root / docs / upgrade / upgrade-0.15.rst @ fba7aaf4
History | View | Annotate | Download (13.4 kB)
1 |
Upgrade to Synnefo v0.15 |
---|---|
2 |
^^^^^^^^^^^^^^^^^^^^^^^^ |
3 |
|
4 |
Prerequisites |
5 |
============== |
6 |
|
7 |
Before upgrading to v0.15 there are two steps that must be performed, relative |
8 |
with Cyclades networking service. |
9 |
|
10 |
Add unique name to the NICs of all Ganeti instances |
11 |
--------------------------------------------------- |
12 |
|
13 |
Since Ganeti 2.8, it is supported to give a name to NICs of Ganeti instances |
14 |
and refer to them with their name, and not only by their index. Synnefo v0.15 |
15 |
assigns a unique name to each NIC and refers to them by their unique name. |
16 |
Before upgrading to v0.15, Synnefo must assign names to all existing NICs. |
17 |
This can easily be performed with a helper script that is shipped with Synnefo |
18 |
v0.14.10: |
19 |
|
20 |
.. code-block:: console |
21 |
|
22 |
cyclades.host$ /usr/lib/synnefo/tools/add_unique_name_to_nics |
23 |
|
24 |
.. note:: If you are not upgrading from v0.14.10, you can find the migration |
25 |
script here XXX. |
26 |
|
27 |
|
28 |
Extend public networks to all Ganeti backends |
29 |
--------------------------------------------- |
30 |
|
31 |
Before v0.15, each public network of Cyclades existed in one of the Ganeti |
32 |
backends. In order to support dynamic addition and removal of public IPv4 |
33 |
address across VMs, each public network must exist in all Ganeti backends. |
34 |
|
35 |
If you are using more than one Ganeti backends, before upgrading to v0.15 you |
36 |
must ensure that the network configuration to all Ganeti backends is identical |
37 |
and appropriate to support all public networks of Cyclades. |
38 |
|
39 |
|
40 |
Upgrade Steps |
41 |
============= |
42 |
|
43 |
The upgrade to v0.15 consists in the following steps: |
44 |
|
45 |
1. Bring down services and backup databases. |
46 |
|
47 |
2. Upgrade packages, migrate the databases and configure settings. |
48 |
|
49 |
3. Create floating IP pools |
50 |
|
51 |
4. Register services and resources. |
52 |
|
53 |
5. Bring up all services. |
54 |
|
55 |
.. warning:: |
56 |
|
57 |
It is strongly suggested that you keep separate database backups |
58 |
for each service after the completion of each step. |
59 |
|
60 |
1. Bring web services down, backup databases |
61 |
============================================ |
62 |
|
63 |
1. All web services must be brought down so that the database maintains a |
64 |
predictable and consistent state during the migration process:: |
65 |
|
66 |
$ service gunicorn stop |
67 |
$ service snf-dispatcher stop |
68 |
$ service snf-ganeti-eventd stop |
69 |
|
70 |
2. Backup databases for recovery to a pre-migration state. |
71 |
|
72 |
3. Keep the database servers running during the migration process. |
73 |
|
74 |
|
75 |
2. Upgrade Synnefo and configure settings |
76 |
========================================= |
77 |
|
78 |
2.1 Install the new versions of packages |
79 |
---------------------------------------- |
80 |
|
81 |
:: |
82 |
|
83 |
astakos.host$ apt-get install \ |
84 |
python-objpool \ |
85 |
snf-common \ |
86 |
python-astakosclient \ |
87 |
snf-django-lib \ |
88 |
snf-webproject \ |
89 |
snf-branding \ |
90 |
snf-astakos-app |
91 |
|
92 |
cyclades.host$ apt-get install \ |
93 |
python-objpool \ |
94 |
snf-common \ |
95 |
python-astakosclient \ |
96 |
snf-django-lib \ |
97 |
snf-webproject \ |
98 |
snf-branding \ |
99 |
snf-pithos-backend \ |
100 |
snf-cyclades-app |
101 |
|
102 |
pithos.host$ apt-get install \ |
103 |
python-objpool \ |
104 |
snf-common \ |
105 |
python-astakosclient \ |
106 |
snf-django-lib \ |
107 |
snf-webproject \ |
108 |
snf-branding \ |
109 |
snf-pithos-backend \ |
110 |
snf-pithos-app \ |
111 |
snf-pithos-webclient |
112 |
|
113 |
ganeti.node$ apt-get install \ |
114 |
python-objpool \ |
