelerateSDSSDS_Deploy_Win>dir
 
Directory of c:SpectrumAccelerateSDSSDS_Deploy_Win
 
01/06/15 08:32 PM 343 copywin.py
02/20/15 08:50 AM <DIR> deploy
02/20/15 08:32 AM <DIR> DLLs
02/20/15 08:50 AM <DIR> Lib
02/20/15 08:31 AM <DIR> Libs
01/06/15 08:32 PM 14,528 LICENSE
01/06/15 08:32 PM 1,857 Microsoft.VC90.CRT.manifest
01/06/15 08:32 PM 224,768 msvcm90.dll
01/06/15 08:32 PM 568,832 msvcp90.dll
01/06/15 08:32 PM 655,872 msvcr90.dll
01/06/15 08:32 PM 359,882 NEWS
01/06/15 08:32 PM 27,136 python.exe
01/06/15 08:32 PM 2,448,384 python27.dll
01/06/15 08:32 PM 357,888 pythoncom27.dll
01/06/15 08:32 PM 8,192 pythoncomloader27.dll
01/06/15 08:32 PM 10,752 PythonService.exe
01/06/15 08:32 PM 27,136 pythonw.exe
01/06/15 08:32 PM 109,056 pywintypes27.dll
01/06/15 08:32 PM 54,979 README
02/20/15 08:31 AM <DIR> Scripts
02/20/15 08:32 AM <DIR> VMware OVF Tool
02/19/15 05:33 PM 9,585 windows_deploy.py
01/27/15 05:11 PM 48 xiv_sds_deployment_win.cmd
01/06/15 08:32 PM 30 xpyv.cmd
There are only two files that must be manipulated with during the deployment process: the deployment executable file (xiv_sds_deployment_win.cmd) and the XML template file, which must be customized for the target environment. A template of the deployment configuration XML is in the deploy folder and can be edited with a text editor.
Example 3-4 shows the contents of the SDS_Deploy_Win/deploy folder. The deployment configuration XML is highlighted in bold for clarity.
Example 3-4 Locating the .xml template
c:SpectrumAccelerateSDSSDS_Deploy_Windeploy>dir
 
Directory of c:SpectrumAccelerateSDSSDS_Deploy_Windeploy
 
02/19/15 05:29 PM 11,535 attach_disks_on_esx.py
02/20/15 10:19 AM 98,432 deploy_sds.log
02/19/15 05:29 PM 82,670 deploy_sds.py
02/20/15 08:50 AM 67,369 deploy_sds.pyc
02/19/15 05:33 PM 123 deploy_ver
02/19/15 05:29 PM 11,991 esxi_verifications.py
02/20/15 09:10 AM 10,597 esxi_verifications.pyc
02/20/15 08:54 AM <DIR> new_outputs
02/20/15 09:10 AM 0 rng_errors.txt
02/19/15 05:29 PM 11,001 Sample_RDM.ovf
02/19/15 05:29 PM 1,701 sds_machine_template.xml
02/19/15 05:29 PM 5,539 sds_validations.py
02/20/15 08:50 AM 9,069 sds_validations.pyc
02/19/15 05:29 PM 1,130 sds_xml.py
02/20/15 08:50 AM 1,254 sds_xml.pyc
02/19/15 05:29 PM 7,203 validate_sds_xml.py
02/20/15 08:50 AM 5,424 validate_sds_xml.pyc
02/19/15 05:29 PM 4,234,240 xiv_local_storage.vmdk
02/19/15 05:29 PM 1,163,526,144 xiv_sw_image.vmdk
 
3.3.2 Customizing the IBM Spectrum Accelerate configuration XML file
To customize the configuration file, open the XML template file for editing in a text editor. Some text editors do not parse the formatting correctly and display the file with all options in a single line. If this issue is the case, use a different text editor to help simplify the process.
Customize the XML file to contain a configuration for the environment where the IBM Spectrum Accelerate is deployed.
 
