1 Ganeti administrator's guide
2 ============================
4 Documents Ganeti version |version|
8 .. highlight:: shell-example
13 Ganeti is a virtualization cluster management software. You are expected
14 to be a system administrator familiar with your Linux distribution and
15 the Xen or KVM virtualization environments before using it.
17 The various components of Ganeti all have man pages and interactive
18 help. This manual though will help you getting familiar with the system
19 by explaining the most common operations, grouped by related use.
21 After a terminology glossary and a section on the prerequisites needed
22 to use this manual, the rest of this document is divided in sections
23 for the different targets that a command affects: instance, nodes, etc.
25 .. _terminology-label:
30 This section provides a small introduction to Ganeti terminology, which
31 might be useful when reading the rest of the document.
36 A set of machines (nodes) that cooperate to offer a coherent, highly
37 available virtualization service under a single administration domain.
42 A physical machine which is member of a cluster. Nodes are the basic
43 cluster infrastructure, and they don't need to be fault tolerant in
44 order to achieve high availability for instances.
46 Node can be added and removed (if they host no instances) at will from
47 the cluster. In a HA cluster and only with HA instances, the loss of any
48 single node will not cause disk data loss for any instance; of course,
49 a node crash will cause the crash of the its primary instances.
51 A node belonging to a cluster can be in one of the following roles at a
54 - *master* node, which is the node from which the cluster is controlled
55 - *master candidate* node, only nodes in this role have the full cluster
56 configuration and knowledge, and only master candidates can become the
58 - *regular* node, which is the state in which most nodes will be on
59 bigger clusters (>20 nodes)
60 - *drained* node, nodes in this state are functioning normally but the
61 cannot receive new instances; the intention is that nodes in this role
62 have some issue and they are being evacuated for hardware repairs
63 - *offline* node, in which there is a record in the cluster
64 configuration about the node, but the daemons on the master node will
65 not talk to this node; any instances declared as having an offline
66 node as either primary or secondary will be flagged as an error in the
67 cluster verify operation
69 Depending on the role, each node will run a set of daemons:
71 - the :command:`ganeti-noded` daemon, which control the manipulation of
72 this node's hardware resources; it runs on all nodes which are in a
74 - the :command:`ganeti-confd` daemon (Ganeti 2.1+) which runs on all
75 nodes, but is only functional on master candidate nodes; this daemon
76 can be disabled at configuration time if you don't need its
78 - the :command:`ganeti-rapi` daemon which runs on the master node and
79 offers an HTTP-based API for the cluster
80 - the :command:`ganeti-masterd` daemon which runs on the master node and
81 allows control of the cluster
83 Beside the node role, there are other node flags that influence its
86 - the *master_capable* flag denotes whether the node can ever become a
87 master candidate; setting this to 'no' means that auto-promotion will
88 never make this node a master candidate; this flag can be useful for a
89 remote node that only runs local instances, and having it become a
90 master is impractical due to networking or other constraints
91 - the *vm_capable* flag denotes whether the node can host instances or
92 not; for example, one might use a non-vm_capable node just as a master
93 candidate, for configuration backups; setting this flag to no
94 disallows placement of instances of this node, deactivates hypervisor
95 and related checks on it (e.g. bridge checks, LVM check, etc.), and
96 removes it from cluster capacity computations
102 A virtual machine which runs on a cluster. It can be a fault tolerant,
103 highly available entity.
105 An instance has various parameters, which are classified in three
106 categories: hypervisor related-parameters (called ``hvparams``), general
107 parameters (called ``beparams``) and per network-card parameters (called
108 ``nicparams``). All these parameters can be modified either at instance
109 level or via defaults at cluster level.
114 The are multiple options for the storage provided to an instance; while
115 the instance sees the same virtual drive in all cases, the node-level
116 configuration varies between them.
118 There are five disk templates you can choose from:
121 The instance has no disks. Only used for special purpose operating
122 systems or for testing.
125 The instance will use plain files as backend for its disks. No
126 redundancy is provided, and this is somewhat more difficult to
127 configure for high performance.
130 The instance will use LVM devices as backend for its disks. No
131 redundancy is provided.
134 .. note:: This is only valid for multi-node clusters using DRBD 8.0+
136 A mirror is set between the local node and a remote one, which must be
137 specified with the second value of the --node option. Use this option
138 to obtain a highly available instance that can be failed over to a
139 remote node should the primary one fail.
141 .. note:: Ganeti does not support DRBD stacked devices:
142 DRBD stacked setup is not fully symmetric and as such it is
143 not working with live migration.
146 The instance will use Volumes inside a RADOS cluster as backend for its
147 disks. It will access them using the RADOS block device (RBD).
152 A framework for using external (user-provided) scripts to compute the
153 placement of instances on the cluster nodes. This eliminates the need to
154 manually specify nodes in instance add, instance moves, node evacuate,
157 In order for Ganeti to be able to use these scripts, they must be place
158 in the iallocator directory (usually ``lib/ganeti/iallocators`` under
159 the installation prefix, e.g. ``/usr/local``).
161 “Primary” and “secondary” concepts
162 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
164 An instance has a primary and depending on the disk configuration, might
165 also have a secondary node. The instance always runs on the primary node
166 and only uses its secondary node for disk replication.
168 Similarly, the term of primary and secondary instances when talking
169 about a node refers to the set of instances having the given node as
170 primary, respectively secondary.
175 Tags are short strings that can be attached to either to cluster itself,
176 or to nodes or instances. They are useful as a very simplistic
177 information store for helping with cluster administration, for example
178 by attaching owner information to each instance after it's created::
180 $ gnt-instance add … %instance1%
181 $ gnt-instance add-tags %instance1% %owner:user2%
183 And then by listing each instance and its tags, this information could
184 be used for contacting the users of each instance.
189 While not directly visible by an end-user, it's useful to know that a
190 basic cluster operation (e.g. starting an instance) is represented
191 internally by Ganeti as an *OpCode* (abbreviation from operation
192 code). These OpCodes are executed as part of a *Job*. The OpCodes in a
193 single Job are processed serially by Ganeti, but different Jobs will be
194 processed (depending on resource availability) in parallel. They will
195 not be executed in the submission order, but depending on resource
196 availability, locks and (starting with Ganeti 2.3) priority. An earlier
197 job may have to wait for a lock while a newer job doesn't need any locks
198 and can be executed right away. Operations requiring a certain order
199 need to be submitted as a single job, or the client must submit one job
200 at a time and wait for it to finish before continuing.
202 For example, shutting down the entire cluster can be done by running the
203 command ``gnt-instance shutdown --all``, which will submit for each
204 instance a separate job containing the “shutdown instance” OpCode.
210 You need to have your Ganeti cluster installed and configured before you
211 try any of the commands in this document. Please follow the
212 :doc:`install` for instructions on how to do that.
220 The add operation might seem complex due to the many parameters it
221 accepts, but once you have understood the (few) required parameters and
222 the customisation capabilities you will see it is an easy operation.