115 |
snf-common \ |
116 |
snf-cyclades-gtools \ |
117 |
snf-pithos-backend |
118 |
|
119 |
.. note:: |
120 |
|
121 |
Make sure `snf-webproject' has the same version with snf-common |
122 |
|
123 |
.. note:: |
124 |
|
125 |
Installing the packages will cause services to start. Make sure you bring |
126 |
them down again (at least ``gunicorn``, ``snf-dispatcher``) |
127 |
|
128 |
2.2 Sync and migrate the database |
129 |
--------------------------------- |
130 |
|
131 |
.. note:: |
132 |
|
133 |
If you are asked about stale content types during the migration process, |
134 |
answer 'no' and let the migration finish. |
135 |
|
136 |
:: |
137 |
|
138 |
astakos-host$ snf-manage syncdb |
139 |
astakos-host$ snf-manage migrate |
140 |
|
141 |
cyclades-host$ snf-manage syncdb |
142 |
cyclades-host$ snf-manage migrate |
143 |
|
144 |
pithos-host$ pithos-migrate upgrade head |
145 |
|
146 |
2.3 Update configuration files |
147 |
------------------------------ |
148 |
|
149 |
The ``ASTAKOS_BASE_URL`` setting has been replaced (both in Cyclades and |
150 |
Pithos services) with the ``ASTAKOS_AUTH_URL`` setting. |
151 |
|
152 |
For Cyclades service we have to change the ``20-snf-cyclades-app-api.conf`` |
153 |
file, remove the ``ASTAKOS_BASE_URL`` setting and replace it with |
154 |
``ASTAKOS_AUTH_URL``. Typically it is sufficient to add ``/identity/v2.0`` |
155 |
at the end of base url to get the auth url. For example if base url had the |
156 |
value of 'https://accounts.example.synnefo.org/' then the ``ASTAKOS_AUTH_URL`` |
157 |
setting will have the value of |
158 |
'https://accounts.example.synnefo.org/identity/v2.0'. |
159 |
|
160 |
For Pithos service we have to change the ``20-snf-pithos-app-settings.conf`` |
161 |
file in the same way as above. |
162 |
|
163 |
2.4 Upgrade vncauthproxy and configure snf-cyclades-app |
164 |
------------------------------------------------------- |
165 |
|
166 |
Synnefo v0.15 adds support for snf-vncauthproxy >= 1.5 and drops support for |
167 |
older versions. You will have to upgrade snf-vncauthproxy to v1.5 and |
168 |
configure the authentication (users) file (``/var/lib/vncauthproxy/users``). |
169 |
|
170 |
In case you're upgrading from an older snf-vncauthproxy version or if it's the |
171 |
first time you're installing snf-vncauthproxy, you will need to add a |
172 |
vncauthproxy user (see below for more information on user management) and |
173 |
restart vncauthproxy daemon. |
174 |
|
175 |
To manage the authentication file, you can use the vncauthproxy-passwd tool, |
176 |
to easily add, update and delete users. |
177 |
|
178 |
To add a user: |
179 |
|
180 |
.. code-block:: console |
181 |
|
182 |
# vncauthproxy-passwd /var/lib/vncauthproxy/users synnefo |
183 |
|
184 |
You will be prompted for a password. |
185 |
|
186 |
You should also configure the new ``CYCLADES_VNCAUTHPROXY_OPTS`` setting in |
187 |
``snf-cyclades-app``, to provide the user and password configured for |
188 |
``Synnefo`` in the vncauthproxy authentication file and enable SSL support if |
189 |
snf-vncauthproxy is configured to run with SSL enabled for the control socket. |
190 |
|
191 |
.. warning:: The vncauthproxy daemon requires a restart for the changes in the |
192 |
authentication file to take effect. |
193 |
|
194 |
.. warning:: If you fail to provide snf-vncauthproxy with a valid |
195 |
authentication file, or in case the configuration of vncauthproxy and the |
196 |
vncauthproxy snf-cyclades-app settings don't match (ie not having SSL enabled |
197 |
on both), VNC console access will not be functional. |
198 |
|
199 |
Finally, snf-vncauthproxy-1.5 adds a dedicated user and group to be used by the |
200 |
vncauthproxy daemon. The Debian default file has changed accordingly (``CHUID`` |
201 |
option in ``/etc/default/vncauthproxy``). The Debian default file now also |
202 |
includes a ``DAEMON_OPTS`` variable which is used to pass any necessary / extra |