Important: Ensure that the editing process does not change the file type. The deployment kit must be a .xml file for the deployment to be successful.
Example 3-5 on page 47 shows a customized XML file that can be used to deploy an IBM Spectrum Accelerate system.
This example does not include a vCenter, which is why the <vcenter_host> stanza is missing. The example includes information for four ESXi servers and deploys the IBM Spectrum Accelerate system as a four-module system.
Example 3-5 Example of a customized XML file
<?XIVSDS xml_version="1.4.0"?>
<sds_machine
name="ITSO_SA_1"
interconnect_mtu="9000"
vm_gateway="10.0.20.1"
vm_netmask="255.255.255.0"
data_disks="6"
memory_gb="24"
icn="1234567"
ssd_disks="0"
off_premise="no"
num_cores="4"
enable_diagnostic_mode="yes"
 
<esx_servers>
<server
hostname="10.0.20.134"
username="root"
password="password"
Data store="ds1"
mgmt_network="Management Network"
iscsi_network="xiv_sds2_mgmt_isci"
vm_mgmt_ip_address="10.0.20.114"
interconnect_ip_address="14.170.0.114"
interconnect_ip_netmask="255.255.255.0"
interconnect_network="xiv_sds2_interconnect"
/>
<server
hostname="10.0.20.135"
username="root"
password="password"
Data store="ds2"
mgmt_network="Management Network"
iscsi_network="xiv_sds2_mgmt_isci"
vm_mgmt_ip_address="10.0.20.115"
interconnect_ip_address="14.170.0.115"
interconnect_ip_netmask="255.255.255.0"
interconnect_network="xiv_sds2_interconnect"
/>
 
<server
hostname="10.0.20.136"
username="root"
password="password"
Data store="ds3"
mgmt_network="Management Network"
iscsi_network="xiv_sds2_mgmt_isci"
vm_mgmt_ip_address="10.0.20.116"
interconnect_ip_address="14.170.0.116"
interconnect_ip_netmask="255.255.255.0"
interconnect_network="xiv_sds2_interconnect"
/>
 
<server
hostname="10.0.20.137"
username="root"
password="password"
Data store="ds4"
mgmt_network="Management Network"
iscsi_network="xiv_sds2_mgmt_isci"
vm_mgmt_ip_address="10.0.20.117"
interconnect_ip_address="14.170.0.117"
interconnect_ip_netmask="255.255.255.0"
interconnect_network="xiv_sds2_interconnect"
/>
 
</esx_servers>
</sds_machine>
3.3.3 Deploying IBM Spectrum Accelerate from the Windows command line
After the deployment kit is extracted and the deployment configuration XML is modified, the deployment process can be started. Run the xiv_sds_deployment_win.cmd -h command to see the possible options for deployment.
Example 3-6 shows the available options when the -h argument is provided. The -c or --config=FILE argument is the only option that is required to start the deployment as it provides the path of the deployment configuration XML file. The -w or --web-ui and --report-ui-progress options were introduced with IBM Spectrum Accelerate V11.5.4, and they open a Deployment Kit Web UI, as described in 4.4, “Deployment by using Deployment Kit Web UI” on page 69.
Example 3-6 Deployment command options
Y:xiv_sds_deployment_kit-11.5.4-internal-p20161108_133518-git-355a5b4885beSDS_Deploy_Win>xiv_sds_deployment_win.cmd -h
Usage: windows_deploy.py [options]
 