224 The add operation requires at minimum five parameters:
226 - the OS for the instance
228 - the disk count and size
229 - the node specification or alternatively the iallocator to use
230 - and finally the instance name
232 The OS for the instance must be visible in the output of the command
233 ``gnt-os list`` and specifies which guest OS to install on the instance.
235 The disk template specifies what kind of storage to use as backend for
236 the (virtual) disks presented to the instance; note that for instances
237 with multiple virtual disks, they all must be of the same type.
239 The node(s) on which the instance will run can be given either manually,
240 via the ``-n`` option, or computed automatically by Ganeti, if you have
241 installed any iallocator script.
243 With the above parameters in mind, the command is::
246 -n %TARGET_NODE%:%SECONDARY_NODE% \
248 -t %DISK_TEMPLATE% -s %DISK_SIZE% \
251 The instance name must be resolvable (e.g. exist in DNS) and usually
252 points to an address in the same subnet as the cluster itself.
254 The above command has the minimum required options; other options you
255 can give include, among others:
257 - The maximum/minimum memory size (``-B maxmem``, ``-B minmem``)
258 (``-B memory`` can be used to specify only one size)
260 - The number of virtual CPUs (``-B vcpus``)
262 - Arguments for the NICs of the instance; by default, a single-NIC
263 instance is created. The IP and/or bridge of the NIC can be changed
264 via ``--nic 0:ip=IP,bridge=BRIDGE``
266 See the manpage for gnt-instance for the detailed option list.
268 For example if you want to create an highly available instance, with a
269 single disk of 50GB and the default memory size, having primary node
270 ``node1`` and secondary node ``node3``, use the following command::
272 $ gnt-instance add -n node1:node3 -o debootstrap -t drbd -s 50G \
275 There is a also a command for batch instance creation from a
276 specification file, see the ``batch-create`` operation in the
277 gnt-instance manual page.
279 Regular instance operations
280 +++++++++++++++++++++++++++
285 Removing an instance is even easier than creating one. This operation is
286 irreversible and destroys all the contents of your instance. Use with
289 $ gnt-instance remove %INSTANCE_NAME%
291 .. _instance-startup-label:
296 Instances are automatically started at instance creation time. To
297 manually start one which is currently stopped you can run::
299 $ gnt-instance startup %INSTANCE_NAME%
301 Ganeti will start an instance with up to its maximum instance memory. If
302 not enough memory is available Ganeti will use all the available memory
303 down to the instance minumum memory. If not even that amount of memory
304 is free Ganeti will refuse to start the instance.
306 Note, that this will not work when an instance is in a permanently
307 stopped state ``offline``. In this case, you will first have to
308 put it back to online mode by running::
310 $ gnt-instance modify --online %INSTANCE_NAME%
312 The command to stop the running instance is::
314 $ gnt-instance shutdown %INSTANCE_NAME%
316 If you want to shut the instance down more permanently, so that it
317 does not require dynamically allocated resources (memory and vcpus),
318 after shutting down an instance, execute the following::
320 $ gnt-instance modify --offline %INSTANCE_NAME%
322 .. warning:: Do not use the Xen or KVM commands directly to stop
323 instances. If you run for example ``xm shutdown`` or ``xm destroy``
324 on an instance Ganeti will automatically restart it (via
325 the :command:`ganeti-watcher` command which is launched via cron).
330 There are two ways to get information about instances: listing
331 instances, which does a tabular output containing a given set of fields
332 about each instance, and querying detailed information about a set of
335 The command to see all the instances configured and their status is::
339 The command can return a custom set of information when using the ``-o``
340 option (as always, check the manpage for a detailed specification). Each
341 instance will be represented on a line, thus making it easy to parse
342 this output via the usual shell utilities (grep, sed, etc.).
344 To get more detailed information about an instance, you can run::
346 $ gnt-instance info %INSTANCE%
348 which will give a multi-line block of information about the instance,
349 it's hardware resources (especially its disks and their redundancy
350 status), etc. This is harder to parse and is more expensive than the
351 list operation, but returns much more detailed information.
353 Changing an instance's runtime memory
354 +++++++++++++++++++++++++++++++++++++
356 Ganeti will always make sure an instance has a value between its maximum
357 and its minimum memory available as runtime memory. As of version 2.6
358 Ganeti will only choose a size different than the maximum size when
359 starting up, failing over, or migrating an instance on a node with less
360 than the maximum memory available. It won't resize other instances in
361 order to free up space for an instance.
363 If you find that you need more memory on a node any instance can be
364 manually resized without downtime, with the command::
366 $ gnt-instance modify -m %SIZE% %INSTANCE_NAME%
368 The same command can also be used to increase the memory available on an
369 instance, provided that enough free memory is available on its node, and
370 the specified size is not larger than the maximum memory size the
371 instance had when it was first booted (an instance will be unable to see
372 new memory above the maximum that was specified to the hypervisor at its
373 boot time, if it needs to grow further a reboot becomes necessary).
378 You can create a snapshot of an instance disk and its Ganeti
379 configuration, which then you can backup, or import into another
380 cluster. The way to export an instance is::
382 $ gnt-backup export -n %TARGET_NODE% %INSTANCE_NAME%
385 The target node can be any node in the cluster with enough space under
386 ``/srv/ganeti`` to hold the instance image. Use the ``--noshutdown``
387 option to snapshot an instance without rebooting it. Note that Ganeti
388 only keeps one snapshot for an instance - any previous snapshot of the
389 same instance existing cluster-wide under ``/srv/ganeti`` will be
390 removed by this operation: if you want to keep them, you need to move
391 them out of the Ganeti exports directory.
393 Importing an instance is similar to creating a new one, but additionally
394 one must specify the location of the snapshot. The command is::
396 $ gnt-backup import -n %TARGET_NODE% \
397 --src-node=%NODE% --src-dir=%DIR% %INSTANCE_NAME%
399 By default, parameters will be read from the export information, but you
400 can of course pass them in via the command line - most of the options
401 available for the command :command:`gnt-instance add` are supported here
404 Import of foreign instances
405 +++++++++++++++++++++++++++
407 There is a possibility to import a foreign instance whose disk data is
408 already stored as LVM volumes without going through copying it: the disk
411 For this, ensure that the original, non-managed instance is stopped,
412 then create a Ganeti instance in the usual way, except that instead of
413 passing the disk information you specify the current volumes::
415 $ gnt-instance add -t plain -n %HOME_NODE% ... \
416 --disk 0:adopt=%lv_name%[,vg=%vg_name%] %INSTANCE_NAME%
418 This will take over the given logical volumes, rename them to the Ganeti
419 standard (UUID-based), and without installing the OS on them start
420 directly the instance. If you configure the hypervisor similar to the
421 non-managed configuration that the instance had, the transition should
422 be seamless for the instance. For more than one disk, just pass another
423 disk parameter (e.g. ``--disk 1:adopt=...``).
425 Instance kernel selection
426 +++++++++++++++++++++++++
428 The kernel that instances uses to bootup can come either from the node,
429 or from instances themselves, depending on the setup.
