XenServer Installation Guide

Release 5.0.0

Table of Contents

1. Introducing XenServer
1.1. About this document
1.2. How this Guide relates to other documentation
1.3. New Features in XenServer 5.0.0
1.3.1. New Guest Support
1.3.2. Business Continuity
1.3.3. Storage and Provisioning
1.3.4. Usability and Reliability
1.3.5. Performance and Hardware
2. System Requirements
2.1. XenServer host system requirements
2.2. XenCenter requirements
2.3. VM support
3. XenServer Licensing
4. Installing XenServer
4.1. Installing the XenServer host
4.2. Installing XenCenter
4.3. Installation and deployment scenarios
4.3.1. XenServer hosts with local storage
4.3.2. XenServer hosts with shared NFS storage
4.3.3. XenServer hosts with shared iSCSI storage
5. Updating, upgrading, or freshening XenServer
5.1. Preparing XenServer hosts for upgrade
5.2. Applying updates
5.2.1. Applying updates using the CLI
5.3. Reinstalling the same version
5.4. Rolling upgrade between major versions
5.4.1. Backup
5.4.2. Performing a rolling pool upgrade
5.5. Upgrading from version 3.2 to 4.0.1
5.6. Upgrading VHD files from version 4.0.1
5.7. Backing up and restoring XenServer hosts and VMs
5.7.1. Backing up Virtual Machine metadata
5.7.2. Backing up XenServer hosts
5.7.3. Backing up VMs
A. Troubleshooting
B. PXE installation of XenServer host
B.1. Setting up the PXE boot environment
B.2. Creating an answerfile for unattended PXE installation
B.3. Installation media repository format
B.3.1. Presence of installation media repositories
B.3.2. Installation media repository metadata
B.3.3. Package metadata
B.3.4. Example files
B.3.5. Notes on best practice
C. Xen Memory Usage
C.1. Setting Control Domain Memory
Index

Table of Contents

1.1. About this document
1.2. How this Guide relates to other documentation
1.3. New Features in XenServer 5.0.0
1.3.1. New Guest Support
1.3.2. Business Continuity
1.3.3. Storage and Provisioning
1.3.4. Usability and Reliability
1.3.5. Performance and Hardware

Thank you for choosing XenServer™ from Citrix Systems, Inc., the creators of the Xen® hypervisor and leaders of the open source Xen project.

This document is an installation guide for XenServer, the platform virtualization solution from Citrix. The XenServer package contains all you need for creating a network of virtual x86 computers running on Xen®, the open-source paravirtualizing hypervisor with near-native performance.

This document contains procedures to guide you through the installation, configuration, and initial operation of XenServer. This document also contains information about troubleshooting problems that might occur during installation, and where to go for further information.

This document is primarily aimed at system administrators who need to set up XenServer hosts on physical servers. Other documentation shipped with this release includes:

This release of XenServer contains the following major new features. For a detailed list, please refer to the XenServer 5.0.0 Release Notes. XenServer 5.0.0 includes a number of new features and ongoing improvements in hardware support, performance and scalability.

The following new guest operating systems are supported for installation:

Linux guests have also been refreshed to their latest upstream versions, including the bundled XenServer kernels which fix stability issues not addressed in the upstream releases.

XenServer supports several features to guarantee service uptime in the event of infrastructure failure. Firstly, resource pools can be configured for automated high-availability. This deals with individual host failures by restarting VMs that were running on that host onto the next available machine in the resource pool. Notable features include:

Integrated disaster recovery to enable regular backups of virtual machine metadata. When combined with SAN storage replication, this provides a fast way to migrate entire resource pools to another physical site and continue running services with little interruption. It also permits the use of storage repositories which include all of the metadata for VMs installed on them, permitting a "transportable VM" model across resource pools.

Fibre Channel and iSCSI multipath support, configurable from the XenCenter GUI. This ensures that redundant storage links to an FC SAN or ISCSI SR permit link failures without loss of service.

Improved network reliability by support active/active NIC aggregation. Existing active/passive NIC bonds on older installations will be upgraded to aggregation mode for active/active usage, permitting full use of all available bandwidth while still maintaining redundant links.

Multiple management network interfaces can be defined in the control domain, and individual ones can be dedicated for use by network storage (e.g. iSCSI or NFS). This helps improve isolation between VMs and storage infrastructure traffic.

One of the most common reasons for system failure is operator error through misconfiguration. XenServer now provides e-mail and XenCenter alerting for potentially dangerous configurations, for example when VM performance will be degraded or a resource pool is overcommitted with respect to high availability.

Support has been added for QLogic and Emulex 8GB Fibre Channel HBA devices, as well as bug-fixes and stability improvements for other HBA devices.

Diskless virtual machines running on Citrix Provisioning Server now support more advanced PXE configurations, such as DHCP Proxy ARP responses or wide area DHCP relays. Support for PXE servers such as Altiris and Windows Deployment Services (WDS) has been improved.

There is CLI-only support for quiesced, fully consistent VM storage snapshots for Windows VMs. The storage backend must support fast-cloning and snapshot, e.g. the NetApp and Equalogic storage repositories.

The VHD-based storage repositories have undergone significant tuning, improving performance and reducing resource utilization for the local VHD and NFS storage repositories.

This release of XenServer 5.0.0 provides a preview of storage repository support for Dell EqualLogic arrays. This storage repository type supports fast-cloning, snapshots and high-performance LUN-per-VDI access by directly controlling the storage array.

Performance statistics are now persisted in a round-robin database on the XenServer pool, with decreasing granularity as time passes. This enables long-term trending and resource optimization for complex data centre deployments.

The XenCenter GUI supports connecting to both 4.1 and 5.0.0 hosts, and has several new features:

Enhanced support for keeping hosts up to date with the latest improvements:

A menu-driven text console is now present when connecting to the main screen of the XenServer host. This new interface provides an alternative to XenCenter for common operations, or when networking is not otherwise available. The boot process is also much more user-friendly now, with a graphical splash screen and progress indicator.

The core Xen hypervisor has been upgraded to a version based off the stable 3.2 releases. Major improvements are present in:

Many hardware device drivers have been refreshed to their latest stable versions, including MegaRAID SAS, MPT Fusion, Broadcom and 3Ware RAID cards. For 10Gb networking adapters, support has been added for Neterion and Broadcom cards, and Chelsio performance has been significantly improved.

Table of Contents

2.1. XenServer host system requirements
2.2. XenCenter requirements
2.3. VM support

XenServer requires at least two separate physical x86 computers: one to be the XenServer host, and the other to run the XenCenter application. The XenServer host machine is dedicated entirely to the task of hosting VMs and is not used for other applications. The computer that runs XenCenter can be any general-purpose Windows computer that satisfies the hardware requirements, and can be used to run other applications simultaneously.

The XenServer host is a 64-bit x86 server-class machine devoted to hosting multiple VMs. This machine runs a stripped-down Linux operating system with a Xen-enabled kernel which controls the interaction between the virtualized devices seen by VMs and the physical hardware.

The following are the system requirements for the XenServer host:

CPUsOne or more 64-bit x86 CPU(s), 1.5 GHz minimum, 2 GHz or faster multicore CPU recommended

To support VMs running Windows, an Intel VT or AMD-V 64-bit x86-based system with one or more (up to 32) CPUs is required.

Note

To run Windows VMs, hardware support for virtualization must be enabled on the XenServer host. This is an option in the BIOS. It is possible your BIOS might have virtualization support disabled. Consult your BIOS documentation for more details.

To support VMs running supported paravirtualized Linux, a standard 64-bit x86-based system with one or more (up to 32) CPUs is required.

RAM 1 GB minimum, 2 GB or more recommended
Disk spaceLocally attached storage (PATA, SATA, SCSI) with 16 GB of disk space minimum, 60 GB of disk space recommended

General disk space requirements for VMs:

  • Product installation creates two 4GB partitions for the XenServer host control domain; remaining space is available for VMs
  • VMs based on the Debian templates are allocated a 4GB root device, and a 512MB swap device
  • Linux VMs are allocated a root device of 8 GB
  • Windows Vista VMs are allocated a root device of 16 GB, and other Windows VMs default to 8 GB.
Network100 Mbit/s or faster network interface card (NIC). A gigabit NIC is recommended for faster P2V and export/import data transfers and for live relocation of VMs.

The remote XenCenter application for managing the XenServer host can be installed and run on any Windows 2003, Windows XP, Windows Vista workstation or laptop.