Options:
-h, --help show this help message and exit
-c FILE, --config=FILE
Deploy based on the specified XML configuration file
(full path).
-f, --force Allows the deployment script to delete existing VMs
that have the same name.
-T, --esxi-verifications-only
Run ESXi verifications only.
-t, --allow-configuring-esxi
Allow changing ESXi settings.
-a, --add-module Add one or more virtual XIV modules (storage nodes).
Can be used only with -c|--config.
-V, --verbose Run in verbose mode. Can be used only with
-c|--config.
-v, --version Display the version number of the embedded XIV
microcode.
-n, --no-startup Deploy, but do not turn the VMs.
-b, --batch Assume 'N' on all user prompts.
-d, --diagnostic-only
Run diagnostic mode and then clean up.
-i, --get-ilmt-tag Returns ILMT tag and exits.
-w, --web-ui Runs deployment through a web UI.
--override=VARVALUE INTERNAL: override a variable from deploy_sds.py
--import-over=FILENAME
INTERNAL: import a python file to override
deploy_sds.py variables/functions (caution!)
--diagnostic-defaults-file=XML_FILE
INTERNAL: inject defaults to diagnostic kit from
provided XML file
--allow-lsa-on-hdd INTERNAL: Allows installing LSA on HDD as data drive
(rather than SSD)
--report-ui-progress INTERNAL: Reports progress when deploying with UI
When an IBM Spectrum Accelerate system is deployed, the -V or --verbose argument helps to provide more information about the progress and status of the deployment. Example 3-7 shows an example of running the deployment command in verbose mode.
Example 3-7 Deploying an IBM Spectrum Accelerate system with verbosity turned on
c:SpectrumAccelerateSDSSDS_Deploy_Win>xiv_sds_deployment_win.cmd -V -c deploy_itso_sds1_1.xml
Example 3-8 includes an excerpt of verbose output during deployment. The output shows the creation of the RDM disk devices and the installation of the IBM Spectrum Accelerate system software to the module.
Example 3-8 Sample output of the script for a single module
2015-02-20 10:08:34: Creating direct attach disks on ESXi server 159.xx.y09.30: 6 data disks and 0 SSD disks
2015-02-20 10:08:44: [159.xx.y09.30] DIRECT ATTACH SCRIPT OUTPUT :Device mpx.vmhba32:C0:T0:L0 is of type CD-ROM - skipped
2015-02-20 10:08:44: [159.xx.y09.30] DIRECT ATTACH SCRIPT OUTPUT :Adding Device /vmfs/devices/disks/naa.5000c5005686599b as DATA disk #1
2015-02-20 10:08:44: [159.xx.y09.30] DIRECT ATTACH SCRIPT OUTPUT :Adding Device /vmfs/devices/disks/naa.5000c50056855703 as DATA disk #2
2015-02-20 10:08:44: [159.xx.y09.30] DIRECT ATTACH SCRIPT OUTPUT :Adding Device /vmfs/devices/disks/naa.5000c5005686073f as DATA disk #3
2015-02-20 10:08:44: [159.xx.y09.30] DIRECT ATTACH SCRIPT OUTPUT :Adding Device /vmfs/devices/disks/naa.5000c5005684365f as DATA disk #4
2015-02-20 10:08:44: [159.xx.y09.30] DIRECT ATTACH SCRIPT OUTPUT :Adding Device /vmfs/devices/disks/naa.5000c50056853ba3 as DATA disk #5
2015-02-20 10:08:44: [159.xx.y09.30] DIRECT ATTACH SCRIPT OUTPUT :Adding Device /vmfs/devices/disks/naa.5000c50056855477 as DATA disk #6
2015-02-20 10:08:44: Going to deploy VM itso_sds1_module_2
Opening OVF source: new_outputs/vmdk/tmp_ovf.ovf
Opening VI target: vi://[email protected]:443/
Deploying to VI: vi://[email protected]:443/
Transfer Completed
The deployment of the IBM Spectrum Accelerate system is completed when the installation process finishes for all of the modules in the system. Diagnostics are run as the last step of the deployment process if the configuration XML contained the option to do so.
Example 3-9 shows what completion of the deployment process looks like on a four-module IBM Spectrum Accelerate system with six HDDs and no SSDs.
Example 3-9 Sample of successful deployment
2015-02-20 10:18:13: Deploy of All modules completed successfully!
2015-02-20 10:18:13: Adding 6 data disks and 0 ssd disks to VM itso_sds1_module_1
2015-02-20 10:18:29: Adding 6 data disks and 0 ssd disks to VM itso_sds1_module_2
2015-02-20 10:18:46: Adding 6 data disks and 0 ssd disks to VM itso_sds1_module_3
2015-02-20 10:19:02: Adding 6 data disks and 0 ssd disks to VM itso_sds1_module_4
2015-02-20 10:19:19: Adding disks to All VMs completed successfully
2015-02-20 10:19:19: Turning on VM: 'itso_sds1_module_1' on ESXi host: '159.xx.y09.56'
2015-02-20 10:19:20: Turning on VM: 'itso_sds1_module_2' on ESXi host: '159.xx.y09.30'
2015-02-20 10:19:21: Turning on VM: 'itso_sds1_module_3' on ESXi host: '159.xx.y09.28'
2015-02-20 10:19:22: Turning on VM: 'itso_sds1_module_4' on ESXi host: '159.xx.y09.29'
2015-02-20 10:19:39: Module 1 started successfully
2015-02-20 10:19:39: Module 2 started successfully
2015-02-20 10:19:40: Module 3 started successfully
2015-02-20 10:19:40: Module 4 started successfully
2015-02-20 10:19:40: System (non-unique) serial number: 9021379
2015-02-20 10:19:40: System's machine unique ID (UUID): b750cae8fd6b419a92a597678902df33
2015-02-20 10:19:40: All Done
3.4 Linux command-line deployment procedure
The Linux deployment kit is delivered as a bash script with a self-contained archive. The deployment kit must be copied to the Linux system that is used for deployment. Place the deployment kit in a directory where the user has adequate permissions.
This section reviews the required steps to deploy an IBM Spectrum Accelerate system from a Linux server.
3.4.1 Obtaining a deployment XML configuration file from the deployment kit
The Linux deployment kit contains a deployment configuration XML template. Run the script with the -e argument to extract the template XML from the deployment kit. Example 3-10 shows the extraction of the configuration XML template from the deployment kit.
Example 3-10 Extracting the deployment configuration XML template from the deployment kit
itsouser: ~$ ./xiv_sds_deployment_kit-11.5.1.c-git-8f8231f5a544.bash -e
Searching embedded archive for example files... Please wait.
SDS_Deploy/deploy/sds_machine_template.xml
Examples available in ./XIV_Deployment_Examples/
itsouser: ~$ ls ~/XIV_Deployment_Examples
sds_machine_template.xml
 