434 With Xen PVM, there are three options.
436 First, you can use a kernel from the node, by setting the hypervisor
439 - ``kernel_path`` to a valid file on the node (and appropriately
441 - ``kernel_args`` optionally set to a valid Linux setting (e.g. ``ro``)
442 - ``root_path`` to a valid setting (e.g. ``/dev/xvda1``)
443 - ``bootloader_path`` and ``bootloader_args`` to empty
445 Alternatively, you can delegate the kernel management to instances, and
446 use either ``pvgrub`` or the deprecated ``pygrub``. For this, you must
447 install the kernels and initrds in the instance and create a valid GRUB
448 v1 configuration file.
450 For ``pvgrub`` (new in version 2.4.2), you need to set:
452 - ``kernel_path`` to point to the ``pvgrub`` loader present on the node
453 (e.g. ``/usr/lib/xen/boot/pv-grub-x86_32.gz``)
454 - ``kernel_args`` to the path to the GRUB config file, relative to the
455 instance (e.g. ``(hd0,0)/grub/menu.lst``)
456 - ``root_path`` **must** be empty
457 - ``bootloader_path`` and ``bootloader_args`` to empty
459 While ``pygrub`` is deprecated, here is how you can configure it:
461 - ``bootloader_path`` to the pygrub binary (e.g. ``/usr/bin/pygrub``)
462 - the other settings are not important
464 More information can be found in the Xen wiki pages for `pvgrub
465 <http://wiki.xensource.com/xenwiki/PvGrub>`_ and `pygrub
466 <http://wiki.xensource.com/xenwiki/PyGrub>`_.
471 For KVM also the kernel can be loaded either way.
473 For loading the kernels from the node, you need to set:
475 - ``kernel_path`` to a valid value
476 - ``initrd_path`` optionally set if you use an initrd
477 - ``kernel_args`` optionally set to a valid value (e.g. ``ro``)
479 If you want instead to have the instance boot from its disk (and execute
480 its bootloader), simply set the ``kernel_path`` parameter to an empty
481 string, and all the others will be ignored.
486 .. note:: This section only applies to multi-node clusters
488 .. _instance-change-primary-label:
490 Changing the primary node
491 +++++++++++++++++++++++++
493 There are three ways to exchange an instance's primary and secondary
494 nodes; the right one to choose depends on how the instance has been
495 created and the status of its current primary node. See
496 :ref:`rest-redundancy-label` for information on changing the secondary
497 node. Note that it's only possible to change the primary node to the
498 secondary and vice-versa; a direct change of the primary node with a
499 third node, while keeping the current secondary is not possible in a
500 single step, only via multiple operations as detailed in
501 :ref:`instance-relocation-label`.
503 Failing over an instance
504 ~~~~~~~~~~~~~~~~~~~~~~~~
506 If an instance is built in highly available mode you can at any time
507 fail it over to its secondary node, even if the primary has somehow
508 failed and it's not up anymore. Doing it is really easy, on the master
509 node you can just run::
511 $ gnt-instance failover %INSTANCE_NAME%
513 That's it. After the command completes the secondary node is now the
514 primary, and vice-versa.
516 The instance will be started with an amount of memory between its
517 ``maxmem`` and its ``minmem`` value, depending on the free memory on its
518 target node, or the operation will fail if that's not possible. See
519 :ref:`instance-startup-label` for details.
521 If the instance's disk template is of type rbd, then you can specify
522 the target node (which can be any node) explicitly, or specify an
523 iallocator plugin. If you omit both, the default iallocator will be
524 used to determine the target node::
526 $ gnt-instance failover -n %TARGET_NODE% %INSTANCE_NAME%
528 Live migrating an instance
529 ~~~~~~~~~~~~~~~~~~~~~~~~~~
531 If an instance is built in highly available mode, it currently runs and
532 both its nodes are running fine, you can at migrate it over to its
533 secondary node, without downtime. On the master node you need to run::
535 $ gnt-instance migrate %INSTANCE_NAME%
537 The current load on the instance and its memory size will influence how
538 long the migration will take. In any case, for both KVM and Xen
539 hypervisors, the migration will be transparent to the instance.
541 If the destination node has less memory than the instance's current
542 runtime memory, but at least the instance's minimum memory available
543 Ganeti will automatically reduce the instance runtime memory before
544 migrating it, unless the ``--no-runtime-changes`` option is passed, in
545 which case the target node should have at least the instance's current
548 If the instance's disk template is of type rbd, then you can specify
549 the target node (which can be any node) explicitly, or specify an
550 iallocator plugin. If you omit both, the default iallocator will be
551 used to determine the target node::
553 $ gnt-instance migrate -n %TARGET_NODE% %INSTANCE_NAME%
555 Moving an instance (offline)
556 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
558 If an instance has not been create as mirrored, then the only way to
559 change its primary node is to execute the move command::
561 $ gnt-instance move -n %NEW_NODE% %INSTANCE%
563 This has a few prerequisites:
565 - the instance must be stopped
566 - its current primary node must be on-line and healthy
567 - the disks of the instance must not have any errors
569 Since this operation actually copies the data from the old node to the
570 new node, expect it to take proportional to the size of the instance's
571 disks and the speed of both the nodes' I/O system and their networking.
576 Disk failures are a common cause of errors in any server
577 deployment. Ganeti offers protection from single-node failure if your
578 instances were created in HA mode, and it also offers ways to restore
579 redundancy after a failure.
581 Preparing for disk operations
582 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
584 It is important to note that for Ganeti to be able to do any disk
585 operation, the Linux machines on top of which Ganeti must be consistent;
586 for LVM, this means that the LVM commands must not return failures; it
587 is common that after a complete disk failure, any LVM command aborts
588 with an error similar to::
591 /dev/sdb1: read failed after 0 of 4096 at 0: Input/output error
592 /dev/sdb1: read failed after 0 of 4096 at 750153695232: Input/output error
593 /dev/sdb1: read failed after 0 of 4096 at 0: Input/output error
594 Couldn't find device with uuid 't30jmN-4Rcf-Fr5e-CURS-pawt-z0jU-m1TgeJ'.
595 Couldn't find all physical volumes for volume group xenvg.
597 Before restoring an instance's disks to healthy status, it's needed to
598 fix the volume group used by Ganeti so that we can actually create and
599 manage the logical volumes. This is usually done in a multi-step
602 #. first, if the disk is completely gone and LVM commands exit with
603 “Couldn't find device with uuid…” then you need to run the command::
605 $ vgreduce --removemissing %VOLUME_GROUP%
607 #. after the above command, the LVM commands should be executing
608 normally (warnings are normal, but the commands will not fail
611 #. if the failed disk is still visible in the output of the ``pvs``
612 command, you need to deactivate it from allocations by running::
614 $ pvs -x n /dev/%DISK%
616 At this point, the volume group should be consistent and any bad
617 physical volumes should not longer be available for allocation.
619 Note that since version 2.1 Ganeti provides some commands to automate
620 these two operations, see :ref:`storage-units-label`.