The following are the system requirements for XenCenter:

Operating systemWindows XP, Windows Server 2003, or Windows Vista
.NET frameworkversion 2.0 or above
CPU Speed750 MHz minimum, 1 GHz or faster recommended
RAM1 GB minimum, 2 GB or more recommended
Disk space100 MB minimum
Network interface card100Mb or faster NIC

Windows VMs can be created only on XenServer hosts equipped with Intel VT-enabled or AMD-V CPUs. All Windows VMs are created via installing the operating system from either the Microsoft installation media in the XenServer host physical CD/DVD-ROM drive or a network-accessible ISO image to the appropriate template.

Linux VMs do not require XenServer hosts that are equipped with Intel VT-enabled or AMD-V CPUs.

For a list of supported Windows and Linux distributions, refer to the XenServer Virtual Machine Installation Guide.

When XenServer 5.0.0 is first installed, the product is enabled with the XenServer Express Edition feature set by default. This feature set enables basic XenServer capabilities, such as single server management through XenCenter and support for up to four concurrently running virtual machines. However, to enable the more advanced capabilities found in XenServer Standard, Enterprise, and Platinum Editions, a corresponding license key-in the form of a license file-must be installed on each XenServer host system.

The XenServer license key is provided in the form of a license file with the .xslic extension. This unique file will be issued to your organization by Citrix and can be installed on a XenServer host system in a number of different ways including:

To install a license file using XenCenter

  1. Select the server in the Resources pane.
  2. On the Server menu, click Install License Key.
  3. Locate the license file and click Open. By default, only license files with the .xslic extension will be displayed.

To install a license file using the xe CLI

  • On the server console, enter the command
    host-license-add license-file=<path/license_filename>

Important

For XenServer Enterprise Edition and higher, each host system in a resource pool must be individually licensed. For example, if you are supporting four XenServer hosts in a resource pool, you must apply four unique license files to each of the four host systems.

Q: The XenServer (annual, not-for-resale, trial) license has expired. What is going to happen?
Q: The license file from a previous version of XenServer does not work with XenServer 5.0.0
Q: This XenServer 4.0.1 license file has a .txt extension, but the product licensing instructions reference a license file with the .xslic extension. Does this mean that the XenServer 4.0 license is incompatible?
Q: Is there a way to manually install a XenServer license file without using XenCenter?

Q:

The XenServer (annual, not-for-resale, trial) license has expired. What is going to happen?

A:

Upon license expiration, a XenServer host reverts back to the XenServer Express Edition feature set. If the license on a XenServer host expires while the system is still running, all active virtual machines continue to run as long as the host system is not disrupted. However, new VMs cannot be launched. If the host system is disrupted (through a power failure, system restart, and so on), XenServer reverts to Express Edition-level functionality upon restart.

Note

Citrix strongly encourages customers who opt for the XenServer annual license to renew their new annual license before the expiration date to ensure the greatest degree of continuity. XenCenter and e-mail alerts will be generated daily from 30 days before the license is due to expire, to give you enough notice to upgrade.

Q:

The license file from a previous version of XenServer does not work with XenServer 5.0.0

A:

All license files from XenServer 3.x are incompatible with XenServer 5.0.0. XenServer 3.x customers under valid software maintenance or Subscription Advantage agreements will receive a valid XenServer 4.x license file from Citrix and these newly issued license files should be used in conjunction with XenServer 5.0.0.

XenServer 4.0.1 and 4.1.0 license files are forward-compatible with XenServer 5.0.0.

Q:

This XenServer 4.0.1 license file has a .txt extension, but the product licensing instructions reference a license file with the .xslic extension. Does this mean that the XenServer 4.0 license is incompatible?

A:

No. In general, XenServer 4.0 license files are forward-compatible with XenServer 5.0.0 and as a result XenCenter can import valid XenServer 4.x license files of any extension. For some administrators, it may be easier to rename an older XenServer 4.0 license key with a .txt extension to a file with the .xslic extension prior to applying the license file in XenCenter.

Q:

Is there a way to manually install a XenServer license file without using XenCenter?

A:

Yes. First, license files may be manually installed using the xe CLI. The host-license-add command allows a local license file to be installed on a particular XenServer host. For more information about using xe, refer to the XenServer Administrator's Guide. Additionally, you can use Secure Copy (SCP) to upload a license file from a system where the license file resides to a XenServer host. The target path on the XenServer host system must be /etc/xensource/license. Citrix strongly recommends using SCP to apply a license file only as a last resort, such as when the XenCenter console or xe CLI are unavailable.

Table of Contents

4.1. Installing the XenServer host
4.2. Installing XenCenter
4.3. Installation and deployment scenarios
4.3.1. XenServer hosts with local storage
4.3.2. XenServer hosts with shared NFS storage
4.3.3. XenServer hosts with shared iSCSI storage

Any XenServer network, from the simplest to the most complex deployment, is made up of one or more XenServer hosts, each running some number of VMs, and one or more workstations running XenCenter to administer the XenServer hosts.

In order to create resource pools and enable XenMotion (live migration of VMs), a means of shared storage also needs to be deployed on the network. This version of the XenServer product family supports Fibre Channel, NetApp filers, LVM over iSCSI, and NFS shared storage.

This chapter describes installing XenServer host software on physical servers, installing XenCenter on Windows workstations, and connecting them to form the infrastructure for a network of Virtual Machines.

The first sections describe the installation of XenServer host and XenCenter, which are common to all deployments. The following sections describe several common installation and deployment scenarios and provide information that is specific to each scenario.

Installers for both the XenServer host and XenCenter are on the installation media. The installation media also includes:

The XenServer host consists of a Xen-enabled Linux operating system, a management agent, VM templates, and a local storage repository reserved for VMs. The XenServer host must be installed on a dedicated 64-bit x86 server. XenServer is not supported in a dual-boot configuration with any other operating system.

You can install the XenServer host from the installation CDs or set up a network-accessible TFTP server to boot from via PXE. For details about setting up a TFTP server for PXE-booting the installer, see Appendix B, PXE installation of XenServer host.

Note

Do not install any other operating system in a dual-boot configuration with the XenServer host; this is an unsupported configuration.

The main installation CD contains the basic packages to set up the XenServer host on a physical host, and to create Windows VMs by using the Windows installation CDs. The XenServer package also contains a separate CD containing support for creating Linux VMs (including complete built-in distributions of Debian Sarge and Etch), and six CDs containing source code for the included open-source software.

The installer has an upgrade choice if it is run on a server that already has a previous version of XenServer on it. It parallels the first-time installation process but several setup steps are bypassed, and the existing settings for networking configuration, system time setting, etc. will be retained.

If you only want to create Windows VMs, you can install XenServer using just the first CD. If you want to install Linux VMs, be sure to

  1. Download the Linux Pack ISO
  2. Burn it to a physical CD if installing from a DVD/CD drive, or set it up for PXE installation as described in Appendix B, PXE installation of XenServer host

Note

If, after installing XenServer without Linux support, you decide later to add it, mount the Linux Pack installation CD or ISO image on the XenServer host and run the script install.sh, located in the root of the CD.

To install or upgrade the XenServer host

Warning

If you are performing an upgrade, please ensure you do not have any suspended virtual machines as these may not be resumable after the upgrade. Please ensure that all CD drives of virtual machines have been ejected and are empty.

  1. Boot the computer from the main installation CD, or PXE-boot from your TFTP server if applicable (see Appendix B, PXE installation of XenServer host for details on how to set up the XenServer media for PXE installation).
  2. After the initial boot messages, the installer does some hardware detection and initialization, then presents a screen asking you to select which keyboard keymap you want to use for the installation. In this and the screens that follow, use Tab or Alt+Tab to move between elements, Space to select, and F12 to move to the next screen.Select the desired keymap and choose OK to proceed.
  3. Next, the "Welcome to XenServer" screen appears. Select Install or upgrade XenServer host and choose OK to proceed.
  4. The next screen displays a message telling you that the setup program will install XenServer on the computer, and warns that it will overwrite data on any hard drives that you select to use for the installation. Choose OK to proceed.
  5. The XenServer End User License Agreement (EULA) is displayed. Use the up and down arrow keys to scroll through and read the agreement. Choose Accept EULA to proceed.
  6. At this point, if the computer on which you are installing the XenServer host does not have a CPU which supports hardware virtualization, or if the support is disabled in the BIOS, a message appears to warn you that you will not be able to run Windows VMs. Choose OK to proceed.Note that some systems have bugs in their BIOS software which can result in the setting being incorrect. If you get a spurious warning about a lack of hardware virtualization (or do not see a warning when you expected one), then perform a hard power cycle of the host and restart the installation. You should also check the hardware manufacturer's support site for BIOS upgrades.
  7. If the installer detects a previously-installed version of XenServer host, you are offered the choice to perform a clean installation, or to upgrade the existing version, which preserves any of the VMs present. Choose an installation type as appropriate and select OK to proceed.If you select to upgrade an existing version, you will get a message that the installer is going to create a backup of the existing installation. Select Continue to proceed.
  8. If you have multiple local hard disks, you are asked to choose the Primary Disk for the installation. Select the desired disk and choose OK to proceed. After selecting the primary one, you are also prompted to choose if you want any of the other drives to be formatted for use by XenServer for VM storage. Select and choose OK to proceed.If the computer has a single hard disk, these two screens do not appear.
  9. The next screen asks you to specify the source of the installation packages. If you are installing off the CD, you will most likely want to select Local media (CD-ROM). If you are installing via PXE you will most likely want to select HTTP or FTP or NFS, as appropriate.If you select HTTP or FTP or NFS, you are next prompted to set up Networking so that the installation script can connect to the product repository.If the computer has multiple network interfaces, you are prompted to select one of them to be used to access the XenServer product repository. Select and choose OK to proceed.If the computer has a single network interface, that interface is used to access the XenServer product repository and no prompt is displayed.You can select Automatic configuration (DHCP) to configure the NIC using DHCP, or Static configuration, which prompts you to configure the NIC's properties manually. Following that, you are prompted to provide the URL or NFS server and path where the installation media are, as appropriate.

