1 Ganeti administrator's guide
2 ============================
4 Documents Ganeti version |version|
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
76 - the :command:`ganeti-rapi` daemon which runs on the master node and
77 offers an HTTP-based API for the cluster
78 - the :command:`ganeti-masterd` daemon which runs on the master node and
79 allows control of the cluster
81 Beside the node role, there are other node flags that influence its
84 - the *master_capable* flag denotes whether the node can ever become a
85 master candidate; setting this to 'no' means that auto-promotion will
86 never make this node a master candidate; this flag can be useful for a
87 remote node that only runs local instances, and having it become a
88 master is impractical due to networking or other constraints
89 - the *vm_capable* flag denotes whether the node can host instances or
90 not; for example, one might use a non-vm_capable node just as a master
91 candidate, for configuration backups; setting this flag to no
92 disallows placement of instances of this node, deactivates hypervisor
93 and related checks on it (e.g. bridge checks, LVM check, etc.), and
94 removes it from cluster capacity computations
100 A virtual machine which runs on a cluster. It can be a fault tolerant,
101 highly available entity.
103 An instance has various parameters, which are classified in three
104 categories: hypervisor related-parameters (called ``hvparams``), general
105 parameters (called ``beparams``) and per network-card parameters (called
106 ``nicparams``). All these parameters can be modified either at instance
107 level or via defaults at cluster level.
112 The are multiple options for the storage provided to an instance; while
113 the instance sees the same virtual drive in all cases, the node-level
114 configuration varies between them.
116 There are four disk templates you can choose from:
119 The instance has no disks. Only used for special purpose operating
120 systems or for testing.
123 The instance will use plain files as backend for its disks. No
124 redundancy is provided, and this is somewhat more difficult to
125 configure for high performance.
128 The instance will use LVM devices as backend for its disks. No
129 redundancy is provided.
132 .. note:: This is only valid for multi-node clusters using DRBD 8.0+
134 A mirror is set between the local node and a remote one, which must be
135 specified with the second value of the --node option. Use this option
136 to obtain a highly available instance that can be failed over to a
137 remote node should the primary one fail.
142 A framework for using external (user-provided) scripts to compute the
143 placement of instances on the cluster nodes. This eliminates the need to
144 manually specify nodes in instance add, instance moves, node evacuate,
147 In order for Ganeti to be able to use these scripts, they must be place
148 in the iallocator directory (usually ``lib/ganeti/iallocators`` under
149 the installation prefix, e.g. ``/usr/local``).
151 “Primary” and “secondary” concepts
152 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
154 An instance has a primary and depending on the disk configuration, might
155 also have a secondary node. The instance always runs on the primary node
156 and only uses its secondary node for disk replication.
158 Similarly, the term of primary and secondary instances when talking
159 about a node refers to the set of instances having the given node as
160 primary, respectively secondary.
165 Tags are short strings that can be attached to either to cluster itself,
166 or to nodes or instances. They are useful as a very simplistic
167 information store for helping with cluster administration, for example
168 by attaching owner information to each instance after it's created::
170 gnt-instance add … instance1
171 gnt-instance add-tags instance1 owner:user2
173 And then by listing each instance and its tags, this information could
174 be used for contacting the users of each instance.
179 While not directly visible by an end-user, it's useful to know that a
180 basic cluster operation (e.g. starting an instance) is represented
181 internall by Ganeti as an *OpCode* (abbreviation from operation
182 code). These OpCodes are executed as part of a *Job*. The OpCodes in a
183 single Job are processed serially by Ganeti, but different Jobs will be
184 processed (depending on resource availability) in parallel.
186 For example, shutting down the entire cluster can be done by running the
187 command ``gnt-instance shutdown --all``, which will submit for each
188 instance a separate job containing the “shutdown instance” OpCode.
194 You need to have your Ganeti cluster installed and configured before you
195 try any of the commands in this document. Please follow the
196 :doc:`install` for instructions on how to do that.
204 The add operation might seem complex due to the many parameters it
205 accepts, but once you have understood the (few) required parameters and
206 the customisation capabilities you will see it is an easy operation.
208 The add operation requires at minimum five parameters:
210 - the OS for the instance
212 - the disk count and size
213 - the node specification or alternatively the iallocator to use
214 - and finally the instance name
216 The OS for the instance must be visible in the output of the command
217 ``gnt-os list`` and specifies which guest OS to install on the instance.