622 .. _rest-redundancy-label:
624 Restoring redundancy for DRBD-based instances
625 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
627 A DRBD instance has two nodes, and the storage on one of them has
628 failed. Depending on which node (primary or secondary) has failed, you
629 have three options at hand:
631 - if the storage on the primary node has failed, you need to re-create
633 - if the storage on the secondary node has failed, you can either
634 re-create the disks on it or change the secondary and recreate
635 redundancy on the new secondary node
637 Of course, at any point it's possible to force re-creation of disks even
638 though everything is already fine.
640 For all three cases, the ``replace-disks`` operation can be used::
642 # re-create disks on the primary node
643 $ gnt-instance replace-disks -p %INSTANCE_NAME%
644 # re-create disks on the current secondary
645 $ gnt-instance replace-disks -s %INSTANCE_NAME%
646 # change the secondary node, via manual specification
647 $ gnt-instance replace-disks -n %NODE% %INSTANCE_NAME%
648 # change the secondary node, via an iallocator script
649 $ gnt-instance replace-disks -I %SCRIPT% %INSTANCE_NAME%
650 # since Ganeti 2.1: automatically fix the primary or secondary node
651 $ gnt-instance replace-disks -a %INSTANCE_NAME%
653 Since the process involves copying all data from the working node to the
654 target node, it will take a while, depending on the instance's disk
655 size, node I/O system and network speed. But it is (barring any network
656 interruption) completely transparent for the instance.
658 Re-creating disks for non-redundant instances
659 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
661 .. versionadded:: 2.1
663 For non-redundant instances, there isn't a copy (except backups) to
664 re-create the disks. But it's possible to at-least re-create empty
665 disks, after which a reinstall can be run, via the ``recreate-disks``
668 $ gnt-instance recreate-disks %INSTANCE%
670 Note that this will fail if the disks already exists.
672 Conversion of an instance's disk type
673 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
675 It is possible to convert between a non-redundant instance of type
676 ``plain`` (LVM storage) and redundant ``drbd`` via the ``gnt-instance
679 # start with a non-redundant instance
680 $ gnt-instance add -t plain ... %INSTANCE%
682 # later convert it to redundant
683 $ gnt-instance stop %INSTANCE%
684 $ gnt-instance modify -t drbd -n %NEW_SECONDARY% %INSTANCE%
685 $ gnt-instance start %INSTANCE%
687 # and convert it back
688 $ gnt-instance stop %INSTANCE%
689 $ gnt-instance modify -t plain %INSTANCE%
690 $ gnt-instance start %INSTANCE%
692 The conversion must be done while the instance is stopped, and
693 converting from plain to drbd template presents a small risk, especially
694 if the instance has multiple disks and/or if one node fails during the
695 conversion procedure). As such, it's recommended (as always) to make
696 sure that downtime for manual recovery is acceptable and that the
697 instance has up-to-date backups.
702 Accessing an instance's disks
703 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
705 From an instance's primary node you can have access to its disks. Never
706 ever mount the underlying logical volume manually on a fault tolerant
707 instance, or will break replication and your data will be
708 inconsistent. The correct way to access an instance's disks is to run
709 (on the master node, as usual) the command::
711 $ gnt-instance activate-disks %INSTANCE%
713 And then, *on the primary node of the instance*, access the device that
714 gets created. For example, you could mount the given disks, then edit
715 files on the filesystem, etc.
717 Note that with partitioned disks (as opposed to whole-disk filesystems),
718 you will need to use a tool like :manpage:`kpartx(8)`::
721 $ gnt-instance activate-disks %instance1%
727 $ mount /dev/mapper/… /mnt/
728 # edit files under mnt as desired
734 After you've finished you can deactivate them with the deactivate-disks
735 command, which works in the same way::
737 $ gnt-instance deactivate-disks %INSTANCE%
739 Note that if any process started by you is still using the disks, the
740 above command will error out, and you **must** cleanup and ensure that
741 the above command runs successfully before you start the instance,
742 otherwise the instance will suffer corruption.
744 Accessing an instance's console
745 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
747 The command to access a running instance's console is::
749 $ gnt-instance console %INSTANCE_NAME%
751 Use the console normally and then type ``^]`` when done, to exit.
753 Other instance operations
754 +++++++++++++++++++++++++
759 There is a wrapper command for rebooting instances::
761 $ gnt-instance reboot %instance2%
763 By default, this does the equivalent of shutting down and then starting
764 the instance, but it accepts parameters to perform a soft-reboot (via
765 the hypervisor), a hard reboot (hypervisor shutdown and then startup) or
766 a full one (the default, which also de-configures and then configures
767 again the disks of the instance).
769 Instance OS definitions debugging
770 +++++++++++++++++++++++++++++++++
772 Should you have any problems with instance operating systems the command
773 to see a complete status for all your nodes is::
777 .. _instance-relocation-label:
782 While it is not possible to move an instance from nodes ``(A, B)`` to
783 nodes ``(C, D)`` in a single move, it is possible to do so in a few
786 # instance is located on A, B
787 $ gnt-instance replace -n %nodeC% %instance1%
788 # instance has moved from (A, B) to (A, C)
789 # we now flip the primary/secondary nodes
790 $ gnt-instance migrate %instance1%
791 # instance lives on (C, A)
792 # we can then change A to D via:
793 $ gnt-instance replace -n %nodeD% %instance1%
795 Which brings it into the final configuration of ``(C, D)``. Note that we
796 needed to do two replace-disks operation (two copies of the instance
797 disks), because we needed to get rid of both the original nodes (A and
803 There are much fewer node operations available than for instances, but
804 they are equivalently important for maintaining a healthy cluster.
809 It is at any time possible to extend the cluster with one more node, by
810 using the node add operation::
812 $ gnt-node add %NEW_NODE%
814 If the cluster has a replication network defined, then you need to pass
815 the ``-s REPLICATION_IP`` parameter to this option.
817 A variation of this command can be used to re-configure a node if its
818 Ganeti configuration is broken, for example if it has been reinstalled
821 $ gnt-node add --readd %EXISTING_NODE%
823 This will reinitialise the node as if it's been newly added, but while
824 keeping its existing configuration in the cluster (primary/secondary IP,
825 etc.), in other words you won't need to use ``-s`` here.
827 Changing the node role
828 ++++++++++++++++++++++
830 A node can be in different roles, as explained in the
831 :ref:`terminology-label` section. Promoting a node to the master role is
832 special, while the other roles are handled all via a single command.
834 Failing over the master node
835 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
837 If you want to promote a different node to the master role (for whatever
838 reason), run on any other master-candidate node the command::
840 $ gnt-cluster master-failover
842 and the node you ran it on is now the new master. In case you try to run
843 this on a non master-candidate node, you will get an error telling you
844 which nodes are valid.
