This section describes how to install the single system OpenStack evaluation configuration. The first three methods described install on bare metal. The fourth method installs into a kernel zone.
How to Install Using the Unified Archive File and an AI Install Service
How to Install Using AI Bootable Media Created from the Unified Archive File
How to Install into a Kernel Zone Using the Havana Unified Archive File
This procedure describes how to install Oracle Solaris and OpenStack on bare metal by using the downloaded USB file. This method does not require an AI server and is the most direct method for bare metal installations.
If you have access to an Oracle Solaris 11.2 system, use the usbcopy utility. See the usbcopy(1M) man page.
You must use the usbcopy command from Oracle Solaris 11.2. You cannot use usbcopy from an earlier version of Solaris.
If you do not have access to an Oracle Solaris 11.2 system, you can use the dd command.
Be extremely careful that you correctly identify the proper disk (flash drive) when using dd.
# svcadm disable -t hal
# rmformat
# dd if=/path/image.usb of=/dev/rdsk/device bs=16k
# svcadm enable hal
# dmesg | tail
# dd if=/path/image.usb of=/dev/diskN bs=16k
# diskutil list # diskutil unmountDisk /dev/diskN
# dd if=/path/image.usb of=/dev/diskN bs=16k
The System Configuration Interactive (SCI) Tool is displayed. If you do not see the SCI Tool, press the Enter key or press Ctrl-L to redraw the screen.
This procedure describes how to install Oracle Solaris and OpenStack on bare metal by using the downloaded Unified Archive file and AI.
On your Oracle Solaris AI install server, copy and modify the file /usr/share/auto_install/manifest/default_archive.xml according to your AI setup. In the ARCHIVE software section, specify the location of the downloaded .uar file.
Set up an AI install service using the AI manifest from the previous step. See Part III, Installing Using an Install Server, in Installing Oracle Solaris 11.2 Systems .
ok boot net -install
The SCI Tool is displayed to configure the system. If you do not see the SCI Tool, press the Enter key or press Ctrl-L to redraw the screen.
This procedure describes how to install Oracle Solaris 11.2 and OpenStack on bare metal by creating a bootable AI image. A bootable USB image is created from the downloaded Unified Archive file. See Chapter 5, Automated Installations That Boot From Media, in Installing Oracle Solaris 11.2 Systems for more information about this method.
# archiveadm create-media -s http://pkg.oracle.com/solaris/release \ -f usb -o /workdir/usb-filename \ /workdir/uar-file
where workdir is the location where you downloaded the Unified Archive file. The AI USB file will be created in the same directory.
If you downloaded the USB file to an Oracle Solaris 11.2 system, use the usbcopy command.
If you downloaded the USB file to a system whose Oracle Solaris OS is previous to Oracle Solaris 11.2, use the dd command as follows:
Disable the HAL service.
# svcadm disable -t hal
Insert the flash drive and locate the appropriate device.
# rmformat
Copy the image to the flash drive.
# dd if=/path/image.usb of=/dev/rdsk/device bs=16k
Enable the HAL service.
# svcadm enable hal
You can use the default manifest or create a custom manifest. If you create a custom manifest, store the custom manifest in a location that is reachable by the system being installed.
You are prompted to use the default AI manifest or provide the location of the custom manifest.
The SCI Tool is displayed to configure the system. If you do not see the SCI Tool, press the Enter key or press Ctrl-L to redraw the screen.
This procedure describes how to install Oracle Solaris 11.2 and Havana OpenStack directly into a kernel zone by using the downloaded Unified Archive file.
Before You Begin
Make sure the system that will host the kernel zone satisfies the virtualization requirements specified in OpenStack Installation Requirements.
# zonecfg -z OpenStackKZ create -t SYSsolaris-kz
Ensure your kernel zone has enough virtual CPUs, RAM, storage, and MAC addresses. Non-global zones created inside the kernel zone will be able to consume these extra MAC addresses automatically.
The following example configures the zone with 8 virtual CPUs, a limit of 8 GB of physical memory, and automatic MAC address assignment. See the zonecfg(1M) man page for descriptions of configurable resources.
# zonecfg -z OpenStackKZ zonecfg:OpenStackKZ> add virtual-cpu zonecfg:OpenStackKZ:virtual-cpu> set ncpus=8 zonecfg:OpenStackKZ:virtual-cpu> end zonecfg:OpenStackKZ> select capped-memory zonecfg:OpenStackKZ:capped-memory> set physical=8g zonecfg:OpenStackKZ:capped-memory> end zonecfg:OpenStackKZ> select anet id=0 zonecfg:OpenStackKZ:anet> add mac zonecfg:OpenStackKZ:anet:mac> set mac-address=auto zonecfg:OpenStackKZ:anet:mac> end zonecfg:OpenStackKZ:anet> end zonecfg:OpenStackKZ> exit
# zonecfg -z OpenStackKZ info
The following example shows installing the x86 version of the .uar file.
# zoneadm -z OpenStackKZ install -a /path/uar-file
where uar-file is the Oracle Solaris 11.2 with Havana OpenStack archive.
# zoneadm -z OpenStackKZ boot
# zlogin -C OpenStackKZ
The SCI Tool is displayed to configure the system. If you do not see the SCI Tool, press the Enter key or press Ctrl-L to redraw the screen.
The Unified Archive expects a DHCP server to assign an IP address to the kernel zone. If you are using DHCP, make sure a MAC address is assigned to the kernel zone. Selecting Automatic on the network page of the SCI Tool in the previous step should result in a MAC address being assigned. If you are not using DHCP, make sure to assign an IPv4 address to the kernel zone.
The IP address is important because no system repositories currently exist for kernel zones. With an IP address, the kernel zone can connect to an IPS package repository if required.