219 The disk template specifies what kind of storage to use as backend for
220 the (virtual) disks presented to the instance; note that for instances
221 with multiple virtual disks, they all must be of the same type.
223 The node(s) on which the instance will run can be given either manually,
224 via the ``-n`` option, or computed automatically by Ganeti, if you have
225 installed any iallocator script.
227 With the above parameters in mind, the command is::
230 -n TARGET_NODE:SECONDARY_NODE \
232 -t DISK_TEMPLATE -s DISK_SIZE \
235 The instance name must be resolvable (e.g. exist in DNS) and usually
236 points to an address in the same subnet as the cluster itself.
238 The above command has the minimum required options; other options you
239 can give include, among others:
241 - The memory size (``-B memory``)
243 - The number of virtual CPUs (``-B vcpus``)
245 - Arguments for the NICs of the instance; by default, a single-NIC
246 instance is created. The IP and/or bridge of the NIC can be changed
247 via ``--nic 0:ip=IP,bridge=BRIDGE``
249 See the manpage for gnt-instance for the detailed option list.
251 For example if you want to create an highly available instance, with a
252 single disk of 50GB and the default memory size, having primary node
253 ``node1`` and secondary node ``node3``, use the following command::
255 gnt-instance add -n node1:node3 -o debootstrap -t drbd \
258 There is a also a command for batch instance creation from a
259 specification file, see the ``batch-create`` operation in the
260 gnt-instance manual page.
262 Regular instance operations
263 +++++++++++++++++++++++++++
268 Removing an instance is even easier than creating one. This operation is
269 irreversible and destroys all the contents of your instance. Use with
272 gnt-instance remove INSTANCE_NAME
277 Instances are automatically started at instance creation time. To
278 manually start one which is currently stopped you can run::
280 gnt-instance startup INSTANCE_NAME
282 While the command to stop one is::
284 gnt-instance shutdown INSTANCE_NAME
286 .. warning:: Do not use the Xen or KVM commands directly to stop
287 instances. If you run for example ``xm shutdown`` or ``xm destroy``
288 on an instance Ganeti will automatically restart it (via the
289 :command:`ganeti-watcher` command which is launched via cron).
294 There are two ways to get information about instances: listing
295 instances, which does a tabular output containing a given set of fields
296 about each instance, and querying detailed information about a set of
299 The command to see all the instances configured and their status is::
303 The command can return a custom set of information when using the ``-o``
304 option (as always, check the manpage for a detailed specification). Each
305 instance will be represented on a line, thus making it easy to parse
306 this output via the usual shell utilities (grep, sed, etc.).
308 To get more detailed information about an instance, you can run::
310 gnt-instance info INSTANCE
312 which will give a multi-line block of information about the instance,
313 it's hardware resources (especially its disks and their redundancy
314 status), etc. This is harder to parse and is more expensive than the
315 list operation, but returns much more detailed information.
321 You can create a snapshot of an instance disk and its Ganeti
322 configuration, which then you can backup, or import into another
323 cluster. The way to export an instance is::
325 gnt-backup export -n TARGET_NODE INSTANCE_NAME
328 The target node can be any node in the cluster with enough space under
329 ``/srv/ganeti`` to hold the instance image. Use the ``--noshutdown``
330 option to snapshot an instance without rebooting it. Note that Ganeti
331 only keeps one snapshot for an instance - any previous snapshot of the
332 same instance existing cluster-wide under ``/srv/ganeti`` will be
333 removed by this operation: if you want to keep them, you need to move
334 them out of the Ganeti exports directory.
336 Importing an instance is similar to creating a new one, but additionally
337 one must specify the location of the snapshot. The command is::
339 gnt-backup import -n TARGET_NODE \
340 --src-node=NODE --src-dir=DIR INSTANCE_NAME
342 By default, parameters will be read from the export information, but you
343 can of course pass them in via the command line - most of the options
344 available for the command :command:`gnt-instance add` are supported here
347 Import of foreign instances
348 +++++++++++++++++++++++++++
350 There is a possibility to import a foreign instance whose disk data is
351 already stored as LVM volumes without going through copying it: the disk
354 For this, ensure that the original, non-managed instance is stopped,
355 then create a Ganeti instance in the usual way, except that instead of
356 passing the disk information you specify the current volumes::
358 gnt-instance add -t plain -n HOME_NODE ... \
359 --disk 0:adopt=lv_name[,vg=vg_name] INSTANCE_NAME
361 This will take over the given logical volumes, rename them to the Ganeti
362 standard (UUID-based), and without installing the OS on them start
363 directly the instance. If you configure the hypervisor similar to the
364 non-managed configuration that the instance had, the transition should
365 be seamless for the instance. For more than one disk, just pass another
366 disk parameter (e.g. ``--disk 1:adopt=...``).