846 Changing between the other roles
847 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
849 The ``gnt-node modify`` command can be used to select a new role::
851 # change to master candidate
852 $ gnt-node modify -C yes %NODE%
853 # change to drained status
854 $ gnt-node modify -D yes %NODE%
855 # change to offline status
856 $ gnt-node modify -O yes %NODE%
857 # change to regular mode (reset all flags)
858 $ gnt-node modify -O no -D no -C no %NODE%
860 Note that the cluster requires that at any point in time, a certain
861 number of nodes are master candidates, so changing from master candidate
862 to other roles might fail. It is recommended to either force the
863 operation (via the ``--force`` option) or first change the number of
864 master candidates in the cluster - see :ref:`cluster-config-label`.
869 There are two steps of moving instances off a node:
871 - moving the primary instances (actually converting them into secondary
873 - moving the secondary instances (including any instances converted in
876 Primary instance conversion
877 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
879 For this step, you can use either individual instance move
880 commands (as seen in :ref:`instance-change-primary-label`) or the bulk
881 per-node versions; these are::
883 $ gnt-node migrate %NODE%
884 $ gnt-node evacuate -s %NODE%
886 Note that the instance “move” command doesn't currently have a node
889 Both these commands, or the equivalent per-instance command, will make
890 this node the secondary node for the respective instances, whereas their
891 current secondary node will become primary. Note that it is not possible
892 to change in one step the primary node to another node as primary, while
893 keeping the same secondary node.
895 Secondary instance evacuation
896 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
898 For the evacuation of secondary instances, a command called
899 :command:`gnt-node evacuate` is provided and its syntax is::
901 $ gnt-node evacuate -I %IALLOCATOR_SCRIPT% %NODE%
902 $ gnt-node evacuate -n %DESTINATION_NODE% %NODE%
904 The first version will compute the new secondary for each instance in
905 turn using the given iallocator script, whereas the second one will
906 simply move all instances to DESTINATION_NODE.
911 Once a node no longer has any instances (neither primary nor secondary),
912 it's easy to remove it from the cluster::
914 $ gnt-node remove %NODE_NAME%
916 This will deconfigure the node, stop the ganeti daemons on it and leave
917 it hopefully like before it joined to the cluster.
919 Replication network changes
920 +++++++++++++++++++++++++++
922 The :command:`gnt-node modify -s` command can be used to change the
923 secondary IP of a node. This operation can only be performed if:
925 - No instance is active on the target node
926 - The new target IP is reachable from the master's secondary IP
928 Also this operation will not allow to change a node from single-homed
929 (same primary and secondary ip) to multi-homed (separate replication
930 network) or vice versa, unless:
932 - The target node is the master node and `--force` is passed.
933 - The target cluster is single-homed and the new primary ip is a change
934 to single homed for a particular node.
935 - The target cluster is multi-homed and the new primary ip is a change
936 to multi homed for a particular node.
938 For example to do a single-homed to multi-homed conversion::
940 $ gnt-node modify --force -s %SECONDARY_IP% %MASTER_NAME%
941 $ gnt-node modify -s %SECONDARY_IP% %NODE1_NAME%
942 $ gnt-node modify -s %SECONDARY_IP% %NODE2_NAME%
943 $ gnt-node modify -s %SECONDARY_IP% %NODE3_NAME%
946 The same commands can be used for multi-homed to single-homed except the
947 secondary IPs should be the same as the primaries for each node, for
953 When using LVM (either standalone or with DRBD), it can become tedious
954 to debug and fix it in case of errors. Furthermore, even file-based
955 storage can become complicated to handle manually on many hosts. Ganeti
956 provides a couple of commands to help with automation.
961 This is a command specific to LVM handling. It allows listing the
962 logical volumes on a given node or on all nodes and their association to
963 instances via the ``volumes`` command::
966 Node PhysDev VG Name Size Instance
967 node1 /dev/sdb1 xenvg e61fbc97-….disk0 512M instance17
968 node1 /dev/sdb1 xenvg ebd1a7d1-….disk0 512M instance19
969 node2 /dev/sdb1 xenvg 0af08a3d-….disk0 512M instance20
970 node2 /dev/sdb1 xenvg cc012285-….disk0 512M instance16
971 node2 /dev/sdb1 xenvg f0fac192-….disk0 512M instance18
973 The above command maps each logical volume to a volume group and
974 underlying physical volume and (possibly) to an instance.
976 .. _storage-units-label:
978 Generalized storage handling
979 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
981 .. versionadded:: 2.1
983 Starting with Ganeti 2.1, a new storage framework has been implemented
984 that tries to abstract the handling of the storage type the cluster
987 First is listing the backend storage and their space situation::
989 $ gnt-node list-storage
990 Node Name Size Used Free
991 node1 /dev/sda7 673.8G 0M 673.8G
992 node1 /dev/sdb1 698.6G 1.5G 697.1G
993 node2 /dev/sda7 673.8G 0M 673.8G
994 node2 /dev/sdb1 698.6G 1.0G 697.6G
996 The default is to list LVM physical volumes. It's also possible to list
997 the LVM volume groups::
999 $ gnt-node list-storage -t lvm-vg
1004 Next is repairing storage units, which is currently only implemented for
1005 volume groups and does the equivalent of ``vgreduce --removemissing``::
1007 $ gnt-node repair-storage %node2% lvm-vg xenvg
1008 Sun Oct 25 22:21:45 2009 Repairing storage unit 'xenvg' on node2 ...
1010 Last is the modification of volume properties, which is (again) only
1011 implemented for LVM physical volumes and allows toggling the
1012 ``allocatable`` value::
1014 $ gnt-node modify-storage --allocatable=no %node2% lvm-pv /dev/%sdb1%
1016 Use of the storage commands
1017 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
1019 All these commands are needed when recovering a node from a disk
1022 - first, we need to recover from complete LVM failure (due to missing
1023 disk), by running the ``repair-storage`` command
1024 - second, we need to change allocation on any partially-broken disk
1025 (i.e. LVM still sees it, but it has bad blocks) by running
1027 - then we can evacuate the instances as needed
1033 Beside the cluster initialisation command (which is detailed in the
1034 :doc:`install` document) and the master failover command which is
1035 explained under node handling, there are a couple of other cluster
1036 operations available.
1038 .. _cluster-config-label:
1043 One of the few commands that can be run on any node (not only the
1044 master) is the ``getmaster`` command::
1047 $ gnt-cluster getmaster
1050 It is possible to query and change global cluster parameters via the
1051 ``info`` and ``modify`` commands::
1054 Cluster name: cluster.example.com
1055 Cluster UUID: 07805e6f-f0af-4310-95f1-572862ee939c
1056 Creation time: 2009-09-25 05:04:15
1057 Modification time: 2009-10-18 22:11:47
1058 Master node: node1.example.com
1059 Architecture (this node): 64bit (x86_64)
1062 Default hypervisor: xen-pvm
1063 Enabled hypervisors: xen-pvm
1064 Hypervisor parameters:
1066 root_path: /dev/sda1
1069 - candidate pool size: 10
1071 Default instance parameters:
1075 Default nic parameters:
1080 There various parameters above can be changed via the ``modify``
1081 commands as follows:
1083 - the hypervisor parameters can be changed via ``modify -H
1084 xen-pvm:root_path=…``, and so on for other hypervisors/key/values
1085 - the "default instance parameters" are changeable via ``modify -B
1086 parameter=value…`` syntax
1087 - the cluster parameters are changeable via separate options to the
1088 modify command (e.g. ``--candidate-pool-size``, etc.)
