4 .. highlight:: shell-example
6 This document details the steps needed to upgrade a cluster to newer versions
9 As a general rule the node daemons need to be restarted after each software
10 upgrade; if using the provided example init.d script, this means running the
11 following command on all nodes::
13 $ /etc/init.d/ganeti restart
19 Starting with Ganeti 2.0, upgrades between revisions (e.g. 2.1.0 to 2.1.1)
20 should not need manual intervention. As a safety measure, minor releases (e.g.
21 2.1.3 to 2.2.0) require the ``cfgupgrade`` command for changing the
22 configuration version. Below you find the steps necessary to upgrade between
25 To run commands on all nodes, the `distributed shell (dsh)
26 <http://www.netfort.gr.jp/~dancer/software/dsh.html.en>`_ can be used, e.g.
27 ``dsh -M -F 8 -f /var/lib/ganeti/ssconf_online_nodes gnt-cluster --version``.
29 #. Ensure no jobs are running (master node only)::
33 #. Pause the watcher for an hour (master node only)::
35 $ gnt-cluster watcher pause 1h
37 #. Stop all daemons on all nodes::
39 $ /etc/init.d/ganeti stop
41 #. Backup old configuration (master node only)::
43 $ tar czf /var/lib/ganeti-$(date +\%FT\%T).tar.gz -C /var/lib ganeti
45 #. Install new Ganeti version on all nodes
46 #. Run cfgupgrade on the master node::
48 $ /usr/lib/ganeti/tools/cfgupgrade --verbose --dry-run
49 $ /usr/lib/ganeti/tools/cfgupgrade --verbose
51 (``cfgupgrade`` supports a number of parameters, run it with
52 ``--help`` for more information)
54 #. Upgrade the directory permissions on all nodes::
56 $ /usr/lib/ganeti/ensure-dirs --full-run
58 #. Create the (missing) required users and make users part of the required
61 $ /usr/lib/ganeti/tools/users-setup
63 #. Restart daemons on all nodes::
65 $ /etc/init.d/ganeti restart
67 #. Re-distribute configuration (master node only)::
69 $ gnt-cluster redist-conf
71 #. If you use file storage, check that the ``/etc/ganeti/file-storage-paths``
72 is correct on all nodes. For security reasons it's not copied
73 automatically, but it can be copied manually via::
75 $ gnt-cluster copyfile /etc/ganeti/file-storage-paths
77 #. Restart daemons again on all nodes::
79 $ /etc/init.d/ganeti restart
81 #. Enable the watcher again (master node only)::
83 $ gnt-cluster watcher continue
85 #. Verify cluster (master node only)::
92 For going back between revisions (e.g. 2.1.1 to 2.1.0) no manual
93 intervention is required, as for upgrades.
95 Starting from version 2.8, ``cfgupgrade`` supports ``--downgrade``
96 option to bring the configuration back to the previous stable version.
97 This is useful if you upgrade Ganeti and after some time you run into
98 problems with the new version. You can downgrade the configuration
99 without losing the changes made since the upgrade. Any feature not
100 supported by the old version will be removed from the configuration, of
101 course, but you get a warning about it. If there is any new feature and
102 you haven't changed from its default value, you don't have to worry
103 about it, as it will get the same value whenever you'll upgrade again.
105 The procedure is similar to upgrading, but please notice that you have to
106 revert the configuration **before** installing the old version.
108 #. Ensure no jobs are running (master node only)::
112 #. Pause the watcher for an hour (master node only)::
114 $ gnt-cluster watcher pause 1h
116 #. Stop all daemons on all nodes::
118 $ /etc/init.d/ganeti stop
120 #. Backup old configuration (master node only)::
122 $ tar czf /var/lib/ganeti-$(date +\%FT\%T).tar.gz -C /var/lib ganeti
124 #. Run cfgupgrade on the master node::
126 $ /usr/lib/ganeti/tools/cfgupgrade --verbose --downgrade --dry-run
127 $ /usr/lib/ganeti/tools/cfgupgrade --verbose --downgrade
129 You may want to copy all the messages about features that have been
130 removed during the downgrade, in case you want to restore them when
133 #. Install the old Ganeti version on all nodes
134 #. Restart daemons on all nodes::
136 $ /etc/init.d/ganeti restart
138 #. Re-distribute configuration (master node only)::
140 $ gnt-cluster redist-conf
142 #. Restart daemons again on all nodes::
144 $ /etc/init.d/ganeti restart
146 #. Enable the watcher again (master node only)::
148 $ gnt-cluster watcher continue
150 #. Verify cluster (master node only)::
161 No changes needed except restarting the daemon; but rollback to 2.0.3 might
162 require configuration editing.
164 If you're using Xen-HVM instances, please double-check the network
165 configuration (``nic_type`` parameter) as the defaults might have changed:
166 2.0.4 adds any missing configuration items and depending on the version of the
167 software the cluster has been installed with, some new keys might have been
173 Between 2.0.1 and 2.0.2 there have been some changes in the handling of block
174 devices, which can cause some issues. 2.0.3 was then released which adds two
175 new options/commands to fix this issue.
