| 1 |
[[ch-xen]] |
| 2 |
== Using Xen with DRBD |
| 3 |
|
| 4 |
indexterm:[Xen]This chapter outlines the use of DRBD as a Virtual |
| 5 |
Block Device (VBD) for virtualization environments using the Xen |
| 6 |
hypervisor. |
| 7 |
|
| 8 |
[[s-xen-primer]] |
| 9 |
=== Xen primer |
| 10 |
|
| 11 |
Xen is a virtualization framework originally developed at the |
| 12 |
University of Cambridge (UK), and later being maintained by XenSource, |
| 13 |
Inc. (now a part of Citrix). It is included in reasonably recent |
| 14 |
releases of most Linux distributions, such as Debian GNU/Linux (since |
| 15 |
version 4.0), SUSE Linux Enterprise Server (since release 10), Red Hat |
| 16 |
Enterprise Linux (since release 5), and many others. |
| 17 |
|
| 18 |
Xen uses paravirtualization — a virtualization method |
| 19 |
involving a high degree of cooperation between the virtualization host |
| 20 |
and guest virtual machines — with selected guest operating systems for |
| 21 |
improved performance in comparison to conventional virtualization |
| 22 |
solutions (which are typically based on hardware |
| 23 |
emulation). Xen also supports full hardware emulation |
| 24 |
on CPUs that support the appropriate virtualization extensions; in Xen |
| 25 |
parlance, this is known as HVM ( "hardware-assisted virtual machine"). |
| 26 |
|
| 27 |
NOTE: At the time of writing, CPU extensions supported by Xen for HVM |
| 28 |
are Intel's Virtualization Technology (VT, formerly codenamed |
| 29 |
"Vanderpool"), and AMD's Secure Virtual Machine (SVM, formerly known |
| 30 |
as "Pacifica"). |
| 31 |
|
| 32 |
Xen supports indexterm:[Xen, live migration]_live migration_, which refers to the |
| 33 |
capability of transferring a running guest operating system from one |
| 34 |
physical host to another, without interruption. |
| 35 |
|
| 36 |
When a DRBD resource is used as a replicated Virtual Block Device |
| 37 |
(VBD) for Xen, it serves to make the entire contents of a domU's |
| 38 |
virtual disk available on two servers, which can then be configured |
| 39 |
for automatic fail-over. That way, DRBD does not only provide |
| 40 |
redundancy for Linux servers (as in non-virtualized DRBD deployment |
| 41 |
scenarios), but also for any other operating system that can be |
| 42 |
virtualized under Xen — which, in essence, includes any operating |
| 43 |
system available on 32- or 64-bit Intel compatible architectures. |
| 44 |
|
| 45 |
[[s-xen-drbd-mod-params]] |
| 46 |
=== Setting DRBD module parameters for use with Xen |
| 47 |
|
| 48 |
For Xen Domain-0 kernels, it is recommended to load the DRBD module |
| 49 |
with the parameter `disable_sendpage` set to `1`. To do so, create (or open) the file |
| 50 |
`/etc/modprobe.d/drbd.conf` and enter the following line: |
| 51 |
|
| 52 |
[source,drbd] |
| 53 |
---------------------------- |
| 54 |
options drbd disable_sendpage=1 |
| 55 |
---------------------------- |
| 56 |
|
| 57 |
[[s-xen-create-resource]] |
| 58 |
=== Creating a DRBD resource suitable to act as a Xen VBD |
| 59 |
|
| 60 |
Configuring a DRBD resource that is to be used as a Virtual Block |
| 61 |
Device for Xen is fairly straightforward -- in essence, the typical |
| 62 |
configuration matches that of a DRBD resource being used for any other |
| 63 |
purpose. However, if you want to enable live migration |
| 64 |
for your guest instance, you need to enable indexterm:[Dual-Primary |
| 65 |
mode]<<s-dual-primary-mode,dual-primary mode>>for this resource: |
| 66 |
|
| 67 |
[source,drbd] |
| 68 |
---------------------------- |
| 69 |
resource <resource> { |
| 70 |
net { |
| 71 |
allow-two-primaries yes; |
| 72 |
... |
| 73 |
} |
| 74 |
... |
| 75 |
} |
| 76 |
---------------------------- |
| 77 |
|
| 78 |
Enabling dual-primary mode is necessary because Xen, before initiating |
| 79 |
live migration, checks for write access on all VBDs a resource is |
| 80 |
configured to use on both the source and the destination host for the |
| 81 |
migration. |
| 82 |
|
| 83 |
[[s-xen-configure-domu]] |
| 84 |
=== Using DRBD VBDs |
| 85 |
|
| 86 |
In order to use a DRBD resource as the virtual block device, you must |
| 87 |
add a line like the following to your Xen domU configuration: |
| 88 |
indexterm:[Xen] |
| 89 |
|
| 90 |
[source,drbd] |
| 91 |
---------------------------- |
| 92 |
disk = [ 'drbd:<resource>,xvda,w' ] |
| 93 |
---------------------------- |
| 94 |
|
| 95 |
This example configuration makes the DRBD resource named _resource_ |
| 96 |
available to the domU as `/dev/xvda` in read/write mode (`w`). |
| 97 |
|
| 98 |
Of course, you may use multiple DRBD resources with a single domU. In |
| 99 |
that case, simply add more entries like the one provided in the |
| 100 |
example to the `disk` option, separated by commas. |
| 101 |
|
| 102 |
|
| 103 |
NOTE: There are three sets of circumstances under which you cannot use |
| 104 |
this approach: |
| 105 |
|
| 106 |
* You are configuring a fully virtualized (HVM) domU. |
| 107 |
|
| 108 |
* You are installing your domU using a graphical installation utility, |
| 109 |
_and_ that graphical installer does not support the `drbd:` syntax. |