371 .. note:: This section only applies to multi-node clusters
373 .. _instance-change-primary-label:
375 Changing the primary node
376 +++++++++++++++++++++++++
378 There are three ways to exchange an instance's primary and secondary
379 nodes; the right one to choose depends on how the instance has been
380 created and the status of its current primary node. See
381 :ref:`rest-redundancy-label` for information on changing the secondary
382 node. Note that it's only possible to change the primary node to the
383 secondary and vice-versa; a direct change of the primary node with a
384 third node, while keeping the current secondary is not possible in a
385 single step, only via multiple operations as detailed in
386 :ref:`instance-relocation-label`.
388 Failing over an instance
389 ~~~~~~~~~~~~~~~~~~~~~~~~
391 If an instance is built in highly available mode you can at any time
392 fail it over to its secondary node, even if the primary has somehow
393 failed and it's not up anymore. Doing it is really easy, on the master
394 node you can just run::
396 gnt-instance failover INSTANCE_NAME
398 That's it. After the command completes the secondary node is now the
399 primary, and vice-versa.
401 Live migrating an instance
402 ~~~~~~~~~~~~~~~~~~~~~~~~~~
404 If an instance is built in highly available mode, it currently runs and
405 both its nodes are running fine, you can at migrate it over to its
406 secondary node, without downtime. On the master node you need to run::
408 gnt-instance migrate INSTANCE_NAME
410 The current load on the instance and its memory size will influence how
411 long the migration will take. In any case, for both KVM and Xen
412 hypervisors, the migration will be transparent to the instance.
414 Moving an instance (offline)
415 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
417 If an instance has not been create as mirrored, then the only way to
418 change its primary node is to execute the move command::
420 gnt-instance move -n NEW_NODE INSTANCE
422 This has a few prerequisites:
424 - the instance must be stopped
425 - its current primary node must be on-line and healthy
426 - the disks of the instance must not have any errors
428 Since this operation actually copies the data from the old node to the
429 new node, expect it to take proportional to the size of the instance's
430 disks and the speed of both the nodes' I/O system and their networking.
435 Disk failures are a common cause of errors in any server
436 deployment. Ganeti offers protection from single-node failure if your
437 instances were created in HA mode, and it also offers ways to restore
438 redundancy after a failure.
440 Preparing for disk operations
441 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
443 It is important to note that for Ganeti to be able to do any disk
444 operation, the Linux machines on top of which Ganeti must be consistent;
445 for LVM, this means that the LVM commands must not return failures; it
446 is common that after a complete disk failure, any LVM command aborts
447 with an error similar to::
450 /dev/sdb1: read failed after 0 of 4096 at 0: Input/output error
451 /dev/sdb1: read failed after 0 of 4096 at 750153695232: Input/output
453 /dev/sdb1: read failed after 0 of 4096 at 0: Input/output error
454 Couldn't find device with uuid
455 't30jmN-4Rcf-Fr5e-CURS-pawt-z0jU-m1TgeJ'.
456 Couldn't find all physical volumes for volume group xenvg.
458 Before restoring an instance's disks to healthy status, it's needed to
459 fix the volume group used by Ganeti so that we can actually create and
460 manage the logical volumes. This is usually done in a multi-step
463 #. first, if the disk is completely gone and LVM commands exit with
464 “Couldn't find device with uuid…” then you need to run the command::
466 vgreduce --removemissing VOLUME_GROUP
468 #. after the above command, the LVM commands should be executing
469 normally (warnings are normal, but the commands will not fail
472 #. if the failed disk is still visible in the output of the ``pvs``
473 command, you need to deactivate it from allocations by running::
477 At this point, the volume group should be consistent and any bad
478 physical volumes should not longer be available for allocation.
480 Note that since version 2.1 Ganeti provides some commands to automate
481 these two operations, see :ref:`storage-units-label`.