    Note

    To be part of a resource pool, XenServer hosts need to have static IP addresses.

    If you selected Local media, this networking setup appears later in the installation process.If you selected Local media, the next screen asks if you want to install the Linux Pack from a second CD. If you are planning to install VMs that will run Linux operating systems, choose Yes. If you are planning to install only Windows VMs, you can choose No.

    Important

    In a pooled setup, the Linux Pack must be installed either on all of the pool XenServer hosts, or on none of them, so that they are homogeneous.

  10. The next screen asks if you want to verify the integrity of the installation media. If you select Verify installation source, the MD5 checksum of the packages is calculated and checked against the known value. This may take some time. If you select Skip verification, this check is bypassed. Make your selection and choose OK to proceed.
  11. You are next prompted to set a root password. (This will be the password that the XenCenter application will use to connect to the XenServer host.) Type the desired password and type it again to verify it.
  12. If you selected to perform a clean installation, you are next prompted to set up Networking for the management NIC, which is the interface that will be used to connect to the XenCenter.If you selected to upgrade an existing installation, the existing management NIC configuration is used and these screens are bypassed.If the computer has multiple network interfaces, you are prompted to select one of them to be used as the management NIC for the XenServer host software. Select one and choose OK to proceed.If the computer has a single network interface, that interface is used as the management NIC and no prompt is displayed.Next you can select Automatic configuration (DHCP) to configure the NIC using DHCP, or Static configuration, which prompts you to configure the NIC's properties manually.

    Note

    To be part of a resource pool, XenServer hosts need to have static IP addresses.

  13. If you selected to perform a clean installation, you are next prompted to specify the hostname and the configuration for the name service.If you selected to upgrade an existing installation, the existing hostname and name service configuration is used and these screens are bypassed.In the Hostname Configuration section, if you select Automatically set via DHCP, the DHCP server will provide the hostname along with the IP address. If you select Manually specify, enter the desired hostname for the server in the field below.In the DNS Configuration section, if you select Manually specify, enter the IP addresses of your primary (required), secondary (optional), and tertiary (optional) Nameservers in the fields below. Otherwise, select Automatically set up via DHCP to get name service configuration via DHCP.Select OK to proceed.
  14. If you selected to perform a clean installation, you are prompted to select the general geographical area for the Time Zone. Choose from the displayed list of geographical areas, then choose OK to proceed. You are then prompted to select the specific locale for the Time Zone. (Note that this list is long, but if you type the first letter of the desired locale, the selection will jump to the first entry that begins with this letter.) Choose from the displayed list of locales, then choose OK to proceed.If you selected to upgrade an existing installation, the existing time zone and locale is used and these screens are bypassed.
  15. If you selected to perform a clean installation, you are prompted to choose a method of setting the System Time. You can select Using NTP or Manual time entry. Make your selection and choose OK to proceed.If you selected to upgrade an existing installation, the existing method of setting system time is used and this screens is bypassed.
  16. If you selected Using NTP in the preceding step, you are prompted to identify the time server or servers you want to use. You can check NTP is configured by my DHCP server and the time server will be set by DHCP. Otherwise, enter at least one NTP server name or IP address in the fields below. Choose OK to proceed.Otherwise, the installation script moves to the next step; you will be prompted for the manually-entered time later, near the end of the installation.

    Warning

    Currently XenServer assumes that the time setting for the server's BIOS is the current time in UTC, and that the time for the VMs reflects the local time based on the time zone offset specified.

  17. A message is displayed that the installation is ready to proceed and that this will format the primary disk and any other disks selected for VM storage, destroying any data that is currently on them. Select Install XenServer to proceed.A progress bar is displayed as the installation commences. If you chose to set the system date and time manually, a dialog box appears when the progress bar has reached about 90%. Enter the correct numbers in the fields and select OK to proceed.
  18. If you are installing from CD and selected to include support for Linux VMs, you will be prompted to put in the Linux Pack disk. Eject the main disk, put in the Linux Pack disk, and close the CD drawer. Select OK. A screen appears, identifying that this disk contains the Linux Pack. Select Use media to proceed with installing it. Another progress bar is displayed, and when it reaches 100%, a completion message is displayed.If you selected not to install support for Linux VMs, a completion message is displayed.

    Note

    If you decide later to add Linux support, mount the Linux Pack installation CD or ISO image on the XenServer host and run the script install.sh, located in the root of the CD.

  19. Select Reboot. After the splash screen, XenServer displays xsconsole, a convenient system configuration console.

XenCenter is a Windows client application. XenCenter must be installed on a remote machine that can connect to the XenServer host through the network; it cannot run on the same machine as the XenServer host. It can be installed and run on Windows 2003, XP SP2, or Vista. The .NET framework version 2.0 or above must be installed as well.

Note

The XenCenter installation wizard can be used to install previous versions of XenCenter, which can be used to manage previous versions of XenServer. Running multiple versions of XenCenter on a single machine is supported.

To install XenCenter

  1. Before installing XenCenter, be sure to uninstall the previous version if one exists.
  2. Put the Base Pack CD in the drive or browse to the location where you downloaded the installation file XenCenter.msi.
  3. If installing from CD:
    • If Auto-play is enabled for the CD drive, the application installer launches automatically after a few moments.
    • If Auto-play is not enabled for the CD drive, browse to the /client_install directory on the CD and find the file named XenCenter.msi. Then double-click on the file to launch the application installer.
    If installing from the installation file XenCenter.msi, double-click on the file to launch the application installer.
  4. Click next on the first page of the setup wizard. On the Custom Setup page, XenCenter 4.1.0 is shown as a subfeature of XenCenter 5.0.0. If you have any XenServer 4.0 hosts that you want to manage, select it by clicking and choosing Will be installed on local hard drive or Entire feature will be installed on local hard drive. In this case, a separate XenCenter 4.1.0 will also be installed on your computer.If you do not have any XenServer 4.0 hosts that you want to manage, leave this subfeature unselected.Click Next to proceed.
  5. The next page allows you to modify the default destination folder (C:\Program Files\Citrix\XenCenter). Click Browse to change the default installation location, if desired. You can also select whether XenCenter is installed so that every user of the computer can access it, or only the user logged into the current profile. Click Next to proceed.
  6. In the next page, click Install to proceed.When complete, there will be a XenCenter icon on the desktop and a XenCenter item on the All Programs list.

    Note

    The installer will only create a single desktop icon, for XenCenter 5.0.0. XenCenter 4.1.0 appears in the All Programs list on the Start menu

  7. When the installation process is complete, click Finish to close the Setup wizard. There will be a XenCenter icon on the desktop and a XenCenter item on the All Programs list.

Note

By default, XenCenter allows saving of usernames and passwords. To disable this, use the registry Editor, navigate to the key HKEY_CURRENT_USER\Software\Citrix\XenCenter and add a key named AllowCredentialSave with the string value false. This will cause XenCenter to no longer save usernames or passwords, and disables the Save and Restore Connection State dialog box in XenCenter (Tools -> Save and Restore).

Should you need to, XenCenter can be uninstalled from a system quite easily.

Note

If you installed XenCenter 4.1.0 along with XenCenter 5.0.0, the uninstallation process will remove both versions.