203 |
options to the vncauthproxy daemon. In case you're ugprading from an older |
204 |
version of vncauthproxy, you should make sure to 'merge' the new default file |
205 |
with the older one. |
206 |
|
207 |
Check the `documentation |
208 |
<http://www.synnefo.org/docs/snf-vncauthproxy/latest/index.html>`_ of |
209 |
snf-vncauthproxy for more information on upgrading to version 1.5. |
210 |
|
211 |
2.5 Stats configuration |
212 |
----------------------- |
213 |
|
214 |
snf-cyclades-gtools comes with a collectd plugin to collect CPU and network |
215 |
stats for Ganeti VMs and an example collectd configuration. snf-stats-app is a |
216 |
Django (snf-webproject) app that serves the VM stats graphsmm by reading the VM |
217 |
stats (from RRD files) and serves graphs. |
218 |
|
219 |
To enable / deploy VM stats collecting and snf-stats-app see the relevant |
220 |
documentation in the :ref:`admin guide <admin-guide-stats>`. |
221 |
|
222 |
If you were using collectd to collect VM stats on Debian squeeze and you are |
223 |
upgrading to Wheezy, you will need to upgrade your RRD files. Follow the |
224 |
instructions on the collectd v4-to-v5 migration `guide |
225 |
<https://collectd.org/wiki/index.php/V4_to_v5_migration_guide>`_. |
226 |
You will proabably just need to run the `migration script |
227 |
<https://collectd.org/wiki/index.php/V4_to_v5_migration_guide#Migration_script>`_ |
228 |
provided. |
229 |
|
230 |
If you were using a previous version of snf-stats-app, you should also make |
231 |
sure to set the ``STATS_BASE_URL`` setting in ``20-snf-stats-app-settings.conf`` |
232 |
to match your deployment and change the graph URL settings in |
233 |
``20-snf-cyclades-app-api.conf`` accordingly. |
234 |
|
235 |
v0.15 has also introduced the ``CYCLADES_STATS_SECRET_KEY`` and |
236 |
``STATS_SECRET_KEY`` settings. ``CYCLADES_STATS_SECRET_KEY`` in |
237 |
``20-snf-cyclades-app-api.conf`` is used by Cyclades to encrypt the instance id |
238 |
/ hostname in the URLs serving the VM stats. You should set it to a random |
239 |
value / string and make sure that it's the same as the ``STATS_SECRET_KEY`` |
240 |
setting (used to decrypt the instance hostname) in |
241 |
``20-snf-stats-settings.conf`` on your Stats host. |
242 |
|
243 |
3. Create floating IP pools |
244 |
=========================== |
245 |
|
246 |
Synnefo v0.15 introduces floating IPs, which are public IPv4 addresses that can |
247 |
dynamically be added/removed to/from VMs and are quotable via the |
248 |
'cyclades.floating_ip' resource. Connecting a VM to a public network is only |
249 |
allowed if the user has firstly created a floating IP from this network. |
250 |
|
251 |
Floating IPs are created from networks that are marked as Floating IP pools. |
252 |
Creation of floating IP pools is done with the `snf-manage network-create` |
253 |
command using the `--floating-ip-pool` option. |
254 |
|
255 |
Existing networks can be converted to floating IPs using `network-modify` |
256 |
command: |
257 |
|
258 |
.. code-block:: console |
259 |
|
260 |
snf-manage network-modify --floating-ip-pool=True <network_ID> |
261 |
|
262 |
Already allocated public IPv4 addresses are not automatically converted to |
263 |
floating IPs. Existing VMs can keep their IPv4 addresses which will be |
264 |
automatically be released when these VMs will be destroyed. In order to |
265 |
convert existing public IPs to floating IPs run the following command: |
266 |
|
267 |
.. code-block:: console |
268 |
|
269 |
cyclades.host$ /usr/lib/synnefo/tools/update_to_floating_ips |
270 |
|
271 |
or for just one network: |
272 |
|
273 |
.. code-block:: console |
274 |
|
275 |
cyclades.host$ /usr/lib/synnefo/tools/update_to_floating_ips --network-id=<network_ID> |
276 |
|
277 |
4. Register services and resources |
278 |
================================== |
279 |
|
280 |
