root / docs / installing.sgml @ 1005b0c1
History | View | Annotate | Download (14.2 kB)
1 |
<!DOCTYPE article PUBLIC "-//OASIS//DTD DocBook V4.2//EN" [ |
---|---|
2 |
]> |
3 |
<article class="specification"> |
4 |
<articleinfo> |
5 |
<title>Ganeti node/cluster installation tutorial</title> |
6 |
</articleinfo> |
7 |
<para>Documents Ganeti version 1.2</para> |
8 |
|
9 |
<sect1> |
10 |
<title>Introduction</title> |
11 |
|
12 |
<para> |
13 |
Ganeti is a cluster virtualization management system. This |
14 |
document explains how to bootstrap a Ganeti node and create a |
15 |
running cluster. You need to repeat most of the steps in this |
16 |
document for every node you want to install, but of course we |
17 |
recommend creating some semi-automatic procedure if you plan to |
18 |
deploy Ganeti on a medium/large scale. |
19 |
</para> |
20 |
|
21 |
<para>This document is divided into two main sections: |
22 |
|
23 |
<itemizedlist> |
24 |
<listitem> |
25 |
<simpara>Installation of the core system and base |
26 |
components</simpara> |
27 |
</listitem> |
28 |
<listitem> |
29 |
<simpara>Configuration of the environment for |
30 |
Ganeti</simpara> |
31 |
</listitem> |
32 |
</itemizedlist> |
33 |
|
34 |
Each of these is divided into sub-sections. While a full Ganeti |
35 |
system will need all of the steps specified, some are not strictly |
36 |
required for every environment. Which ones they are, and why, is |
37 |
specified in the corresponding sections. |
38 |
</para> |
39 |
|
40 |
<para> |
41 |
While Ganeti itself is distribution-agnostic most of the |
42 |
examples in this document will be targeted at Debian or |
43 |
Debian-derived distributions. You are expected to be familiar |
44 |
with your distribution, its package management system, and Xen |
45 |
before trying to use Ganeti. |
46 |
</para> |
47 |
|
48 |
<para> |
49 |
A basic Ganeti terminology glossary is provided in the |
50 |
introductory section of the "admin guide". Please refer to that |
51 |
if you are uncertain about the terms we are using. |
52 |
</para> |
53 |
|
54 |
</sect1> |
55 |
|
56 |
<sect1> |
57 |
<title>Installing the system and base components</title> |
58 |
|
59 |
<sect2> |
60 |
<title>Installing the base system</title> |
61 |
|
62 |
<para><emphasis role="strong">Mandatory.</emphasis></para> |
63 |
|
64 |
<para> |
65 |
Please install your operating system as you would normally |
66 |
do. The only requirement you need to be aware of at this stage |
67 |
is to partition leaving enough space for a big LVM volume |
68 |
group which will then host your instance file systems. You can |
69 |
even create the volume group at installation time, of course: |
70 |
the default volume group name Ganeti 1.2 uses is "xenvg" but |
71 |
you may name it differently should you wish to, as long as the |
72 |
name is the same for all the nodes in the cluster. |
73 |
</para> |
74 |
|
75 |
</sect2> |
76 |
|
77 |
<sect2> |
78 |
<title>Installing Xen</title> |
79 |
|
80 |
<para> |
81 |
<emphasis role="strong">Mandatory:</emphasis> While Ganeti is |
82 |
developed with the ability to modularly run on different |
83 |
virtualization environments in mind the only one currently |
84 |
useable on a live system is <ulink |
85 |
url="http://xen.xensource.com/">Xen</ulink>. |
86 |
</para> |
87 |
|
88 |
<para> |
89 |
Please follow your distribution's recommended way to install |
90 |
and set up Xen, or install Xen from the upstream source, if |
91 |
you wish, following their manual. |
92 |
</para> |
93 |
|
94 |
<para> |
95 |
For example under Debian 4.0 or 3.1+backports you can install |
96 |
the relevant xen-linux-system package, which will pull in both |
97 |
the hypervisor and the relevant kernel. On Ubuntu (from Gutsy |
98 |
on) the package is called ubuntu-xen-server. |
99 |
</para> |
100 |
|
101 |
<para> |
102 |
After installing Xen you need to reboot into your xenified |
103 |