To uninstall XenCenter

  1. Select Control Panel from the Start menu.
  2. In Windows XP or 2003, select Add or Remove Programs. In Windows Vista, select Programs, then select Programs and Features.
  3. A list of programs installed on the computer is displayed. Scroll down if necessary and select XenCenter.
  4. In Windows XP or 2003, click the Remove button. In Windows Vista, select Uninstall from the toolbar above the list of programs.This will remove the Citrix application. At the end, a message is displayed. Click OK to close the message box.

This section describes several common installation and deployment scenarios:

and details the steps that differ between scenarios.

The simplest use of XenServer is to set up a simple network of VMs running on one or more XenServer hosts without shared storage. This, of course, means that live relocation of VMs from one XenServer host to another is not possible, as this requires shared storage.

Requirements

Basic procedure

  1. Install XenServer host software on server(s)
  2. Install XenCenter on workstation(s)
  3. Run XenCenter and connect to XenServer hosts

Adding shared storage to the XenServer network enables grouping of XenServer hosts into resource pools, enabling live relocation of VMs and sharing of server resources.

Requirements

Note

To be part of a resource pool, the XenServer hosts and the server or servers providing the shared NFS storage need to have static IP addresses.

Basic procedure

  1. Install XenServer host software on server(s)
  2. Install XenCenter on workstation(s)
  3. Set up the NFS server
  4. Run XenCenter and connect to XenServer hosts
  5. Choose one XenServer host as a pool master and join other XenServer hosts to its pool.
  6. Create an SR on the NFS share at the pool level

For this procedure, a server running a typical Linux distribution is assumed as the NFS server. Consult your Linux distribution documentation for further information.

Set up NFS share on NFS server

  1. Check to see if the portmap daemon is installed and running:
    # chkconfig --list portmap
portmap         0:off   1:off   2:off   3:on    4:on    5:on    6:off
    Note that in the preceding example, runlevels 3, 4, and 5 say "on". That means that at boot, for runlevels 3, 4 and 5, the portmap daemon is started automatically. If either 3, 4 or 5 say "off," turn them on with the following command:
    chkconfig portmap on
  2. Check to see if the NFS daemon is installed and running:
    # chkconfig --list nfs
nfs             0:off   1:off   2:on    3:on    4:on    5:on    6:off
    If either 3, 4 or 5 say "off," turn them on with the following command:
    chkconfig nfs on
  3. Make a directory for the shared storage to live in:
    mkdir /<vm_share_dir>
  4. Edit the file /etc/exports and add the line
    /<vm_share_dir>     *(rw,no_root_squash,sync)
    Save and close the file.
  5. Restart the portmap and nfs daemons as follows:
    service portmap restart
service nfs restart
    The <vm_share_dir> should now be exported on the network and you should be able to use XenCenter to point to it using the Storage wizard. See the XenCenter online help for details.

Create an SR on the NFS share at the pool level

  1. Open a host text console on any XenServer host in the pool.
  2. Create the storage repository on server:/path
    xe sr-create content-type=user type=nfs name-label=<sr_name=> \
					shared=true device-config-server=<server> \
					device-config-serverpath=<path>
    The device-config-server refers to the hostname of the NFS server and device-config-serverpath refers to the path on the server. Since shared is set to true, the shared storage will be automatically connected to every host in the pool and any hosts that subsequently join will also be connected to the storage. The UUID of the created storage repository will be printed on the screen.
  3. Find the UUID of the pool
    xe pool-list
  4. Set the shared storage as the pool-wide default
    xe pool-param-set uuid=<pool_uuid> \
					default-SR=<storage_repository_uuid>
    Since the shared storage has been set as the pool-wide default, all future VMs will have their disks created on shared storage by default.

Adding shared storage to the XenServer network enables grouping of XenServer hosts into resource pools, enabling live relocation of VMs and sharing of server resources.

Requirements

Note

To be part of a resource pool, the XenServer hosts and the server or servers providing the shared iSCSI storage need to have static IP addresses.

Basic procedure

  1. Install XenServer host software on server(s)
  2. Install XenCenter on workstation(s)
  3. Prepare the iSCSI storage
  4. If necessary, enable your iSCSI device for multiple initiators
  5. Run XenCenter and connect to XenServer hosts
  6. Choose one XenServer host as a pool master and join other XenServer hosts to its pool
  7. Configure the iSCSI IQN for each XenServer host
  8. Create an SR on the iSCSI share at the pool level

The details of how to set up iSCSI storage differ between the various iSCSI solutions on the market. In general, though, you need to provide an iSCSI target on the SAN for the VM storage, and then configure XenServer hosts to be able to see and connect to it. This is done by providing a valid iSCSI Qualified Name (IQN) to the iSCSI target and to the iSCSCI initiator on each XenServer host.

Prepare the iSCSI storage

  1. Assign a virtual storage volume on the iSCSI SAN for VM storage
  2. Create IQNs on the SAN for each XenServer host that will use the storage

You can use either XenCenter or the CLI to configure the IQN for each XenServer host and to create the SR. The following describes using the CLI; see the XenServer Help for details on using XenCenter.

Warning

When using the XenCenter to create SRs for iSCSI and NetApp storage, any existing contents of the volume will be destroyed.

To configure the iSCSI IQN for each XenServer host via the CLI

  1. In the Text Console, issue the command:
    xe-set-iscsi-iqn <iscsi_iqn>
    Alternatively, the CLI can be used directly:
    xe host-param-set uuid=<host_uuid> other-config-iscsi_iqn=<iscsi_iqn>
  2. Repeat for each XenServer host in the pool.

To create an SR on the iSCSI share at the pool level via the CLI

  1. In the Text Console of any server in the pool, issue the command:
    xe sr-create name-label=<name_for_sr>
content-type=user device-config-target=<iscsi_server_ip_address> 
device-config-targetIQN=<iscsi_target_iqn>
device-config-localIQN=<iscsi_local_iqn>
type=lvmoiscsi shared=true device-config-LUNid=<lun_id>
    The device-config-target argument refers to the hostname or IP address of the iSCSI server. The device-config-LUNid argument can be a comma-separated list of LUN IDs. Since the shared argument is set to true, the shared storage will be automatically connected to every host in the pool and any hosts that subsequently join will also be connected to the storage.The command returns the UUID of the created storage repository.
  2. Find the UUID of the pool by issuing the command
    xe pool-list
  3. Set the shared storage as the pool-wide default as follows:
    xe pool-param-set uuid=<pool_uuid> default-SR=<iscsi_shared_sr_uuid>
    Now that the shared storage has been set as the pool-wide default, all future VMs will have their disks created on shared storage by default.

Table of Contents

5.1. Preparing XenServer hosts for upgrade
5.2. Applying updates
5.2.1. Applying updates using the CLI
5.3. Reinstalling the same version
5.4. Rolling upgrade between major versions
5.4.1. Backup
5.4.2. Performing a rolling pool upgrade
5.5. Upgrading from version 3.2 to 4.0.1
5.6. Upgrading VHD files from version 4.0.1
5.7. Backing up and restoring XenServer hosts and VMs
5.7.1. Backing up Virtual Machine metadata
5.7.2. Backing up XenServer hosts
5.7.3. Backing up VMs

This chapter documents how to update (apply minor update patches), upgrade from an earlier version, or freshen (reinstall the same version) your XenServer installation.

Before performing maintenance operations on a XenServer host that is part of a resource pool, you should disable it (which prevents any VMs from being started on it), then migrate its VMs to another XenServer host in the pool. This can most readily be accomplished by placing the XenServer host into Maintenance Mode using XenCenter. See the XenCenter Help for details.

To prepare a XenServer host in a pool for maintenance operations via the CLI

  1. Use the command
    xe host-evacuate uuid=<xenserver_host_uuid>
    This will disable the XenServer host and then migrate any running VMs to other XenServer hosts in the pool.
  2. Perform the desired maintenance operation.
  3. Once the maintenance operation is completed, enable the XenServer host:
    xe host-enable
    then restart any halted VMs and/or resume any suspended VMs.

Before performing maintenance operations on a XenServer host that is not part of a resource pool, you should disable it (which prevents any VMs from being started on it), then either shutdown or suspend its VMs.

Warning

Note that any suspended VM with a CD drive attached (with the Tools ISO or a physical CD in the local physical drive, for example) cannot be resumed after performing an upgrade. To get a suspended VM back into a usable state, you have to perform a "Force Shutdown" on it and then restart it.

To prepare an unpooled XenServer host for upgrade via the CLI

  1. Disable the XenServer host:
    xe host-disable
  2. Shut down or suspend any running VMs using the xe vm-shutdown or xe vm-suspend commands. If you choose to suspend any VMs, check that no CDs are attached to them.
  3. Perform the desired maintenance operation.
  4. Once the maintenance operation is completed, enable the XenServer host:
    xe host-enable
    then restart any halted VMs and/or resume any suspended VMs.