483 .. _rest-redundancy-label:
485 Restoring redundancy for DRBD-based instances
486 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
488 A DRBD instance has two nodes, and the storage on one of them has
489 failed. Depending on which node (primary or secondary) has failed, you
490 have three options at hand:
492 - if the storage on the primary node has failed, you need to re-create
494 - if the storage on the secondary node has failed, you can either
495 re-create the disks on it or change the secondary and recreate
496 redundancy on the new secondary node
498 Of course, at any point it's possible to force re-creation of disks even
499 though everything is already fine.
501 For all three cases, the ``replace-disks`` operation can be used::
503 # re-create disks on the primary node
504 gnt-instance replace-disks -p INSTANCE_NAME
505 # re-create disks on the current secondary
506 gnt-instance replace-disks -s INSTANCE_NAME
507 # change the secondary node, via manual specification
508 gnt-instance replace-disks -n NODE INSTANCE_NAME
509 # change the secondary node, via an iallocator script
510 gnt-instance replace-disks -I SCRIPT INSTANCE_NAME
511 # since Ganeti 2.1: automatically fix the primary or secondary node
512 gnt-instance replace-disks -a INSTANCE_NAME
514 Since the process involves copying all data from the working node to the
515 target node, it will take a while, depending on the instance's disk
516 size, node I/O system and network speed. But it is (baring any network
517 interruption) completely transparent for the instance.
519 Re-creating disks for non-redundant instances
520 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
522 .. versionadded:: 2.1
524 For non-redundant instances, there isn't a copy (except backups) to
525 re-create the disks. But it's possible to at-least re-create empty
526 disks, after which a reinstall can be run, via the ``recreate-disks``
529 gnt-instance recreate-disks INSTANCE
531 Note that this will fail if the disks already exists.
533 Conversion of an instance's disk type
534 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
536 It is possible to convert between a non-redundant instance of type
537 ``plain`` (LVM storage) and redundant ``drbd`` via the ``gnt-instance
540 # start with a non-redundant instance
541 gnt-instance add -t plain ... INSTANCE
543 # later convert it to redundant
544 gnt-instance stop INSTANCE
545 gnt-instance modify -t drbd -n NEW_SECONDARY INSTANCE
546 gnt-instance start INSTANCE
548 # and convert it back
549 gnt-instance stop INSTANCE
550 gnt-instance modify -t plain INSTANCE
551 gnt-instance start INSTANCE
553 The conversion must be done while the instance is stopped, and
554 converting from plain to drbd template presents a small risk, especially
555 if the instance has multiple disks and/or if one node fails during the
556 conversion procedure). As such, it's recommended (as always) to make
557 sure that downtime for manual recovery is acceptable and that the
558 instance has up-to-date backups.
563 Accessing an instance's disks
564 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
566 From an instance's primary node you can have access to its disks. Never
567 ever mount the underlying logical volume manually on a fault tolerant
568 instance, or will break replication and your data will be
569 inconsistent. The correct way to access an instance's disks is to run
570 (on the master node, as usual) the command::
572 gnt-instance activate-disks INSTANCE
574 And then, *on the primary node of the instance*, access the device that
575 gets created. For example, you could mount the given disks, then edit
576 files on the filesystem, etc.
578 Note that with partitioned disks (as opposed to whole-disk filesystems),
579 you will need to use a tool like :manpage:`kpartx(8)`::
581 node1# gnt-instance activate-disks instance1
584 node3# kpartx -l /dev/…
585 node3# kpartx -a /dev/…
586 node3# mount /dev/mapper/… /mnt/
587 # edit files under mnt as desired
589 node3# kpartx -d /dev/…
593 After you've finished you can deactivate them with the deactivate-disks
594 command, which works in the same way::
596 gnt-instance deactivate-disks INSTANCE
598 Note that if any process started by you is still using the disks, the
599 above command will error out, and you **must** cleanup and ensure that
600 the above command runs successfully before you start the instance,
601 otherwise the instance will suffer corruption.
603 Accessing an instance's console
604 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
606 The command to access a running instance's console is::
608 gnt-instance console INSTANCE_NAME
610 Use the console normally and then type ``^]`` when done, to exit.
612 Other instance operations
613 +++++++++++++++++++++++++
618 There is a wrapper command for rebooting instances::
620 gnt-instance reboot instance2
622 By default, this does the equivalent of shutting down and then starting
623 the instance, but it accepts parameters to perform a soft-reboot (via
624 the hypervisor), a hard reboot (hypervisor shutdown and then startup) or
625 a full one (the default, which also de-configures and then configures
626 again the disks of the instance).