1090 For detailed option list see the :manpage:`gnt-cluster(8)` man page.
1092 The cluster version can be obtained via the ``version`` command::
1093 $ gnt-cluster version
1094 Software version: 2.1.0
1095 Internode protocol: 20
1096 Configuration format: 2010000
1100 This is not very useful except when debugging Ganeti.
1102 Global node commands
1103 ++++++++++++++++++++
1105 There are two commands provided for replicating files to all nodes of a
1106 cluster and for running commands on all the nodes::
1108 $ gnt-cluster copyfile %/path/to/file%
1109 $ gnt-cluster command %ls -l /path/to/file%
1111 These are simple wrappers over scp/ssh and more advanced usage can be
1112 obtained using :manpage:`dsh(1)` and similar commands. But they are
1113 useful to update an OS script from the master node, for example.
1115 Cluster verification
1116 ++++++++++++++++++++
1118 There are three commands that relate to global cluster checks. The first
1119 one is ``verify`` which gives an overview on the cluster state,
1120 highlighting any issues. In normal operation, this command should return
1121 no ``ERROR`` messages::
1123 $ gnt-cluster verify
1124 Sun Oct 25 23:08:58 2009 * Verifying global settings
1125 Sun Oct 25 23:08:58 2009 * Gathering data (2 nodes)
1126 Sun Oct 25 23:09:00 2009 * Verifying node status
1127 Sun Oct 25 23:09:00 2009 * Verifying instance status
1128 Sun Oct 25 23:09:00 2009 * Verifying orphan volumes
1129 Sun Oct 25 23:09:00 2009 * Verifying remaining instances
1130 Sun Oct 25 23:09:00 2009 * Verifying N+1 Memory redundancy
1131 Sun Oct 25 23:09:00 2009 * Other Notes
1132 Sun Oct 25 23:09:00 2009 - NOTICE: 5 non-redundant instance(s) found.
1133 Sun Oct 25 23:09:00 2009 * Hooks Results
1135 The second command is ``verify-disks``, which checks that the instance's
1136 disks have the correct status based on the desired instance state
1139 $ gnt-cluster verify-disks
1141 Note that this command will show no output when disks are healthy.
1143 The last command is used to repair any discrepancies in Ganeti's
1144 recorded disk size and the actual disk size (disk size information is
1145 needed for proper activation and growth of DRBD-based disks)::
1147 $ gnt-cluster repair-disk-sizes
1148 Sun Oct 25 23:13:16 2009 - INFO: Disk 0 of instance instance1 has mismatched size, correcting: recorded 512, actual 2048
1149 Sun Oct 25 23:13:17 2009 - WARNING: Invalid result from node node4, ignoring node results
1151 The above shows one instance having wrong disk size, and a node which
1152 returned invalid data, and thus we ignored all primary instances of that
1155 Configuration redistribution
1156 ++++++++++++++++++++++++++++
1158 If the verify command complains about file mismatches between the master
1159 and other nodes, due to some node problems or if you manually modified
1160 configuration files, you can force an push of the master configuration
1161 to all other nodes via the ``redist-conf`` command::
1163 $ gnt-cluster redist-conf
1165 This command will be silent unless there are problems sending updates to
1172 It is possible to rename a cluster, or to change its IP address, via the
1173 ``rename`` command. If only the IP has changed, you need to pass the
1174 current name and Ganeti will realise its IP has changed::
1176 $ gnt-cluster rename %cluster.example.com%
1177 This will rename the cluster to 'cluster.example.com'. If
1178 you are connected over the network to the cluster name, the operation
1179 is very dangerous as the IP address will be removed from the node and
1180 the change may not go through. Continue?
1182 Failure: prerequisites not met for this operation:
1183 Neither the name nor the IP address of the cluster has changed
1185 In the above output, neither value has changed since the cluster
1186 initialisation so the operation is not completed.
1191 The job queue execution in Ganeti 2.0 and higher can be inspected,
1192 suspended and resumed via the ``queue`` command::
1194 $ gnt-cluster queue info
1195 The drain flag is unset
1196 $ gnt-cluster queue drain
1197 $ gnt-instance stop %instance1%
1198 Failed to submit job for instance1: Job queue is drained, refusing job
1199 $ gnt-cluster queue info
1200 The drain flag is set
1201 $ gnt-cluster queue undrain
1203 This is most useful if you have an active cluster and you need to
1204 upgrade the Ganeti software, or simply restart the software on any node:
1206 #. suspend the queue via ``queue drain``
1207 #. wait until there are no more running jobs via ``gnt-job list``
1208 #. restart the master or another node, or upgrade the software
1209 #. resume the queue via ``queue undrain``
1211 .. note:: this command only stores a local flag file, and if you
1212 failover the master, it will not have effect on the new master.
1218 The :manpage:`ganeti-watcher` is a program, usually scheduled via
1219 ``cron``, that takes care of cluster maintenance operations (restarting
1220 downed instances, activating down DRBD disks, etc.). However, during
1221 maintenance and troubleshooting, this can get in your way; disabling it
1222 via commenting out the cron job is not so good as this can be
1223 forgotten. Thus there are some commands for automated control of the
1224 watcher: ``pause``, ``info`` and ``continue``::
1226 $ gnt-cluster watcher info
1227 The watcher is not paused.
1228 $ gnt-cluster watcher pause %1h%
1229 The watcher is paused until Mon Oct 26 00:30:37 2009.
1230 $ gnt-cluster watcher info
1231 The watcher is paused until Mon Oct 26 00:30:37 2009.
1233 2009-10-25 23:30:47,984: pid=28867 ganeti-watcher:486 DEBUG Pause has been set, exiting
1234 $ gnt-cluster watcher continue
1235 The watcher is no longer paused.
1237 2009-10-25 23:31:04,789: pid=28976 ganeti-watcher:345 DEBUG Archived 0 jobs, left 0
1238 2009-10-25 23:31:05,884: pid=28976 ganeti-watcher:280 DEBUG Got data from cluster, writing instance status file
1239 2009-10-25 23:31:06,061: pid=28976 ganeti-watcher:150 DEBUG Data didn't change, just touching status file
1240 $ gnt-cluster watcher info
1241 The watcher is not paused.
1243 The exact details of the argument to the ``pause`` command are available
1246 .. note:: this command only stores a local flag file, and if you
1247 failover the master, it will not have effect on the new master.
1249 Node auto-maintenance
1250 +++++++++++++++++++++
1252 If the cluster parameter ``maintain_node_health`` is enabled (see the
1253 manpage for :command:`gnt-cluster`, the init and modify subcommands),
1254 then the following will happen automatically:
1256 - the watcher will shutdown any instances running on offline nodes
1257 - the watcher will deactivate any DRBD devices on offline nodes
1259 In the future, more actions are planned, so only enable this parameter
1260 if the nodes are completely dedicated to Ganeti; otherwise it might be
1261 possible to lose data due to auto-maintenance actions.