177 If you use DRBD-type instances and see problems in instance start or
178 activate-disks with messages from DRBD about "lower device too small" or
179 similar, it is recoomended to:
181 #. Run ``gnt-instance activate-disks --ignore-size $instance`` for each
182 of the affected instances
183 #. Then run ``gnt-cluster repair-disk-sizes`` which will check that
184 instances have the correct disk sizes
191 - Ganeti 1.2.7 is currently installed
192 - All instances have been migrated from DRBD 0.7 to DRBD 8.x (i.e. no
193 ``remote_raid1`` disk template)
194 - Upgrade to Ganeti 2.0.0~rc2 or later (~rc1 and earlier don't have the needed
197 In the below steps, replace :file:`/var/lib` with ``$libdir`` if Ganeti was not
198 installed with this prefix (e.g. :file:`/usr/local/var`). Same for
201 Execution (all steps are required in the order given):
203 #. Make a backup of the current configuration, for safety::
205 $ cp -a /var/lib/ganeti /var/lib/ganeti-1.2.backup
207 #. Stop all instances::
209 $ gnt-instance stop --all
211 #. Make sure no DRBD device are in use, the following command should show no
214 $ gnt-cluster command grep cs: /proc/drbd | grep -v cs:Unconf
216 #. Stop the node daemons and rapi daemon on all nodes (note: should be logged
217 in not via the cluster name, but the master node name, as the command below
218 will remove the cluster ip from the master node)::
220 $ gnt-cluster command /etc/init.d/ganeti stop
222 #. Install the new software on all nodes, either from packaging (if available)
223 or from sources; the master daemon will not start but give error messages
224 about wrong configuration file, which is normal
225 #. Upgrade the configuration file::
227 $ /usr/lib/ganeti/tools/cfgupgrade12 -v --dry-run
228 $ /usr/lib/ganeti/tools/cfgupgrade12 -v
230 #. Make sure ``ganeti-noded`` is running on all nodes (and start it if
232 #. Start the master daemon::
236 #. Check that a simple node-list works::
240 #. Redistribute updated configuration to all nodes::
242 $ gnt-cluster redist-conf
243 $ gnt-cluster copyfile /var/lib/ganeti/known_hosts
245 #. Optional: if needed, install RAPI-specific certificates under
246 :file:`/var/lib/ganeti/rapi.pem` and run::
248 $ gnt-cluster copyfile /var/lib/ganeti/rapi.pem
250 #. Run a cluster verify, this should show no problems::
254 #. Remove some obsolete files::
256 $ gnt-cluster command rm /var/lib/ganeti/ssconf_node_pass
257 $ gnt-cluster command rm /var/lib/ganeti/ssconf_hypervisor
259 #. Update the xen pvm (if this was a pvm cluster) setting for 1.2
262 $ gnt-cluster modify -H xen-pvm:root_path=/dev/sda
264 #. Depending on your setup, you might also want to reset the initrd parameter::
266 $ gnt-cluster modify -H xen-pvm:initrd_path=/boot/initrd-2.6-xenU
268 #. Reset the instance autobalance setting to default::
270 $ for i in $(gnt-instance list -o name --no-headers); do \
271 gnt-instance modify -B auto_balance=default $i; \
274 #. Optional: start the RAPI demon::
278 #. Restart instances::
280 $ gnt-instance start --force-multiple --all
282 At this point, ``gnt-cluster verify`` should show no errors and the migration
288 1.2.4 to any other higher 1.2 version
289 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
291 No changes needed. Rollback will usually require manual edit of the
297 No changes needed. Note that going back from 1.2.4 to 1.2.3 will require manual
298 edit of the configuration file (since we added some HVM-related new
304 No changes needed. Note that the drbd7-to-8 upgrade tool does a disk format
305 change for the DRBD metadata, so in theory this might be **risky**. It is
306 advised to have (good) backups before doing the upgrade.
316 No changes needed. Only some bugfixes and new additions that don't affect
319 1.2.0 beta 3 to 1.2.0
320 ~~~~~~~~~~~~~~~~~~~~~
324 1.2.0 beta 2 to beta 3
325 ~~~~~~~~~~~~~~~~~~~~~~
327 No changes needed. A new version of the debian-etch-instance OS (0.3) has been
328 released, but upgrading it is not required.
330 1.2.0 beta 1 to beta 2
331 ~~~~~~~~~~~~~~~~~~~~~~
333 Beta 2 switched the config file format to JSON. Steps to upgrade:
335 #. Stop the daemons (``/etc/init.d/ganeti stop``) on all nodes
336 #. Disable the cron job (default is :file:`/etc/cron.d/ganeti`)
337 #. Install the new version
338 #. Make a backup copy of the config file
339 #. Upgrade the config file using the following command::
341 $ /usr/share/ganeti/cfgupgrade --verbose /var/lib/ganeti/config.data
343 #. Start the daemons and run ``gnt-cluster info``, ``gnt-node list`` and
344 ``gnt-instance list`` to check if the upgrade process finished successfully
346 The OS definition also need to be upgraded. There is a new version of the
347 debian-etch-instance OS (0.2) that goes along with beta 2.
349 .. vim: set textwidth=72 :