Between releases of XenServer software, Citrix occasionally releases updates to the software. These updates typically contain accumulated bug fixes and feature improvements. When an update is released, it is made accessible on the Internet and an email announcement is sent to all XenServer customers.

Once downloaded, updates can be applied most readily via XenCenter, but can also be applied using the CLI. Updates are applied through the Manage Updates dialog box, under the pool menu. See the XenCenter Help for details.

Updates sometimes have special steps that have to be performed after the update is applied, such as requiring the XenAPI agent to be restarted. Whenever possible, updates will be such that they can be applied without interruption, but in some cases they might require XenServer host or VM restarts to be performed. In cases where a XenServer host restart is required, you can avoid downtime of virtual machines in a pooled environment by applying the update to each server in turn, migrating VMs away from each server in turn as the update is applied. XenCenter can take care of this update sequence automatically on your behalf via the Manage Updates feature. If you are using the CLI, you will have to do this manually using the host-evacuate command.

If using the CLI to perform the update, XenServer hosts to be updated should be prepared for this operation by the procedures in Section 5.1, “Preparing XenServer hosts for upgrade”. If using XenCenter, this will be taken care of automatically where required.

First, the update must be uploaded to the pool or server to which it will be applied. This will cause a UUID (identifier) to be assigned to the update, and information about which servers it has been applied to will be tracked. Once an update has been uploaded to a pool or server, you can use the patch-list and patch-param-list commands to view information about the update. The second stage is to apply the update. We recommend that the patch-pool-apply command be used to do this; this will result in the update being applied on all servers in the pool. Alternatively, the patch-apply command may be used to apply the update to one server in a pool - this may be useful when applying the update and then restarting individual servers in the pool. Pools should not be left in an inconsistent update state (one where updates have been installed on some servers and not others).

Discussion of procedures using the CLI below assume a basic knowledge of the usage of the xe tool. For information about this, please see the Administrator's Guide.

The update procedure is essentially the same for both a single server and pool scenario, except that in a pooled scenario you must ensure that the update is applied to all servers in the pool. This will be achieved either by using the patch-pool-apply command, or by executing the patch-apply once for each host. These are described below.

To apply an update to a XenServer host or XenServer host pool using the CLI

  1. Download the update to a local directory. Note the path to the update file you have downloaded. (It is also possible to download the update directly to an appropriate location on the server, e.g. /root, using standard Linux commands, but it is usually best to download it to a remote client.)
  2. Upload the update to your server or pool. An example CLI command to do this might be:
    xe -s my_server -u root -pw root_password patch-upload \
file-name=update_file b89249c7-feba-41c5-8838-911ded969add
    Here, the -s -u, and -pw options refer to the server, the username (which would usually be root), and the password, as usual - these would be omitted if running the command directly from the XenServer host console.Once you have executed the above command, you will be given the UUID of the uploaded update. This UUID will be used to specify the update that is to be applied.
  3. Be sure to follow any guidance regarding the update before continuing, in particular any information provided about whether VMs should be moved away from the server or that the server should be restarted after applying the update. As always, we recommend taking appropriate backup measures before making modifications to system software. To automatically move VMs to other servers, you can make use of the host-evacuate CLI command.
  4. Apply the update to the pool. A command like the following may be used to do this:
    xe patch-pool-apply uuid=b89249c7-feba-41c5-8838-911ded969add
    This will apply the update to all servers in the pool. Alternatively, if you need to restart servers and perform the update in a rolling manner, you can apply the update to an individual server by running a command like the following:
    xe patch-apply host-uuid=ebf17583-d8c5-4835-999a-e0a29716207d \
uuid=b89249c7-feba-41c5-8838-911ded969add
  5. Verify that the update was applied by using the patch-list command again. Now the hosts field should contain the host UUID.

After an update is applied to a XenServer host, a small file containing the same information stored on the master from the xe patch-upload command is written to a subdirectory of the machine's patch directory. This enables XenServer hosts later ejected from the pool to repopulate their databases with information about updates already applied.

To save space on the master, large updates can be deleted from disk using the xe patch-clean command. (The update information stored in the master's database, though, is always retained.) These updates can be uploaded again using xe patch-upload if required.

The following procedure describes how to "freshen," or reinstall the current version of the XenServer host over an existing installation of XenServer host 5.0.0, and preserve settings on VMs.

Warning

When reinstalling your host, be aware that

To reinstall XenServer host from version 5.0.0

  1. Perform an orderly shutdown on the VMs hosted on the XenServer host. If any of your VMs are in the suspended state, resume them first, and then perform an orderly shutdown on them too.To shut down all the VMs automatically, you can also type service xapi-domains stop into the control domain terminal.
  2. Reboot the XenServer host, and boot from the Installation CD.
  3. The installation script will identify the version and prompt you whether you want to reinstall over the existing installation and preserve VMs. Select OK to proceed with the installation.
  4. Follow the rest of the installation procedure as described in Section 4.1, “Installing the XenServer host”.
  5. Run XenCenter and connect to the upgraded XenServer host.

XenServer allows you to upgrade a pool of XenServer hosts to the next major version, while keeping VMs on that pool running and thus avoiding downtime of your services. This is achieved by upgrading on a host-by-host basis, with only one XenServer host offline at a time.

Rolling upgrade is supported from:

Note

This means that to upgrade from 4.0.x to 5.0.x, you first upgrade 4.0.x to 4.1.x, and then upgrade to 5.0.x.

You can use XenCenter or the command line interface to migrate VMs running on a XenServer host running an older version of the product to one running either the same version or higher. It is not possible to migrate VMs located on a XenServer host with a newer XenServer version to one running an older version.

You should plan your upgrade path carefully, as we strongly advise against running a mixed-mode pool (one with multiple versions of XenServer co-existing) for longer than necessary. This is because the pool will be operating in a degraded state during the upgrade: all VMs will continue to function as normal, but control operations other than migration might not be available. In particular, it is not safe to perform storage-related operations such as adding, removing or resizing virtual disks in this mode.

It is strongly recommended that you take a backup of the state of your existing pool using the pool-dump-database CLI command (see the XenServer Administrator's Guide). This will allow you revert a partially complete rolling upgrade back to its original state without losing any VM data.

Note that since it is not possible to migrate a VM from an upgraded XenServer host to a XenServer host running an older version of XenServer, it may be necessary to shut down VMs if you need to revert the rolling upgrade for any reason.

Warning

Disable HA functionality before performing a rolling upgrade.

The correct sequence for upgrading a pool of XenServer installations to a newer version is as follows:

  1. Eject any CDs from your Virtual Machines before starting the rolling upgrade. Having CDs inserted during rolling upgrade can prevent migrations from working correctly, and due to the mode of operation of the pool while the rolling upgrade is taking place, it is required that this be done before the rolling upgrade is started.
  2. Upgrade your XenCenter to the latest version. The newer version will continue to operate fine against older versions of XenServer hosts.
  3. Verify that there are no VMs in the Suspended state. This is indicated in XenCenter by a blue paused icon. Any suspended VM with a CD drive attached (with the Tools ISO or a physical CD in the local physical drive, for example) cannot be resumed after performing an upgrade. To get a suspended VM back into a usable state, you have to perform a "Force Shutdown" on it and then restart it.
  4. Migrate all VMs running on the pool master to other XenServer hosts using XenMotion. The pool master is identified in XenCenter as being the topmost server in the pool, and shows Server type: Master in the General tab when selected.
  5. Shut down the pool master using XenCenter or the command line interface. This will cause your pool to enter emergency mode. VMs will continue to run, but you will be unable to perform control operations. This is expected behavior.
  6. Boot the pool master using your XenServer installation media or network and follow the instructions for doing an upgrade (see Chapter 4, Installing XenServer).
  7. On restarting your pool master, after a few minutes your pool will leave emergency mode and normal service will be restored.
  8. You are now ready to upgrade a second XenServer host. You should select a XenServer host still running an old version of XenServer and migrate the VMs running on this XenServer host to the one you have just upgraded. Do not attempt to migrate a VM from an upgraded XenServer host to one that has not yet been upgraded. You will see an error message if you attempt to do this, and your VM will continue running without being migrated.
  9. Upgrade the member XenServer host you have just freed up following a similar procedure as for the master; shut down the member using XenCenter or the command line interface (your pool will not enter emergency mode this time), then upgrade the server software using your product media or remote installation repository.
  10. Repeat the previous two steps for each member XenServer host in the pool.
  11. Now that you have upgraded the XenServer host software on your pool, it is important to upgrade the Citrix Tools for Virtual Machines on each VM. This will enable new functionality and ensure the stability of your VMs. Running old versions of Citrix Tools for Virtual Machines on newer XenServer installations is not a supported configuration except for during the upgrade process. Please refer to the XenCenter Help, or the XenServer Virtual Machine Installation Guide for details on how to perform the upgrade of Citrix Tools for Virtual Machines for Windows and Linux VMs.