628 Instance OS definitions debugging
629 +++++++++++++++++++++++++++++++++
631 Should you have any problems with instance operating systems the command
632 to see a complete status for all your nodes is::
636 .. _instance-relocation-label:
641 While it is not possible to move an instance from nodes ``(A, B)`` to
642 nodes ``(C, D)`` in a single move, it is possible to do so in a few
645 # instance is located on A, B
646 node1# gnt-instance replace -n nodeC instance1
647 # instance has moved from (A, B) to (A, C)
648 # we now flip the primary/secondary nodes
649 node1# gnt-instance migrate instance1
650 # instance lives on (C, A)
651 # we can then change A to D via:
652 node1# gnt-instance replace -n nodeD instance1
654 Which brings it into the final configuration of ``(C, D)``. Note that we
655 needed to do two replace-disks operation (two copies of the instance
656 disks), because we needed to get rid of both the original nodes (A and
662 There are much fewer node operations available than for instances, but
663 they are equivalently important for maintaining a healthy cluster.
668 It is at any time possible to extend the cluster with one more node, by
669 using the node add operation::
671 gnt-node add NEW_NODE
673 If the cluster has a replication network defined, then you need to pass
674 the ``-s REPLICATION_IP`` parameter to this option.
676 A variation of this command can be used to re-configure a node if its
677 Ganeti configuration is broken, for example if it has been reinstalled
680 gnt-node add --readd EXISTING_NODE
682 This will reinitialise the node as if it's been newly added, but while
683 keeping its existing configuration in the cluster (primary/secondary IP,
684 etc.), in other words you won't need to use ``-s`` here.
686 Changing the node role
687 ++++++++++++++++++++++
689 A node can be in different roles, as explained in the
690 :ref:`terminology-label` section. Promoting a node to the master role is
691 special, while the other roles are handled all via a single command.
693 Failing over the master node
694 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
696 If you want to promote a different node to the master role (for whatever
697 reason), run on any other master-candidate node the command::
699 gnt-cluster master-failover
701 and the node you ran it on is now the new master. In case you try to run
702 this on a non master-candidate node, you will get an error telling you
703 which nodes are valid.
705 Changing between the other roles
706 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
708 The ``gnt-node modify`` command can be used to select a new role::
710 # change to master candidate
711 gnt-node modify -C yes NODE
712 # change to drained status
713 gnt-node modify -D yes NODE
714 # change to offline status
715 gnt-node modify -O yes NODE
716 # change to regular mode (reset all flags)
717 gnt-node modify -O no -D no -C no NODE
719 Note that the cluster requires that at any point in time, a certain
720 number of nodes are master candidates, so changing from master candidate
721 to other roles might fail. It is recommended to either force the
722 operation (via the ``--force`` option) or first change the number of
723 master candidates in the cluster - see :ref:`cluster-config-label`.
728 There are two steps of moving instances off a node:
730 - moving the primary instances (actually converting them into secondary
732 - moving the secondary instances (including any instances converted in
735 Primary instance conversion
736 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
738 For this step, you can use either individual instance move
739 commands (as seen in :ref:`instance-change-primary-label`) or the bulk
740 per-node versions; these are::
742 gnt-node migrate NODE
743 gnt-node evacuate NODE
745 Note that the instance “move” command doesn't currently have a node
748 Both these commands, or the equivalent per-instance command, will make
749 this node the secondary node for the respective instances, whereas their
750 current secondary node will become primary. Note that it is not possible
751 to change in one step the primary node to another node as primary, while
752 keeping the same secondary node.
754 Secondary instance evacuation
755 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
757 For the evacuation of secondary instances, a command called
758 :command:`gnt-node evacuate` is provided and its syntax is::
760 gnt-node evacuate -I IALLOCATOR_SCRIPT NODE
761 gnt-node evacuate -n DESTINATION_NODE NODE
763 The first version will compute the new secondary for each instance in
764 turn using the given iallocator script, whereas the second one will
765 simply move all instances to DESTINATION_NODE.
770 Once a node no longer has any instances (neither primary nor secondary),
771 it's easy to remove it from the cluster::
773 gnt-node remove NODE_NAME
775 This will deconfigure the node, stop the ganeti daemons on it and leave
776 it hopefully like before it joined to the cluster.