Important: A deployment configuration XML file that is obtained as an exported XML file from the IBM XIV Management GUI can be used to simplify the configuration process.
3.4.2 Customizing the IBM Spectrum Accelerate configuration XML file
The deployment configuration XML template must be edited to reflect the environment where the IBM Spectrum Accelerate system is installed. Open the deployment configuration XML template file in a simple text editor. Customize the XML file to reflect the correct environment and wanted system settings.
Example 3-5 on page 47 shows a customized XML file that can be used to deploy an IBM Spectrum Accelerate system. This example does not include SSD drives or the use of a vCenter. It includes information for four ESXi servers.
After the deployment configuration XML file is customized, create a backup to prevent accidental deletion later.
3.4.3 Deploying IBM Spectrum Accelerate from the Linux command line
After the XML file is edited, it must be placed in a directory where the user has permissions. Place the configuration XML in the same directory as the deployment kit. This placement simplifies the commands to start the deployment process.
The Linux deployment kit is delivered as a bash script that contains the deployment archives. The deployment kit can be run directly with arguments that alter the behavior of the installation process. Use the -h or --help argument to see a list of the available arguments and their usage. Example 3-11 shows the available options when -h or --help is used. The -w or --web-ui options were introduced with IBM Spectrum Accelerate V11.5.4, and they open a Deployment Web UI, as described in 4.4, “Deployment by using Deployment Kit Web UI” on page 69.
Example 3-11 Deployment script usage
[root@SCB5 Deploy_Kits]# ./xiv_sds_deployment_kit-11.5.4-internal-p20161108_133518-git-355a5b4885be.bash -h
Usage: ./xiv_sds_deployment_kit-11.5.4-internal-p20161108_133518-git-355a5b4885be.bash [options]
-c|--config XML_PATH : Deploy based on the specified XML configuration file (full path).
-n|--no-startup : Deploy on the ESXi hosts without turning on the virtual machines.
-f|--force : Allows the deployment script to delete existing VMs that have the same name.
-a|--add-module(s) : Add one or more virtual XIV modules (storage nodes). Can be used only with -c|--config.
-e|--examples : Extract the XML template file.
-V|--verbose : Run in verbose mode. Can be used only with -c|--config.
-h|--help : Display this help text.
-v|--version : Display the version number of the embedded XIV microcode.
-d|--diagnostic-only : Run system diagnostics and then clean up.
-i|--get-ilmt-tag : Returns ILMT tag and exits.
-T|--esxi-verifications-only : Run ESXi verifications only
-t|--allow-configuring-esxi : Allow changing ESXi settings.
-w|--web-ui : Start the web-based UI.
The deployment process is ready to begin when the deployment configuration XML is customized and the physical environment is validated.
Begin installing the IBM Spectrum Accelerate system by running the xiv_sds_deployment_kit-<version>.bash -V -c <CONFIG_XML> command.
Example 3-12 on page 52 shows running the deployment process on a Linux host. This example also shows the installation process of a four module system with 11 disk drives SSD caching support.
In Example 3-12, the -f (--force) option is used to replace virtual machines with the same name. The -V (--verbose) option is used to see more output during the deployment process. The -c (--config) option is used to specify the deployment configuration XML file. The output is broken up into different sections to highlight what you might see during the deployment process.
 