dom0 system. Again on some distributions this might involve |
104 |
configuring GRUB appropriately. |
105 |
</para> |
106 |
|
107 |
</sect2> |
108 |
|
109 |
<sect2> |
110 |
<title>Installing DRBD</title> |
111 |
|
112 |
<para> |
113 |
Recommended: <ulink url="http://www.drbd.org/">DRBD</ulink> |
114 |
is required if you want to use the high availability (HA) |
115 |
features of Ganeti, but optional if you don't require HA or |
116 |
only run Ganeti on single-node clusters. You can upgrade a |
117 |
non-HA cluster to an HA one later, but you might need to |
118 |
export and reimport all your instances to take advantage of |
119 |
the new features. |
120 |
</para> |
121 |
|
122 |
<para> |
123 |
Now the bad news: unless your distribution already provides it |
124 |
installing DRBD might involve recompiling your kernel or |
125 |
anyway fiddling with it. Hopefully at least the xenified |
126 |
kernel source to start from will be provided. |
127 |
</para> |
128 |
|
129 |
<para> |
130 |
Under Debian you can just install the drbd0.7-module-source |
131 |
and drbd0.7-utils packages, and your kernel source, and then |
132 |
run module-assistant to compile the drbd0.7 module. The |
133 |
following commands should do it: |
134 |
</para> |
135 |
|
136 |
<screen> |
137 |
m-a update |
138 |
m-a a-i drbd0.7 |
139 |
</screen> |
140 |
|
141 |
<para> |
142 |
The good news is that you don't need to configure DRBD at all. |
143 |
Ganeti will do it for you for every instance you set up. If |
144 |
you have the DRBD utils installed and the module in your |
145 |
kernel you're fine. Please check that your system is |
146 |
configured to load the module at every boot. |
147 |
</para> |
148 |
|
149 |
</sect2> |
150 |
|
151 |
<sect2> |
152 |
<title>Other required software</title> |
153 |
|
154 |
<para>Besides Xen and DRBD, you will need to install the |
155 |
following:</para> |
156 |
|
157 |
<itemizedlist> |
158 |
<listitem> |
159 |
<simpara><ulink url="http://sourceware.org/lvm2/">LVM |
160 |
version 2</ulink></simpara> |
161 |
</listitem> |
162 |
<listitem> |
163 |
<simpara><ulink |
164 |
url="http://www.openssl.org/">OpenSSL</ulink></simpara> |
165 |
</listitem> |
166 |
<listitem> |
167 |
<simpara><ulink |
168 |
url="http://www.openssh.com/portable.html">OpenSSH</ulink></simpara> |
169 |
</listitem> |
170 |
<listitem> |
171 |
<simpara><ulink url="http://bridge.sourceforge.net/">Bridge |
172 |
utilities</ulink></simpara> |
173 |
</listitem> |
174 |
<listitem> |
175 |
<simpara><ulink |
176 |
url="http://fping.sourceforge.net/">fping</ulink></simpara> |
177 |
</listitem> |
178 |
<listitem> |
179 |
<simpara><ulink |
180 |
url="http://developer.osdl.org/dev/iproute2">iproute2</ulink></simpara> |
181 |
</listitem> |
182 |
<listitem> |
183 |
<simpara><ulink |
184 |
url="ftp://ftp.inr.ac.ru/ip-routing/iputils-current.tar.gz">arping</ulink> |
185 |
(part of iputils package)</simpara> |
186 |
</listitem> |
187 |
<listitem> |
188 |
<simpara><ulink |
189 |
url="http://www.kernel.org/pub/linux/utils/raid/mdadm/">mdadm</ulink> |
190 |
(Linux Software Raid tools)</simpara> |
191 |
</listitem> |
192 |
<listitem> |
193 |
<simpara><ulink url="http://www.python.org">Python 2.4</ulink></simpara> |
194 |
</listitem> |
195 |
<listitem> |
196 |
<simpara><ulink url="http://twistedmatrix.com/">Python |
197 |
Twisted library</ulink> - the core library is |
198 |
enough</simpara> |
199 |
</listitem> |
200 |
<listitem> |
201 |
<simpara><ulink |
202 |
url="http://pyopenssl.sourceforge.net/">Python OpenSSL |
203 |
bindings</ulink></simpara> |
204 |
</listitem> |
205 |
</itemizedlist> |
206 |
|
207 |
<para>These programs are supplied as part of most Linux |
208 |
distributions, so usually they can be installed via apt or |
209 |
similar methods. Also many of them will already be installed on |
210 |
a standard machine. On Debian Etch you can use this command line |
211 |
to install all of them:</para> |
212 |
|
213 |