Note

In the unlikely event that a host fails (hardware failure, for example) during the rolling upgrade process, it is necessary to use the xe host-forget command to forget the host. Failure to do so will result in XenServer remaining stuck in rolling upgrade mode indefinitely.

The following procedure describes how to install the current version of the XenServer host over an existing installation of XenServer host 3.2.0.

When upgrading from version 3.2.0 to the current version, be aware of the following:

To upgrade XenServer host from version 3.2

  1. Perform an orderly shutdown on the VMs hosted on the XenServer host.If any of your VMs are in the Suspended state, Resume them first, and then perform an orderly shutdown on them too.
  2. Reboot the XenServer host, and boot from the 4.0.1 Installation CD.
  3. The installation script will identify the older version and prompt you whether you want to install over the existing 3.2 installation and preserve VMs. Select OK to proceed with the installation.
  4. Follow the rest of the installation procedure as described in the 4.0.1 installation guide.
  5. Run XenCenter and connect to the upgraded XenServer host.
  6. To upgrade the drivers for a Windows VM, select the "Install Tools" menu option and open its console. Run the xensetup.exe installation program to upgrade your paravirtualized drivers. When finished, reboot the VM.To upgrade the kernel and guest utilities for Linux VMs, follow the instructions in the XenServer 4.0.1 Virtual Machine Installation Guide.Repeat for all other Windows VMs.

Note that there is no direct upgrade path to the current version from 3.2. You must first upgrade to 4.0.1, and then proceed to the next version. This will ensure that your VMs are preserved correctly.

The EXT and NFS storage repository types store data in the Microsoft Virtual Hard Disk (VHD) format under the 4.0.1 version of XenServer that has a free space bitmap written in with the wrong byte order. Version 4.1 and higher provides an upgrade utility vhd-update, located in /usr/sbin/.

Whilst it is not necessary to use this tool under XenServer, you may need to run it in order to safely interchange your Version 4.0.1 VHD files with a third party tool that uses the VHD format. If you wish to use this utility, you should ensure that the VHD files are not being used while you are working with them, and that all the relevant VHD files are upgraded (such as the parent VHD of a snapshot). Working with this tool should be considered an advanced operation, so appropriate backup measures should be taken.

The command

vhd-update -f <filename>

copies all bitmaps in filename to filename.journal, then writes all bitmaps, transformed appropriately, from filename.journal to filename, and finally deletes filename.journal.

If the update is interrupted for any reason, it can be resumed by running:

vhd-update -f <filename> -j <filename>.journal

This will validate the specified journal file, then proceed to transform the bitmaps and rewrite filename.

There is also a rollback operation

vhd-update -f <filename> -j <filename>.journal -r

which will write the bitmaps from filename.journal to filename without transforming them (this is probably only useful for testing and debugging).

Note

We recommend that, whenever possible, you leave the installed state of XenServer hosts unaltered. That is, do not install any additional packages or start additional services on XenServer hosts, and treat them as if they are appliances. The best way to restore, then, is to re-install XenServer host software from the installation media. If you have multiple XenServer hosts, the best approach is to configure a PXE boot server and appropriate answerfiles for this purpose (see Appendix B, PXE installation of XenServer host).

For VMs, the best approach is to install backup agents on them, just as if they were standard physical servers. For Windows VMs, as of this release we have tested CA BrightStor ARCserve Backup, and Symantec NetBackup and Backup Exec.

For more information about backup tools tested, best practices, and backups in general, see the Citrix Knowledge Base.

XenServer hosts use a per-host database to store metadata about VMs and associated resources such as storage and networking. When combined with storage repositories, this database forms the complete view of all VMs available across the pool. Thus, it is important to understand how to backup this database in order to recover from physical hardware failure and other disaster scenarios.

This section first describes how to backup metadata for single-host installations, and then for more complex pool setups.

The CLI must be used to backup the pool database. To obtain a consistent pool metadata backup file, run xe pool-dump-database against the XenServer host and archive the resulting file. The backup file will contain sensitive authentication information about the pool, so ensure it is securely stored.

To restore the pool database, use the xe pool-restore-database from a previous dump file. If your XenServer host has died completely, then you must first do a fresh install, and then run the xe pool-restore-database command against the freshly installed XenServer host.

After a restoration of the pool database, some VMs may still be registered as being “suspended”, but if the storage repository with their suspended memory state (defined in the suspend-VDI-uuid field) was a local SR, it will no longer be available since the host has been reinstalled. To reset these VMs back to the halted state so that they can be started up again, use the xe vm-shutdown vm=vm_name -force command, or use the xe vm-reset-powerstate vm=vm_name -force command.

Note that XenServer hosts restored using this method will have their UUIDs preserved. Thus, if you restore to a different physical machine while the original XenServer host is still running, there will be a UUID clash. The main observable effect of this clash will be that XenCenter will refuse to connect to the second XenServer host. Pool database backup is not the recommended mechanism for cloning physical hosts; you should use the automated installation support for that (see Appendix B, PXE installation of XenServer host).

In a pool scenario, the master host provides an authoritative database which is synchronously mirrored by all the member hosts in the pool. This provides a degree of built-in redundancy to a pool; the master can be replaced by any member since each of them have an accurate version of the pool database. Please refer to the XenServer Administrator's Guide for more information on how to transition a member into becoming a master host.

This level of protection may not be sufficient; for example, if your shared storage containing the VM data is backed up in multiple sites, but your local server storage (containing the pool metadata) is not. To fully recreate a pool given just a set of shared storage, you must first backup the xe pool-dump-database against the master host, and archive this file.

To subsequently restore this backup on a brand new set of hosts

  1. Install a fresh set of XenServer hosts from the installation media, or via PXE.
  2. Use the xe pool-restore-database on the host designated to be the new master.
  3. Run the xe host-forget command on the new master to remove the old member machines.
  4. Use the xe pool-join command on the member hosts to connect them to the new cluster.

Please refer to the "Coping with machine failures" section of the XenServer Administrator's Guide for specific restoration scenarios.

This section describes the XenServer host control domain backup and restore procedures. These procedures do not back up the storage repositories that house the VMs, but only the privileged control domain that runs Xen and the XenServer agent.

Note that since the privileged control domain is best left as installed, without customizing it with other packages, we recommend you set up a PXE boot environment to cleanly perform a fresh installation from the XenServer media as a recovery strategy. In many cases you will not need to backup the control domain at all, but just save the pool metadata (see Section 5.7.1, “Backing up Virtual Machine metadata”). This backup method should always be considered complementary to backing up the pool metadata.

Another approach is to run the XenServer installation twice, selecting to back up the existing installation when prompted. This will create a pristine copy of the freshly-installed control domain that can later be restored if necessary by using the installation CD and choosing the Restore option.

Using the xe commands host-backup and host-restore is another approach that you can take. The xe host-backup command archives the active partition to a file you specify, and the xe host-restore command extracts an archive created by xe host-backup over the host's currently inactive disk partition. This partition can then be made active by booting off the installation CD and choosing the Restore option.

After completing the above steps and rebooting the host, you must ensure that the VM meta-data is restored to a consistent state. This can be achieved by running xe pool-restore-database on /var/backup/pool-database-${DATE}. This file is created by xe host-backup using xe pool-dump-database prior to archiving the running filesystem, in order to snapshot a consistent state of the VM metadata.

To back up a XenServer host

To restore a running XenServer host

  1. If you want to restore a XenServer host from a specific backup, run the following command while the XenServer host is up and reachable:
    xe host-restore file-name=<filename> -h <hostname> -u root -pw <password>;
    This restores the compressed image back to the hard disk of the XenServer host on which the command is run (not the host on which filename resides). In this context “restore” is something of a misnomer, as the word usually suggests that the backed-up state has been put fully in place. The restore command here only unpacks the compressed backup file and restores it to its normal form, but it is written to another partition (/dev/sda2) and does not overwrite the current version of the filesystem.
  2. To use the restored version of the root filesystem, you need to reboot the XenServer host using the XenServer installation CD and select the Restore from backup option.After the restore from backup is completed, reboot the XenServer host and it will start up from the restored image.Finally, restore the VM meta-data using
    xe pool-database-restore file-name=/var/backup/pool-database-*

Note

Restoring from a backup as described here does not destroy the backup partition.