781 When using LVM (either standalone or with DRBD), it can become tedious
782 to debug and fix it in case of errors. Furthermore, even file-based
783 storage can become complicated to handle manually on many hosts. Ganeti
784 provides a couple of commands to help with automation.
789 This is a command specific to LVM handling. It allows listing the
790 logical volumes on a given node or on all nodes and their association to
791 instances via the ``volumes`` command::
793 node1# gnt-node volumes
794 Node PhysDev VG Name Size Instance
795 node1 /dev/sdb1 xenvg e61fbc97-….disk0 512M instance17
796 node1 /dev/sdb1 xenvg ebd1a7d1-….disk0 512M instance19
797 node2 /dev/sdb1 xenvg 0af08a3d-….disk0 512M instance20
798 node2 /dev/sdb1 xenvg cc012285-….disk0 512M instance16
799 node2 /dev/sdb1 xenvg f0fac192-….disk0 512M instance18
801 The above command maps each logical volume to a volume group and
802 underlying physical volume and (possibly) to an instance.
804 .. _storage-units-label:
806 Generalized storage handling
807 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
809 .. versionadded:: 2.1
811 Starting with Ganeti 2.1, a new storage framework has been implemented
812 that tries to abstract the handling of the storage type the cluster
815 First is listing the backend storage and their space situation::
817 node1# gnt-node list-storage
818 Node Name Size Used Free
819 node1 /dev/sda7 673.8G 0M 673.8G
820 node1 /dev/sdb1 698.6G 1.5G 697.1G
821 node2 /dev/sda7 673.8G 0M 673.8G
822 node2 /dev/sdb1 698.6G 1.0G 697.6G
824 The default is to list LVM physical volumes. It's also possible to list
825 the LVM volume groups::
827 node1# gnt-node list-storage -t lvm-vg
832 Next is repairing storage units, which is currently only implemented for
833 volume groups and does the equivalent of ``vgreduce --removemissing``::
835 node1# gnt-node repair-storage node2 lvm-vg xenvg
836 Sun Oct 25 22:21:45 2009 Repairing storage unit 'xenvg' on node2 ...
838 Last is the modification of volume properties, which is (again) only
839 implemented for LVM physical volumes and allows toggling the
840 ``allocatable`` value::
842 node1# gnt-node modify-storage --allocatable=no node2 lvm-pv /dev/sdb1
844 Use of the storage commands
845 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
847 All these commands are needed when recovering a node from a disk
850 - first, we need to recover from complete LVM failure (due to missing
851 disk), by running the ``repair-storage`` command
852 - second, we need to change allocation on any partially-broken disk
853 (i.e. LVM still sees it, but it has bad blocks) by running
855 - then we can evacuate the instances as needed
861 Beside the cluster initialisation command (which is detailed in the
862 :doc:`install` document) and the master failover command which is
863 explained under node handling, there are a couple of other cluster
864 operations available.
866 .. _cluster-config-label:
871 One of the few commands that can be run on any node (not only the
872 master) is the ``getmaster`` command::
874 node2# gnt-cluster getmaster
878 It is possible to query and change global cluster parameters via the
879 ``info`` and ``modify`` commands::
881 node1# gnt-cluster info
882 Cluster name: cluster.example.com
883 Cluster UUID: 07805e6f-f0af-4310-95f1-572862ee939c
884 Creation time: 2009-09-25 05:04:15
885 Modification time: 2009-10-18 22:11:47
886 Master node: node1.example.com
887 Architecture (this node): 64bit (x86_64)
890 Default hypervisor: xen-pvm
891 Enabled hypervisors: xen-pvm
892 Hypervisor parameters:
897 - candidate pool size: 10
899 Default instance parameters:
903 Default nic parameters:
908 There various parameters above can be changed via the ``modify``
911 - the hypervisor parameters can be changed via ``modify -H
912 xen-pvm:root_path=…``, and so on for other hypervisors/key/values
913 - the "default instance parameters" are changeable via ``modify -B
914 parameter=value…`` syntax
915 - the cluster parameters are changeable via separate options to the
916 modify command (e.g. ``--candidate-pool-size``, etc.)