<screen> |
214 |
# apt-get install lvm2 ssh bridge-utils iproute iputils-arping \ |
215 |
> fping python2.4 python-twisted-core python-pyopenssl openssl |
216 |
</screen> |
217 |
</sect2> |
218 |
|
219 |
</sect1> |
220 |
|
221 |
|
222 |
<sect1> |
223 |
<title>Setting up the environment for Ganeti</title> |
224 |
|
225 |
<sect2> |
226 |
<title>Configuring the network</title> |
227 |
|
228 |
<para>Ganeti relies on Xen running in "bridge mode", which means the |
229 |
instances network interfaces will be attached to a software bridge |
230 |
running in dom0. Xen by default creates such a bridge at startup, but |
231 |
your distribution might have a different way to do things. |
232 |
</para> |
233 |
|
234 |
<para> |
235 |
In Debian, in order to enable the default Xen behaviour, you |
236 |
have to edit <filename>/etc/xen/xend-config.sxp</filename> and |
237 |
replace <computeroutput>(network-script |
238 |
network-dummy)</computeroutput> with |
239 |
<computeroutput>(network-script |
240 |
network-bridge)</computeroutput>. The recommended Debian way to |
241 |
configure things, though, is to edit your |
242 |
<filename>/etc/network/interfaces</filename> file and substitute |
243 |
your normal ethernet stanza with something like:</para> |
244 |
|
245 |
<screen> |
246 |
auto br0 |
247 |
iface br0 inet static |
248 |
address <replaceable>YOUR_IP_ADDRESS</replaceable> |
249 |
netmask <replaceable>YOUR_NETMASK</replaceable> |
250 |
network <replaceable>YOUR_NETWORK</replaceable> |
251 |
broadcast <replaceable>YOUR_BROADCAST_ADDRESS</replaceable> |
252 |
gateway <replaceable>YOUR_GATEWAY</replaceable> |
253 |
bridge_ports <replaceable>eth0</replaceable> |
254 |
bridge_stp off |
255 |
bridge_fd 0 |
256 |
</screen> |
257 |
|
258 |
<para> |
259 |
Beware that the default name Ganeti uses is |
260 |
<hardware>xen-br0</hardware> (which was used in Xen 2.0) |
261 |
while Xen 3.0 uses <hardware>xenbr0</hardware> by |
262 |
default. The default bridge your cluster will use for new |
263 |
instances can be specified at cluster initialization time. |
264 |
</para> |
265 |
|
266 |
</sect2> |
267 |
|
268 |
<sect2> |
269 |
<title>Configuring LVM</title> |
270 |
|
271 |
<para> |
272 |
If you haven't configured your LVM volume group at install |
273 |
time you need to do it before trying to initialize the Ganeti |
274 |
cluster. This is done by formatting the devices/partitions you |
275 |
want to use for it and then adding them to the relevant volume |
276 |
group: |
277 |
</para> |
278 |
|
279 |
<screen> |
280 |
pvcreate /dev/sda4 |
281 |
pvcreate /dev/sdb |
282 |
pvcreate /dev/sdc1 |
283 |
vgcreate xenvg /dev/sda4 /dev/sdb /dev/sdc1 |
284 |
</screen> |
285 |
|
286 |
<para> |
287 |
If you want to add a device later you can do so with the |
288 |
<citerefentry><refentrytitle>vgextend</refentrytitle> |
289 |
<manvolnum>8</manvolnum></citerefentry> command: |
290 |
</para> |
291 |
|
292 |
<screen> |
293 |
pvcreate /dev/sdd |
294 |
vgextend xenvg /dev/sdd |
295 |
</screen> |
296 |
|
297 |
<para> |
298 |
As said before you may choose a different name for the volume group, |
299 |
as long as you stick to the same name on all the nodes of a cluster. |
300 |
</para> |
301 |
</sect2> |
302 |
|
303 |
<sect2> |
304 |
<title>Installing Ganeti</title> |
305 |
|
306 |
<para> |
307 |
It's now time to install the Ganeti software itself. You can |
308 |
do it from source, with the usual steps (note that the |
309 |
<option>localstatedir</option> options must be set to |
310 |
<filename class="directory">/var</filename>): |
311 |
</para> |
312 |
|
313 |
<screen> |
314 |
./configure --localstatedir=/var |
315 |
make |
316 |
make install |
317 |
mkdir /srv/ganeti/ /srv/ganeti/os /srv/ganeti/export |
318 |
</screen> |
319 |
|
320 |
<para> |
321 |
You also need to copy from the source archive the file |
322 |
<filename>docs/examples/ganeti.initd</filename> to |
323 |