To restart a crashed XenServer host

  1. If your XenServer host is crashed and not reachable anymore, you need to use the XenServer installation CD to do an upgrade install (see Section 5.5, “Upgrading from version 3.2 to 4.0.1”). When that is completed, reboot the machine and make sure your host is reachable with XenCenter or remote CLI.
  2. Then proceed with the procedure on restoring a running XenServer host above.

VMs are best backed up using standard backup tools running on them individually. For Windows VMs, we have tested CA BrightStor ARCserve Backup.

If you experience odd behavior, crashes, or have other issues during installation, this chapter is meant to help you solve the problem if possible and, failing that, describes where logs are located and other information that can help your Citrix Solution Provider and Citrix track and resolve the issue.

Note

We recommend that you follow the troubleshooting information in this chapter solely under the guidance of your Citrix Solution Provider or Citrix Support.

Citrix provides two forms of support: you can receive free self-help support via the Support site, or you may purchase our Support Services and directly submit requests by filing an online Support Case. Our free web-based resources include product documentation, a Knowledge Base, and discussion forums.

The XenServer host installation CD runs Linux, so most standard Linux commands can be used to diagnose installation problems. There are three virtual terminals available during installation, which display the installation menu, an interactive console and an event log, respectively. Use the ALT + F1-F3 keys to toggle between the virtual terminals.

You can check some basic things in the interactive terminal:

  • fdisk lists all disks that can be seen as a result of the loaded storage device drivers. If a particular device driver did not load, for example, the driver for a RAID card, then the disks attached to that card will not appear in the output from the fdisk command.
  • ifconfig shows the network configuration of physical NICs, including their IP addresses, netmasks, and gateway.
  • ping can be used to verify network connectivity from the XenServer host to a remote IP address and vice-versa.

You should use the two additional virtual terminals solely under the guidance of your Citrix Solution Provider.

Installation logs are written to /install/tmp/

Table of Contents

B.1. Setting up the PXE boot environment
B.2. Creating an answerfile for unattended PXE installation
B.3. Installation media repository format
B.3.1. Presence of installation media repositories
B.3.2. Installation media repository metadata
B.3.3. Package metadata
B.3.4. Example files
B.3.5. Notes on best practice

This appendix describes setting up a TFTP server to enable PXE booting of XenServer host installations. It also describes the use of an XML answerfile, which allows you to perform unattended installations.

To create a PXE environment, you need:

These can all co-exist on the same server, or be distributed on different servers on the network.

Additionally, each system that you want to PXE boot and install the XenServer on needs a PXE boot-enabled Ethernet card.

The following steps assume that the Linux server or servers you will use have RPM support.

To set up a TFTP server for PXE booting

  1. TFTP requires SYSLINUX 3.11 or above. SYSLINUX is a collection of boot loaders for the Linux operating system which operates on Linux EXT2/EXT3 file systems, MS-DOS FAT file systems, network servers using PXE firmware, and CD-ROMs. Make sure you have SYSLINUX version 3.11 or above installed on your system with the command
    #rpm -q syslinux
    If you have an earlier version, you can download an appropriate later version from ftp://ftp.kernel.org/pub/linux/utils/boot/syslinux/RPMS/i386/, then install it using the command
    #rpm -Uvh syslinux.-.rpm
  2. Check if the tftp server package is installed:
    #rpm -q tftp-server
    If not, use system-config-packages and install.
  3. Edit the file /etc/xinetd.d/tftp to change the line
    disable = yes
    to
    disable = no
  4. Restart the xinetd service, which manages tftp:
    # service xinetd restart
  5. Make a directory inside /tftpboot called xenserver.
  6. Copy the files mboot.c32 and pxelinux.0 from /usr/lib/syslinux to the /tftboot directory.
  7. Copy the files install.img, vmlinuz, and xen.gz from the Base Pack CD (found in the root of the Base Pack CD, and in its /boot directory respectively), and place them in /tftpboot/xenserver.
  8. Make a directory called pxelinux.cfg inside /tftboot and create a file named default. The file contents depend on how you want to configure your PXE boot environment. For example, you might have a configuration file like the following:

    Note

    The backslashes at the ends of lines in the example PXE configuration files shown below denote continuation of lines; do not actually include them in your PXE configuration file.

    Also note that the three hyphens in the examples are necessary parts of the mboot.c32 loader syntax, and not including them will cause PXE boot attempts to fail.

    default xenserver
label xenserver
   kernel mboot.c32
   append path/to/boot/directory/xen.gz dom0_mem=752M com1=115200,8n1i \
      console=com1,tty --- path/to/boot/directory/vmlinuz \
      console=ttyS0,115200n8 console=tty0 \
	  --- path/to/boot/directory/install.img
    (where path/to/boot/directory is the directory where you copied install.img, vmlinuz, and xen.gz files in the previous step). This will start an installation on any machine that boots from this server. Someone would then need to manually respond to the prompts to complete the installation. Alternatively, you might have a configuration file like the following:
    default xenserver-auto
label xenserver-auto
   kernel mboot.c32
   append path/to/boot/directory/xen.gz dom0_mem=752M com1=115200,8n1 \
      console=com1,tty --- path/to/boot/directory/vmlinuz \
      console=ttyS0,115200n8 console=tty0 \
      answerfile=http://pxehost.example.com/4.1.0-answerfile \
      install --- path/to/boot/directory/install.img
    This will perform an unattended installation using the answerfile at the URL specified.

    Note

    The above examples show how to configure the installer to run on the physical console, tty0. It is recommended that you place the console= entry of the console you wish to use for the installation as the last entry on the line, so if you wanted to install over serial then, in the examples above, the two entries would be reversed, as console=tty0 console=ttyS0,115200n8

    For details on creating an answerfile for unattended installation, see     Section B.2, “Creating an answerfile for unattended PXE installation”. For more information on PXE configuration file syntax, see the SYSLINUX website.

Please refer to your server operating system manual for details for your specific operating system. The information here is a guide that can be used for Red Hat, Fedora, and some other RPM-based distributions.

To set up a DHCP server

  1. On the server that you will be using for DHCP, check if you have DHCP installed by issuing the command
    rpm -qa dhcp
    If not, install using system-config-packages.
  2. Configure the dhcp server. Refer to article 4221 in the Red Hat Knowledge base for details.
  3. Add these lines to the end of the existing dhcpd.conf file where is your tftp server address:
    allow booting;
allow bootp;
class "pxeclients" {
    match if substring(option vendor-class-identifier, 0, 9) = "PXEClient";
    next-server ;
    filename "pxelinux.0";
}
  4. Restart the dhcpd service:
    # service dhcpd restart

To set up the installation media host

  1. On the server where you are going to house the installation files, copy the contents of the packages directories from the Base Pack CD to a location where they are exported by HTTP, FTP, or NFS. For example, you might make a directory in the document root of a webserver named XenServer_5.0.0 and then copy the directory packages.main from the Base Pack disk to XenServer_5.0.0/packages.main.
  2. If Linux support is also desired, copy packages.linux from the Linux Pack disk to XenServer_5.0.0/packages.linux. This structure allows you to install either both packages by having the answerfile's source element contain the enclosing directory XenServer_5.0.0, or you can install just the base pack (no support for Linux VMs) by putting in the path to XenServer_5.0.0/packages.main.

For example, to install both packages from the webserver http://pxehost.example.com where the packages are in the directories mentioned above relative to the server's document root, the answerfile would contain this source element:

<source type="url">http://pxehost.example.com/XenServer_4.1.0</source>

or, to install just the basic pack and skip Linux support:

<source type="url">
http://pxehost.example.com/XenServer_4.1.0/packages.main
</source>

In either case you can specify a username and password if required, for example:

<source type="url">
http://<username>:<password>@pxehost.example.com/XenServer_4.1.0/packages.main
<source>

To prepare the destination system

  1. Start the system and enter the Boot Menu (F12 in most BIOSes) and select to boot from your Ethernet card.
  2. The system should then PXE boot from the installation source you set up, and the installation script will commence. If you have set up an answerfile, the installation can proceed unattended.

In order to perform installations in an unattended fashion, you need to create an XML answerfile.

Here is an example answerfile:

<?xml version="1.0"?>
   <installation>
      <primary-disk>sda</primary-disk>
      <guest-disk>sdb</guest-disk>
      <guest-disk>sdc</guest-disk>
      <keymap>us</keymap>
      <root-password>mypassword</root-password>
      <source type="url">http://pxehost.example.com</source>
	  <post-install-script type="url">
	  http://pxehost.example.com/myscripts/post-install-script
	  </post-install-script>
      <admin-interface name="eth0" proto="dhcp" />
      <timezone>Europe/London</timezone>
   </installation>

All nodes should be within a root node named installation.