918 For detailed option list see the :manpage:`gnt-cluster(8)` man page.
920 The cluster version can be obtained via the ``version`` command::
921 node1# gnt-cluster version
922 Software version: 2.1.0
923 Internode protocol: 20
924 Configuration format: 2010000
928 This is not very useful except when debugging Ganeti.
933 There are two commands provided for replicating files to all nodes of a
934 cluster and for running commands on all the nodes::
936 node1# gnt-cluster copyfile /path/to/file
937 node1# gnt-cluster command ls -l /path/to/file
939 These are simple wrappers over scp/ssh and more advanced usage can be
940 obtained using :manpage:`dsh(1)` and similar commands. But they are
941 useful to update an OS script from the master node, for example.
946 There are three commands that relate to global cluster checks. The first
947 one is ``verify`` which gives an overview on the cluster state,
948 highlighting any issues. In normal operation, this command should return
949 no ``ERROR`` messages::
951 node1# gnt-cluster verify
952 Sun Oct 25 23:08:58 2009 * Verifying global settings
953 Sun Oct 25 23:08:58 2009 * Gathering data (2 nodes)
954 Sun Oct 25 23:09:00 2009 * Verifying node status
955 Sun Oct 25 23:09:00 2009 * Verifying instance status
956 Sun Oct 25 23:09:00 2009 * Verifying orphan volumes
957 Sun Oct 25 23:09:00 2009 * Verifying remaining instances
958 Sun Oct 25 23:09:00 2009 * Verifying N+1 Memory redundancy
959 Sun Oct 25 23:09:00 2009 * Other Notes
960 Sun Oct 25 23:09:00 2009 - NOTICE: 5 non-redundant instance(s) found.
961 Sun Oct 25 23:09:00 2009 * Hooks Results
963 The second command is ``verify-disks``, which checks that the instance's
964 disks have the correct status based on the desired instance state
967 node1# gnt-cluster verify-disks
969 Note that this command will show no output when disks are healthy.
971 The last command is used to repair any discrepancies in Ganeti's
972 recorded disk size and the actual disk size (disk size information is
973 needed for proper activation and growth of DRBD-based disks)::
975 node1# gnt-cluster repair-disk-sizes
976 Sun Oct 25 23:13:16 2009 - INFO: Disk 0 of instance instance1 has mismatched size, correcting: recorded 512, actual 2048
977 Sun Oct 25 23:13:17 2009 - WARNING: Invalid result from node node4, ignoring node results
979 The above shows one instance having wrong disk size, and a node which
980 returned invalid data, and thus we ignored all primary instances of that
983 Configuration redistribution
984 ++++++++++++++++++++++++++++
986 If the verify command complains about file mismatches between the master
987 and other nodes, due to some node problems or if you manually modified
988 configuration files, you can force an push of the master configuration
989 to all other nodes via the ``redist-conf`` command::
991 node1# gnt-cluster redist-conf
994 This command will be silent unless there are problems sending updates to
1001 It is possible to rename a cluster, or to change its IP address, via the
1002 ``rename`` command. If only the IP has changed, you need to pass the
1003 current name and Ganeti will realise its IP has changed::
1005 node1# gnt-cluster rename cluster.example.com
1006 This will rename the cluster to 'cluster.example.com'. If
1007 you are connected over the network to the cluster name, the operation
1008 is very dangerous as the IP address will be removed from the node and
1009 the change may not go through. Continue?
1011 Failure: prerequisites not met for this operation:
1012 Neither the name nor the IP address of the cluster has changed
1014 In the above output, neither value has changed since the cluster
1015 initialisation so the operation is not completed.
1020 The job queue execution in Ganeti 2.0 and higher can be inspected,
1021 suspended and resumed via the ``queue`` command::
1023 node1~# gnt-cluster queue info
1024 The drain flag is unset
1025 node1~# gnt-cluster queue drain
1026 node1~# gnt-instance stop instance1
1027 Failed to submit job for instance1: Job queue is drained, refusing job
1028 node1~# gnt-cluster queue info
1029 The drain flag is set
1030 node1~# gnt-cluster queue undrain
1032 This is most useful if you have an active cluster and you need to
1033 upgrade the Ganeti software, or simply restart the software on any node:
1035 #. suspend the queue via ``queue drain``
1036 #. wait until there are no more running jobs via ``gnt-job list``
1037 #. restart the master or another node, or upgrade the software
1038 #. resume the queue via ``queue undrain``
1040 .. note:: this command only stores a local flag file, and if you
1041 failover the master, it will not have effect on the new master.