1263 Removing a cluster entirely
1264 +++++++++++++++++++++++++++
1266 The usual method to cleanup a cluster is to run ``gnt-cluster destroy``
1267 however if the Ganeti installation is broken in any way then this will
1270 It is possible in such a case to cleanup manually most if not all traces
1271 of a cluster installation by following these steps on all of the nodes:
1273 1. Shutdown all instances. This depends on the virtualisation method
1274 used (Xen, KVM, etc.):
1276 - Xen: run ``xm list`` and ``xm destroy`` on all the non-Domain-0
1278 - KVM: kill all the KVM processes
1279 - chroot: kill all processes under the chroot mountpoints
1281 2. If using DRBD, shutdown all DRBD minors (which should by at this time
1282 no-longer in use by instances); on each node, run ``drbdsetup
1283 /dev/drbdN down`` for each active DRBD minor.
1285 3. If using LVM, cleanup the Ganeti volume group; if only Ganeti created
1286 logical volumes (and you are not sharing the volume group with the
1287 OS, for example), then simply running ``lvremove -f xenvg`` (replace
1288 'xenvg' with your volume group name) should do the required cleanup.
1290 4. If using file-based storage, remove recursively all files and
1291 directories under your file-storage directory: ``rm -rf
1292 /srv/ganeti/file-storage/*`` replacing the path with the correct path
1295 5. Stop the ganeti daemons (``/etc/init.d/ganeti stop``) and kill any
1296 that remain alive (``pgrep ganeti`` and ``pkill ganeti``).
1298 6. Remove the ganeti state directory (``rm -rf /var/lib/ganeti/*``),
1299 replacing the path with the correct path for your installation.
1301 7. If using RBD, run ``rbd unmap /dev/rbdN`` to unmap the RBD disks.
1302 Then remove the RBD disk images used by Ganeti, identified by their
1303 UUIDs (``rbd rm uuid.rbd.diskN``).
1305 On the master node, remove the cluster from the master-netdev (usually
1306 ``xen-br0`` for bridged mode, otherwise ``eth0`` or similar), by running
1307 ``ip a del $clusterip/32 dev xen-br0`` (use the correct cluster ip and
1308 network device name).
1310 At this point, the machines are ready for a cluster creation; in case
1311 you want to remove Ganeti completely, you need to also undo some of the
1312 SSH changes and log directories:
1314 - ``rm -rf /var/log/ganeti /srv/ganeti`` (replace with the correct
1316 - remove from ``/root/.ssh`` the keys that Ganeti added (check the
1317 ``authorized_keys`` and ``id_dsa`` files)
1318 - regenerate the host's SSH keys (check the OpenSSH startup scripts)
1321 Otherwise, if you plan to re-create the cluster, you can just go ahead
1322 and rerun ``gnt-cluster init``.
1327 The tags handling (addition, removal, listing) is similar for all the
1328 objects that support it (instances, nodes, and the cluster).
1333 Note that the set of characters present in a tag and the maximum tag
1334 length are restricted. Currently the maximum length is 128 characters,
1335 there can be at most 4096 tags per object, and the set of characters is
1336 comprised by alphanumeric characters and additionally ``.+*/:@-``.
1341 Tags can be added via ``add-tags``::
1343 $ gnt-instance add-tags %INSTANCE% %a% %b% %c%
1344 $ gnt-node add-tags %INSTANCE% %a% %b% %c%
1345 $ gnt-cluster add-tags %a% %b% %c%
1348 The above commands add three tags to an instance, to a node and to the
1349 cluster. Note that the cluster command only takes tags as arguments,
1350 whereas the node and instance commands first required the node and
1353 Tags can also be added from a file, via the ``--from=FILENAME``
1354 argument. The file is expected to contain one tag per line.
1356 Tags can also be remove via a syntax very similar to the add one::
1358 $ gnt-instance remove-tags %INSTANCE% %a% %b% %c%
1362 $ gnt-instance list-tags
1363 $ gnt-node list-tags
1364 $ gnt-cluster list-tags
1369 It is also possible to execute a global search on the all tags defined
1370 in the cluster configuration, via a cluster command::
1372 $ gnt-cluster search-tags %REGEXP%
1374 The parameter expected is a regular expression (see
1375 :manpage:`regex(7)`). This will return all tags that match the search,
1376 together with the object they are defined in (the names being show in a
1377 hierarchical kind of way)::
1379 $ gnt-cluster search-tags %o%
1381 /instances/instance1 owner:bar
1387 The various jobs submitted by the instance/node/cluster commands can be
1388 examined, canceled and archived by various invocations of the
1389 ``gnt-job`` command.
1391 First is the job list command::
1394 17771 success INSTANCE_QUERY_DATA
1395 17773 success CLUSTER_VERIFY_DISKS
1396 17775 success CLUSTER_REPAIR_DISK_SIZES
1397 17776 error CLUSTER_RENAME(cluster.example.com)
1398 17780 success CLUSTER_REDIST_CONF
1399 17792 success INSTANCE_REBOOT(instance1.example.com)
1401 More detailed information about a job can be found via the ``info``
1404 $ gnt-job info %17776%
1407 Received: 2009-10-25 23:18:02.180569
1408 Processing start: 2009-10-25 23:18:02.200335 (delta 0.019766s)
1409 Processing end: 2009-10-25 23:18:02.279743 (delta 0.079408s)
1410 Total processing time: 0.099174 seconds
1414 Processing start: 2009-10-25 23:18:02.200335
1415 Processing end: 2009-10-25 23:18:02.252282
1417 name: cluster.example.com
1420 [Neither the name nor the IP address of the cluster has changed]
1423 During the execution of a job, it's possible to follow the output of a
1424 job, similar to the log that one get from the ``gnt-`` commands, via the
1427 $ gnt-instance add --submit … %instance1%
1429 $ gnt-job watch %17818%
1430 Output from job 17818 follows
1431 -----------------------------
1432 Mon Oct 26 00:22:48 2009 - INFO: Selected nodes for instance instance1 via iallocator dumb: node1, node2
1433 Mon Oct 26 00:22:49 2009 * creating instance disks...
1434 Mon Oct 26 00:22:52 2009 adding instance instance1 to cluster config
1435 Mon Oct 26 00:22:52 2009 - INFO: Waiting for instance instance1 to sync disks.
1437 Mon Oct 26 00:23:03 2009 creating os for instance instance1 on node node1
1438 Mon Oct 26 00:23:03 2009 * running the instance OS create scripts...
1439 Mon Oct 26 00:23:13 2009 * starting instance...
1442 This is useful if you need to follow a job's progress from multiple
1445 A job that has not yet started to run can be canceled::
1447 $ gnt-job cancel %17810%
1449 But not one that has already started execution::
1451 $ gnt-job cancel %17805%
1452 Job 17805 is no longer waiting in the queue
1454 There are two queues for jobs: the *current* and the *archive*
1455 queue. Jobs are initially submitted to the current queue, and they stay
1456 in that queue until they have finished execution (either successfully or
1457 not). At that point, they can be moved into the archive queue using e.g.