The following is a summary of the elements. All values should be PCDATA within the nodes, unless otherwise stated. Required elements are indicated.

ElementDescriptionRequired?
<primary-disk> The name of the storage device where the Dom0 should be installed, equivalent to the choice made on the Select Primary Disk step of the interactive installation process.

Attributes:

You can specify a gueststorage attribute with possible values yes and no. For example:

<primary-disk gueststorage="no">sda</primary-disk>

If this attribute is not specified, the default is yes. If you specify no, it is possible to automate an installation scenario where no storage repository is created, if, in addition, no guest-disk keys are specified.

Y
<guest-disk>The name of a storage device to be used for storing guests. You should use one of these elements for each extra disk.N
<keymap>The name of the keymap to use during installation.

<keymap>us</keymap>

Y
<root-password>The desired root password for the XenServer host.Y
<source>Where the packages should be installed from.

Attributes:

type: url, nfs, or local

If local, leave the PCDATA empty. For example,

<source type="url">
http://server/packages
</source>
<source type="local" />
<source type="nfs">
server:packages
</source>
Y
<post-install-script>Where the post-install-script is.

Attributes:

type: url, nfs, or local

If url or nfs, put the url or NFS path in the PCDATA; if local, leave the PCDATA empty. For example,

<source type="url">
http://server/scripts
</source>
<source type="local" />
<source type="nfs">
server:scripts
</source>
Y
<admin-interface>The single network interface to be used as the host administration interface.

Attributes:

proto: dhcp or static

name: eth0 for example.

Children:

  • <ip>: The IP address, if proto="static"
  • <subnet-mask>: The subnet mask, if proto="static"
  • <gateway>: The gateway, if proto="static"

All three child elements are required if proto="static"

N
<timezone>In the format used by the TZ variable, e.g. Europe/London, or America/Los_Angeles.Y
<nameserver>The name of a nameserver. You should use one of these elements for each nameserver you wish to nominate.N
<hostname>Specify if you want to manually set a hostname.N
<bootloader>Specify which bootloader to install for startup time. Only change this if you have problems booting. Currently either grub (the default) or extlinux.N

You can also perform automated upgrades by varying the answerfile appropriately. You would set the mode attribute of the installation element to reinstall, specify the disk on which the existing installation lives with the existing-installation element, and leave the primary-disk and guest-disk elements unspecified. For example:

<?xml version="1.0"?>
   <installation mode="reinstall">
      <existing-installation>sda</existing-installation>
      <keymap>us</keymap>
      <root-password>mypassword</root-password>
      <source type="url">http://pxehost.example.com</source>
	  <post-install-script type="url">
	  http://pxehost.example.com/myscripts/post-install-script
	  </post-install-script>
      <admin-interface name="eth0" proto="dhcp" />
      <timezone>Europe/London</timezone>
   </installation>

The repository format described here should be used by installation sources and driver disks.

Given a path, the presence of a Citrix installation media repository is determined by checking for the existence of valid XS-REPOSITORY and XS-PACKAGES files. From a given base, that base is checked, along with the packages, packages.main, packages.linux, and packages.site subdirectories. Thus, a typical installation point will have the following format:

xs-installation
  +-- packages.main
  |   +-- XS-REPOSITORY
  |   +-- XS-PACKAGES
  |   +-- ...
  +-- packages.linux
  |   +-- XS-REPOSITORY
  |   +-- XS-PACKAGES
  |   +-- ...
  +-- packages.site
  |   +-- XS-REPOSITORY
  |   +-- XS-PACKAGES
  |   +-- ...

A typical driver disk will have the following layout:

xs-driver-disk
  +-- XS-REPOSITORY
  +-- XS-PACKAGES

In the first example, given a path to xs-installation, the XenServer installer will detect the presence of three repositories. In the second example, xs-driver-disk, a single repository will be detected.

The XS-REPOSITORY file is used to describe a Citrix-format installation media repository. It has four fields, separated by newlines:

Repository IDs should be alphanumeric strings that provide a machine identifier for the repository. They should be unique within a target product and version. Best practice is to use the form

vendor:repository

Citrix repositories start with xs (for example, xs:main), custom repositories should be custom:my-repo, and third-party add-ons should be identified as such by using an appropriate vendor string. This will help avoid name clashes.

Repository names are presented to the user, so should be a string that identifies the repository in a sensible manner so the user can confirm that they wish to install from it.

The intended target product will be XenServer; version 5.0.0-build.

The XS-PACKAGES file describes the packages in a repository, one line per package. Fields are separated by spaces.

There are three types of packages:

Firmware loading support is currently limited; this will be addressed in a future release.

The first three fields are mandatory: package name, package size, and package checksum (md5). The fourth field is the package type, either tbz2, driver, or firmware. Which type is used dictates the contents of the subsequent fields.

If the type is tbz2, the subsequent fields are required or optional, source filename, and destination (usually just /).

Example:

docs 37750 2ba1783d84d10c71f07469252c555427 tbz2 required docs.tar.bz2 /

If the type is driver, the subsequent fields are source filename and destination (${KERNEL_VERSION} will be substituted with the Xen-kernel version.)

Example:

firmware_example 77001 3452c04dfcc237cde11c63d43e97a303 driver \
firmware_example.ko \
/lib/modules/${KERNEL_VERSION}/extra/firmware_example.ko

If the type is firmware, the subsequent field is destination filename (no path is necessary - it is automatically prefixed with /lib/firmware/).

Example:

firmware 12 6f5902ac237024bdd0c176cb93063dc4 firmware sample_firware.bin

Note

The backslashes at the ends of lines in the examples in this section denote continuation of lines; do not actually include them in a XS-PACKAGES file.

xs:main
Base Pack and extra driver
XenServer
3.2.0-1934

storage-manager 59831 b66672f0aa681bd2b498e3d902f17c04 tbz2 required \
storage-manager.tar.bz2 /
docs 37750 2ba1783d84d10c71f07469252c555427 tbz2 required docs.tar.bz2 /
xgts-main 1133 59dda9c318f4205167350b7ed993b5cd tbz2 required \
xgts-main.tar.bz2 /
pvdrivers-win 524477 37ea0c145f5b0d7a2740ecb69d21ed52 tbz2 required \
pvdrivers-win.tar.bz2 /
dom0fs 169875708 c1a86d705915eda16cca84cccffaca9f tbz2 required \
dom0fs.tar.bz2 /

If a driver disk is used, any tbz2 packages on it will also be installed to the target. However, a copy of the repository will be taken so that the drivers can be loaded at runtime; this copy is placed into memory. Therefore, if you are constructing a driver disk that also includes user-space tools, and if these result in a large repository, it is better to split it up into two repositories and require that people use the packages.site mechanism to install your add-ons. Alternatively, provide a post-install script to install them after the fact.

Table of Contents

C.1. Setting Control Domain Memory

When calculating the memory footprint of a Xen host there are two components that must be taken into consideration. First there is the memory consumed by the Xen hypervisor itself; then there is the memory consumed by the host's control domain. The control domain is a privileged VM that provides low-level services to other VMs, such as providing access to physical devices. It also runs the management tool stack.

If your control domain requires more allocated memory, this can be set using the Xen CLI.

Use the xe vm-memory-target-set command to set the amount of memory available to the host hypervisor.

The xe vm-memory-target-wait command can be used to check if the control domain has managed to attain the requested memory target specified at the last use of the xe vm-memory-target-set command. xe vm-memory-target-wait will not return until the memory target has been reached.

The following fields on a VM define how much memory will be allocated:

namedefaultdescription
memory-actual209715200The actual amount of memory current available for use by the VM
memory-target209715200The target amount of memory as set by using xe vm-memory-target-set
memory-static-max790102016The maximum possible physical memory
memory-dynamic-max790102016The desired maximum memory to be made available
memory-dynamic-min209715200The desired minimum memory to be made available
memory-static-min209715200The minimum possible physical memory

Dynamic memory values must be within the boundaries set by the static memory values. Additionally the target memory must fall in the range between the dynamic memory values.

Note

The amount of memory reported in XenCenter on the General tab in the Xen field may exceed the values set using this mechanism. This is because the amount reported includes the memory used by the control domain, the hypervisor itself, and the crash kernel. The amount of memory used by the hypervisor will be larger for hosts with more memory.

To find out how much host memory is actually available to be assigned to VMs, get the value of the memory-free field of the host, and then use the vm-compute-maximum-memory command to get the actual amount of free memory that can be allocated to the VM:

xe host-list uuid=<host_uuid> params=memory-free
xe vm-compute-maximum-memory vm=<vm_name> total=<host_memory_free_value>