Example 3-12 Deployment from Linux host
[root@SCB5 Deploy_Kits]# ./xiv_sds_deployment_kit-11.5.4-internal-p20161108_133518-git-355a5b4885be.bash -V -f -c team2_xiv_sds2.xml
Flags:
-c team2_xiv_sds2.xml
--force
no
no
Basic checks pass
Un-tarring the package
..................
 
Note: Take care when the -f option is used because it deletes any virtual machines that feature the same name as the virtual machines that are defined in the deployment configuration XML file.
The deployment process proceeds in a similar manner as a command line deployment from a Windows system. If diagnostics were enabled within the deployment configuration XML file, the IBM Spectrum Accelerate system begins the diagnostic process after the virtual machines are powered on.
If the deployment is successful, the message “All Done” is emitted along with the systems 32 character identification number.
Example 3-13 shows the completion of a four module deployment. The non-unique serial number and unique system identifier can be seen.
Example 3-13 Completion of the deployment process
2016-12-03T15:57:39+1:00 [2488]: Deploy of All modules completed successfully!
2016-12-03T15:57:39+1:00 [2766]: Adding 10 data disks and 0 ssd disks to VM XIV_SDS_ATS2_module_1...
2016-12-03T15:57:39+1:00 [2765]: Adding 10 data disks and 0 ssd disks to VM XIV_SDS_ATS2_module_2...
2016-12-03T15:57:39+1:00 [2767]: Adding 10 data disks and 0 ssd disks to VM XIV_SDS_ATS2_module_3...
2016-12-03T15:57:43+1:00 [2488]: Adding disks to all VMs completed successfully.
2016-12-03T15:57:48+1:00 [2779]: Turning on VM: 'XIV_SDS_ATS2_module_1' on ESXi host: '10.0.20.134'...
2016-12-03T15:57:48+1:00 [2780]: Turning on VM: 'XIV_SDS_ATS2_module_2' on ESXi host: '10.0.20.135'...
2016-12-03T15:57:48+1:00 [2781]: Turning on VM: 'XIV_SDS_ATS2_module_3' on ESXi host: '10.0.20.136'...
2016-12-03T15:57:55+1:00 [2779]: Module 1 started successfully.
2016-12-03T15:57:55+1:00 [2780]: Module 2 started successfully.
2016-12-03T15:57:55+1:00 [2781]: Module 3 started successfully.
2016-12-03T15:57:55+1:00 [2488]: Waiting for system to start work
2016-12-03T15:58:43+1:00 [2488]: System state: 'on'
2016-12-03T15:58:43+1:00 [2488]: System (non-unique) serial number: 9065133
2016-12-03T15:58:43+1:00 [2488]: System's machine unique ID (UUID): 8b7de191b5f442d2b932d2f67ab591b8
2016-12-03T15:58:43+1:00 [2488]: System machine name: XIV_SDS_ATS2
2016-12-03T15:58:43+1:00 [2488]: All Done
Completed.
 
Important: Each new deployment generates a unique 32-digit system identifier and a non-unique 7-digit system serial number.
If a deployment is done to replace an installation, the newly deployed system features a new unique identification number and non-unique serial number that is different from the original system.
3.5 Troubleshooting command-line deployment failure
There are several issues that you must be aware of that can cause the deployment to fail.
..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
52.15.176.80