1458 ``gnt-job autoarchive all``. The ``ganeti-watcher`` script will do this
1459 automatically 6 hours after a job is finished. The ``ganeti-cleaner``
1460 script will then remove archived the jobs from the archive directory
1463 Note that ``gnt-job list`` only shows jobs in the current queue.
1464 Archived jobs can be viewed using ``gnt-job info <id>``.
1466 Special Ganeti deployments
1467 --------------------------
1469 Since Ganeti 2.4, it is possible to extend the Ganeti deployment with
1470 two custom scenarios: Ganeti inside Ganeti and multi-site model.
1472 Running Ganeti under Ganeti
1473 +++++++++++++++++++++++++++
1475 It is sometimes useful to be able to use a Ganeti instance as a Ganeti
1476 node (part of another cluster, usually). One example scenario is two
1477 small clusters, where we want to have an additional master candidate
1478 that holds the cluster configuration and can be used for helping with
1479 the master voting process.
1481 However, these Ganeti instance should not host instances themselves, and
1482 should not be considered in the normal capacity planning, evacuation
1483 strategies, etc. In order to accomplish this, mark these nodes as
1484 non-``vm_capable``::
1486 $ gnt-node modify --vm-capable=no %node3%
1488 The vm_capable status can be listed as usual via ``gnt-node list``::
1490 $ gnt-node list -oname,vm_capable
1496 When this flag is set, the cluster will not do any operations that
1497 relate to instances on such nodes, e.g. hypervisor operations,
1498 disk-related operations, etc. Basically they will just keep the ssconf
1499 files, and if master candidates the full configuration.
1504 If Ganeti is deployed in multi-site model, with each site being a node
1505 group (so that instances are not relocated across the WAN by mistake),
1506 it is conceivable that either the WAN latency is high or that some sites
1507 have a lower reliability than others. In this case, it doesn't make
1508 sense to replicate the job information across all sites (or even outside
1509 of a “central” node group), so it should be possible to restrict which
1510 nodes can become master candidates via the auto-promotion algorithm.
1512 Ganeti 2.4 introduces for this purpose a new ``master_capable`` flag,
1513 which (when unset) prevents nodes from being marked as master
1514 candidates, either manually or automatically.
1516 As usual, the node modify operation can change this flag::
1518 $ gnt-node modify --auto-promote --master-capable=no %node3%
1519 Fri Jan 7 06:23:07 2011 - INFO: Demoting from master candidate
1520 Fri Jan 7 06:23:08 2011 - INFO: Promoted nodes to master candidate role: node4
1522 - master_capable -> False
1523 - master_candidate -> False
1525 And the node list operation will list this flag::
1527 $ gnt-node list -oname,master_capable %node1% %node2% %node3%
1533 Note that marking a node both not ``vm_capable`` and not
1534 ``master_capable`` makes the node practically unusable from Ganeti's
1535 point of view. Hence these two flags should be used probably in
1536 contrast: some nodes will be only master candidates (master_capable but
1537 not vm_capable), and other nodes will only hold instances (vm_capable
1538 but not master_capable).
1544 Beside the usual ``gnt-`` and ``ganeti-`` commands which are provided
1545 and installed in ``$prefix/sbin`` at install time, there are a couple of
1546 other tools installed which are used seldom but can be helpful in some
1552 The ``lvmstrap`` tool, introduced in :ref:`configure-lvm-label` section,
1553 has two modes of operation:
1555 - ``diskinfo`` shows the discovered disks on the system and their status
1556 - ``create`` takes all not-in-use disks and creates a volume group out
1559 .. warning:: The ``create`` argument to this command causes data-loss!
1564 The ``cfgupgrade`` tools is used to upgrade between major (and minor)
1565 Ganeti versions. Point-releases are usually transparent for the admin.
1567 More information about the upgrade procedure is listed on the wiki at
1568 http://code.google.com/p/ganeti/wiki/UpgradeNotes.
1570 There is also a script designed to upgrade from Ganeti 1.2 to 2.0,
1571 called ``cfgupgrade12``.
1576 .. note:: This command is not actively maintained; make sure you backup
1577 your configuration before using it
1579 This can be used as an alternative to direct editing of the
1580 main configuration file if Ganeti has a bug and prevents you, for
1581 example, from removing an instance or a node from the configuration
1589 .. warning:: This command will erase existing instances if given as
1592 This tool is used to exercise either the hardware of machines or
1593 alternatively the Ganeti software. It is safe to run on an existing
1594 cluster **as long as you don't pass it existing instance names**.
1596 The command will, by default, execute a comprehensive set of operations
1597 against a list of instances, these being:
1600 - disk replacement (for redundant instances)
1601 - failover and migration (for redundant instances)
1602 - move (for non-redundant instances)
1604 - add disks, remove disk
1605 - add NICs, remove NICs
1606 - export and then import
1610 - and finally removal of the test instances
1612 Executing all these operations will test that the hardware performs
1613 well: the creation, disk replace, disk add and disk growth will exercise
1614 the storage and network; the migrate command will test the memory of the
1615 systems. Depending on the passed options, it can also test that the
1616 instance OS definitions are executing properly the rename, import and
1622 This tool takes the Ganeti configuration and outputs a "sanitized"
1623 version, by randomizing or clearing:
1625 - DRBD secrets and cluster public key (always)
1626 - host names (optional)
1628 - OS names (optional)
1629 - LV names (optional, only useful for very old clusters which still have
1630 instances whose LVs are based on the instance name)
1632 By default, all optional items are activated except the LV name
1633 randomization. When passing ``--no-randomization``, which disables the
1634 optional items (i.e. just the DRBD secrets and cluster public keys are
1635 randomized), the resulting file can be used as a safety copy of the
1636 cluster config - while not trivial, the layout of the cluster can be
1637 recreated from it and if the instance disks have not been lost it
1638 permits recovery from the loss of all master candidates.
1643 See :doc:`separate documentation for move-instance <move-instance>`.
1645 .. TODO: document cluster-merge tool
1648 Other Ganeti projects
1649 ---------------------
1651 Below is a list (which might not be up-to-date) of additional projects
1652 that can be useful in a Ganeti deployment. They can be downloaded from
1653 the project site (http://code.google.com/p/ganeti/) and the repositories
1654 are also on the project git site (http://git.ganeti.org).
1659 The ``ganeti-nbma`` software is designed to allow instances to live on a
1660 separate, virtual network from the nodes, and in an environment where
1661 nodes are not guaranteed to be able to reach each other via multicasting
1662 or broadcasting. For more information see the README in the source
1668 Before Ganeti version 2.5, this was a standalone project; since that
1669 version it is integrated into the Ganeti codebase (see
1670 :doc:`install-quick` for instructions on how to enable it). If you run
1671 an older Ganeti version, you will have to download and build it
1674 For more information and installation instructions, see the README file
1675 in the source archive.
1677 .. vim: set textwidth=72 :