4.1 Re-register service and resource definitions |
281 |
------------------------------------------------ |
282 |
|
283 |
You will need to register again all Synnefo components, updating the |
284 |
service and resource definitions. On the astakos node, run:: |
285 |
|
286 |
astakos-host$ snf-component-register |
287 |
|
288 |
This will detect that the Synnefo components are already registered and ask |
289 |
to re-register. Answer positively. You need to enter the base URL and the UI |
290 |
URL for each component, just like during the initial registration. |
291 |
|
292 |
.. note:: |
293 |
|
294 |
You can run ``snf-manage component-list -o name,ui_url`` to inspect the |
295 |
current registered UI URL. In the default installation, the base URL can |
296 |
be found by stripping ``/ui`` from the UI URL. |
297 |
|
298 |
The meaning of resources ``cyclades.cpu`` and ``cyclades.ram`` has changed: |
299 |
they now denote the number of CPUs and, respectively, RAM of *active* VMs |
300 |
rather than all VMs. To represent total CPUs and total RAM, as previously, |
301 |
new resources ``cyclades.total_cpu`` and ``cyclades.total_ram`` are |
302 |
introduced. We now also control the usage of floating IPs through resource |
303 |
``cyclades.floating_ip``. |
304 |
|
305 |
4.2 Tweek resource settings |
306 |
--------------------------- |
307 |
|
308 |
New resources (``cyclades.total_cpu``, ``cyclades.total_ram``, and |
309 |
``cyclades.floating_ip``) are registered with infinite default base quota. |
310 |
You will probably need to restrict them, especially |
311 |
``cyclades.floating_ip``. In order to change the default for all *future* |
312 |
users, for instance restricting floating IPs to 2, run:: |
313 |
|
314 |
astakos-host$ snf-manage resource-modify cyclades.floating_ip --default-quota 2 |
315 |
|
316 |
Note that this command does not affect *existing* users any more. They can |
317 |
still have infinite floating IPs. You can update base quota of existing |
318 |
users in bulk, possibly excluding some users, with:: |
319 |
|
320 |
astakos-host$ snf-manage user-modify --all --base-quota cyclades.floating_ip 2 --exclude uuid1,uuid2 |
321 |
|
322 |
.. note:: |
323 |
|
324 |
You can inspect base quota with ``snf-manage quota-list`` before applying |
325 |
any changes, for example:: |
326 |
|
327 |
# Get users with cyclades.vm base quota that differ from the default value |
328 |
astakos-host$ snf-manage quota-list --with-custom=True --filter-by "resource=cyclades.vm" |
329 |
|
330 |
# Get users with cyclades.vm base quota greater than 3 |
331 |
astakos-host$ snf-manage quota-list --filter-by "resource=cyclades.vm,base_quota>3" |
332 |
|
333 |
It is now possible to control whether a resource is visible for the users |
334 |
through the API or the UI. Note that the system always checks resource |
335 |
quota, regardless of their visibility. By default, ``cyclades.total_cpu``, |
336 |
``cyclades.total_ram`` and ``astakos.pending_app`` are not visible. You can |
337 |
change this behavior with:: |
338 |
|
339 |
astakos-host$ snf-manage resource-modify <resource> --api-visible=True (or --ui-visible=True) |
340 |
|
341 |
4.3 Update the Quotaholder |
342 |
-------------------------- |
343 |
|
344 |
To update quota for all new or modified Cyclades resources, bring up Astakos:: |
345 |
|
346 |
astakos-host$ service gunicorn start |
347 |
|
348 |
and run on the Cyclades node:: |
349 |
|
350 |
cyclades-host$ snf-manage reconcile-resources-cyclades --fix --force |
351 |
|
352 |
|
353 |
5. Bring all services up |
354 |
======================== |
355 |
|
356 |
After the upgrade is finished, we bring up all services: |
357 |
|
358 |
.. code-block:: console |
359 |
|
360 |
astakos.host # service gunicorn start |
361 |
cyclades.host # service gunicorn start |
362 |
pithos.host # service gunicorn start |
363 |
|
364 |
cyclades.host # service snf-dispatcher start |