| 110 |
|
| 111 |
* You are configuring a domU without the `kernel`, `initrd`, and |
| 112 |
`extra` options, relying instead on `bootloader` and |
| 113 |
`bootloader_args` to use a Xen pseudo-bootloader, _and_ that |
| 114 |
pseudo-bootloader does not support the `drbd:` syntax. |
| 115 |
|
| 116 |
** pygrub+ (prior to Xen 3.3) and `domUloader.py` (shipped with Xen on |
| 117 |
SUSE Linux Enterprise Server 10) are two examples of |
| 118 |
pseudo-bootloaders that do not support the `drbd:` virtual block device |
| 119 |
configuration syntax. |
| 120 |
|
| 121 |
** `pygrub` from Xen 3.3 forward, and the `domUloader.py` version that |
| 122 |
ships with SLES 11 _do_ support this syntax. |
| 123 |
|
| 124 |
Under these circumstances, you must use the traditional `phy:` device |
| 125 |
syntax and the DRBD device name that is associated with your resource, |
| 126 |
not the resource name. That, however, requires that you manage DRBD |
| 127 |
state transitions outside Xen, which is a less flexible approach than |
| 128 |
that provided by the `drbd` resource type. |
| 129 |
|
| 130 |
[[s-manage-domu]] |
| 131 |
=== Starting, stopping, and migrating DRBD-backed domU's |
| 132 |
|
| 133 |
.Starting the domU |
| 134 |
Once you have configured your DRBD-backed domU, you may start it as |
| 135 |
you would any other domU: |
| 136 |
---------------------------- |
| 137 |
# xm create <domU> |
| 138 |
Using config file "/etc/xen/<domU>". |
| 139 |
Started domain <domU> |
| 140 |
---------------------------- |
| 141 |
|
| 142 |
In the process, the DRBD resource you configured as the VBD will be |
| 143 |
promoted to the primary role, and made accessible to Xen as expected. |
| 144 |
|
| 145 |
|
| 146 |
.Stopping the domU |
| 147 |
This is equally straightforward: |
| 148 |
---------------------------- |
| 149 |
# xm shutdown -w <domU> |
| 150 |
Domain <domU> terminated. |
| 151 |
---------------------------- |
| 152 |
|
| 153 |
Again, as you would expect, the DRBD resource is returned to the |
| 154 |
secondary role after the domU is successfully shut down. |
| 155 |
|
| 156 |
|
| 157 |
.Migrating the domU |
| 158 |
This, too, is done using the usual Xen tools: |
| 159 |
---------------------------- |
| 160 |
# xm migrate --live <domU> <destination-host> |
| 161 |
---------------------------- |
| 162 |
|
| 163 |
In this case, several administrative steps are automatically taken in |
| 164 |
rapid succession: |
| 165 |
* The resource is promoted to the primary role on _destination-host_. |
| 166 |
* Live migration of _domU_ is initiated on the local host. |
| 167 |
* When migration to the destination host has completed, the resource |
| 168 |
is demoted to the secondary role locally. |
| 169 |
|
| 170 |
The fact that both resources must briefly run in the primary role on |
| 171 |
both hosts is the reason for having to configure the resource in |
| 172 |
dual-primary mode in the first place. |
| 173 |
|
| 174 |
|
| 175 |
[[s-xen-internal]] |
| 176 |
=== Internals of DRBD/Xen integration |
| 177 |
|
| 178 |
Xen supports two Virtual Block Device types natively: |
| 179 |
|
| 180 |
.`phy` |
| 181 |
This device type is used to hand "physical" block devices, available |
| 182 |
in the host environment, off to a guest domU in an essentially |
| 183 |
transparent fashion. |
| 184 |
|
| 185 |
.`file` |
| 186 |
This device type is used to make file-based block device images |
| 187 |
available to the guest domU. It works by creating a loop block device |
| 188 |
from the original image file, and then handing that block device off |
| 189 |
to the domU in much the same fashion as the `phy` device type does. |
| 190 |
|
| 191 |
If a Virtual Block Device configured in the `disk` option of a domU |
| 192 |
configuration uses any prefix other than `phy:`, `file:`, or no prefix |
| 193 |
at all (in which case Xen defaults to using the `phy` device type), |
| 194 |
Xen expects to find a helper script named ``block-``_prefix_ in the Xen |
| 195 |
scripts directory, commonly `/etc/xen/scripts`. |
| 196 |
|
| 197 |
The DRBD distribution provides such a script for the `drbd` device |
| 198 |
type, named `/etc/xen/scripts/block-drbd`. This script handles the |
| 199 |
necessary DRBD resource state transitions as described earlier in this |
| 200 |
chapter. |
| 201 |
|
| 202 |
[[s-xen-pacemaker]] |
| 203 |
=== Integrating Xen with Pacemaker |
| 204 |
|
| 205 |
indexterm:[Xen]indexterm:[Heartbeat]In order to fully capitalize on |
| 206 |
the benefits provided by having a DRBD-backed Xen VBD's, it is |
| 207 |
recommended to have Pacemaker manage the associated domU's as |
| 208 |
Pacemaker resources. |
| 209 |
|
| 210 |
You may configure a Xen domU as a Pacemaker resource, and automate |
| 211 |
resource failover. To do so, use the Xen OCF resource agent. If you |
| 212 |
are using the `drbd` Xen device type described in this chapter, you |
| 213 |
will _not_ need to configure any separate `drbd` resource for use by |
| 214 |
the Xen cluster resource. Instead, the `block-drbd` helper script will |
| 215 |
do all the necessary resource transitions for you. |