<filename>/etc/init.d/ganeti</filename> and register it into |
324 |
your distribution's startup scripts, for example in Debian: |
325 |
</para> |
326 |
<screen>update-rc.d ganeti defaults 20 80</screen> |
327 |
|
328 |
</sect2> |
329 |
|
330 |
<sect2> |
331 |
<title>Installing the Operating System support packages</title> |
332 |
|
333 |
<para> |
334 |
Another important component for Ganeti are the OS support |
335 |
packages, which let different operating systems be used as |
336 |
instances. You can grab a simple package that allows |
337 |
installing Debian Etch instances on the project web site |
338 |
(after download, untar it and follow the instructions in the |
339 |
<filename>README</filename> file). |
340 |
</para> |
341 |
|
342 |
<para> |
343 |
Alternatively, you can create your own OS definitions, see |
344 |
<citerefentry> |
345 |
<refentrytitle>ganeti-os-interface</refentrytitle> |
346 |
<manvolnum>8</manvolnum> |
347 |
</citerefentry>. |
348 |
</para> |
349 |
|
350 |
</sect2> |
351 |
|
352 |
<sect2> |
353 |
<title>Initializing the cluster</title> |
354 |
|
355 |
<para><emphasis role="strong">Mandatory:</emphasis> only on one |
356 |
node per cluster.</para> |
357 |
|
358 |
|
359 |
<para>The last step is to initialize the cluster. After you've repeated |
360 |
the above process or some semi-automatic form of it on all of your |
361 |
nodes choose one as the master, and execute: |
362 |
</para> |
363 |
|
364 |
<screen> |
365 |
gnt-cluster init <replaceable>CLUSTERNAME</replaceable> |
366 |
</screen> |
367 |
|
368 |
<para> |
369 |
If the node's network interface which will be used for access |
370 |
from outside the cluster is not named |
371 |
<hardware>xen-br0</hardware>, you need to use the |
372 |
<option>--master-netdev=<replaceable>IFNAME</replaceable></option> |
373 |
option, replacing <replaceable>IFNAME</replaceable> with the |
374 |
correct one for your case (e.g. <hardware>xenbr0</hardware>, |
375 |
<hardware>eth0</hardware>, etc.). Usually this will be the |
376 |
same as the default bridge name (see below). |
377 |
</para> |
378 |
|
379 |
<para> |
380 |
Other options you can pass to <command>gnt-cluster |
381 |
init</command> include the default bridge name |
382 |
(<option>-b</option>), the cluster-wide name for the volume |
383 |
group (<option>-g</option>) and the secondary ip address for |
384 |
the initial node should you wish to keep the data replication |
385 |
network separate. Invoke it with <option>--help</option> to |
386 |
see all the possibilities. |
387 |
</para> |
388 |
|
389 |
<para> |
390 |
Note that the cluster name must exist in DNS. You must choose |
391 |
a name different from any of the nodes names for a multi-node |
392 |
cluster. In general the best choice is to have a unique name |
393 |
for a cluster, even if it consists of only one machine, as you |
394 |
will be able to expand it later without any problem. |
395 |
</para> |
396 |
</sect2> |
397 |
|
398 |
<sect2> |
399 |
<title>Joining the nodes to the cluster.</title> |
400 |
|
401 |
<para> |
402 |
<emphasis role="strong">Mandatory:</emphasis> for all the |
403 |
other nodes. |
404 |
</para> |
405 |
|
406 |
<para> |
407 |
If you have already initialized your cluster you need to join the other |
408 |
nodes to it. You can do so by executing the following command on the |
409 |
master node: |
410 |
<screen> |
411 |
gnt-node add <replaceable>NODENAME</replaceable> |
412 |
</screen> |
413 |
|
414 |
The only option is <option>-s</option>, which sets the node's |
415 |
secondary ip address for replication purposes, if you are |
416 |
using a separate replication network. |
417 |
</para> |
418 |
</sect2> |
419 |
|
420 |
</sect1> |
421 |
|
422 |
<sect1> |
423 |
<title>This is it!</title> |
424 |
|
425 |
<para> |
426 |
Now you can follow the admin guide to use your new Ganeti |
427 |
cluster. |
428 |
</para> |
429 |
|
430 |
</sect1> |
431 |
|
432 |
|
433 |
</article> |