1047 The :manpage:`ganeti-watcher` is a program, usually scheduled via
1048 ``cron``, that takes care of cluster maintenance operations (restarting
1049 downed instances, activating down DRBD disks, etc.). However, during
1050 maintenance and troubleshooting, this can get in your way; disabling it
1051 via commenting out the cron job is not so good as this can be
1052 forgotten. Thus there are some commands for automated control of the
1053 watcher: ``pause``, ``info`` and ``continue``::
1055 node1~# gnt-cluster watcher info
1056 The watcher is not paused.
1057 node1~# gnt-cluster watcher pause 1h
1058 The watcher is paused until Mon Oct 26 00:30:37 2009.
1059 node1~# gnt-cluster watcher info
1060 The watcher is paused until Mon Oct 26 00:30:37 2009.
1061 node1~# ganeti-watcher -d
1062 2009-10-25 23:30:47,984: pid=28867 ganeti-watcher:486 DEBUG Pause has been set, exiting
1063 node1~# gnt-cluster watcher continue
1064 The watcher is no longer paused.
1065 node1~# ganeti-watcher -d
1066 2009-10-25 23:31:04,789: pid=28976 ganeti-watcher:345 DEBUG Archived 0 jobs, left 0
1067 2009-10-25 23:31:05,884: pid=28976 ganeti-watcher:280 DEBUG Got data from cluster, writing instance status file
1068 2009-10-25 23:31:06,061: pid=28976 ganeti-watcher:150 DEBUG Data didn't change, just touching status file
1069 node1~# gnt-cluster watcher info
1070 The watcher is not paused.
1073 The exact details of the argument to the ``pause`` command are available
1076 .. note:: this command only stores a local flag file, and if you
1077 failover the master, it will not have effect on the new master.
1079 Node auto-maintenance
1080 +++++++++++++++++++++
1082 If the cluster parameter ``maintain_node_health`` is enabled (see the
1083 manpage for :command:`gnt-cluster`, the init and modify subcommands),
1084 then the following will happen automatically:
1086 - the watcher will shutdown any instances running on offline nodes
1087 - the watcher will deactivate any DRBD devices on offline nodes
1089 In the future, more actions are planned, so only enable this parameter
1090 if the nodes are completely dedicated to Ganeti; otherwise it might be
1091 possible to lose data due to auto-maintenance actions.
1093 Removing a cluster entirely
1094 +++++++++++++++++++++++++++
1096 The usual method to cleanup a cluster is to run ``gnt-cluster destroy``
1097 however if the Ganeti installation is broken in any way then this will
1100 It is possible in such a case to cleanup manually most if not all traces
1101 of a cluster installation by following these steps on all of the nodes:
1103 1. Shutdown all instances. This depends on the virtualisation method
1104 used (Xen, KVM, etc.):
1106 - Xen: run ``xm list`` and ``xm destroy`` on all the non-Domain-0
1108 - KVM: kill all the KVM processes
1109 - chroot: kill all processes under the chroot mountpoints
1111 2. If using DRBD, shutdown all DRBD minors (which should by at this time
1112 no-longer in use by instances); on each node, run ``drbdsetup
1113 /dev/drbdN down`` for each active DRBD minor.
1115 3. If using LVM, cleanup the Ganeti volume group; if only Ganeti created
1116 logical volumes (and you are not sharing the volume group with the
1117 OS, for example), then simply running ``lvremove -f xenvg`` (replace
1118 'xenvg' with your volume group name) should do the required cleanup.
1120 4. If using file-based storage, remove recursively all files and
1121 directories under your file-storage directory: ``rm -rf
1122 /srv/ganeti/file-storage/*`` replacing the path with the correct path
1125 5. Stop the ganeti daemons (``/etc/init.d/ganeti stop``) and kill any
1126 that remain alive (``pgrep ganeti`` and ``pkill ganeti``).
1128 6. Remove the ganeti state directory (``rm -rf /var/lib/ganeti/*``),
1129 replacing the path with the correct path for your installation.
1131 On the master node, remove the cluster from the master-netdev (usually
1132 ``xen-br0`` for bridged mode, otherwise ``eth0`` or similar), by running
1133 ``ip a del $clusterip/32 dev xen-br0`` (use the correct cluster ip and
1134 network device name).
1136 At this point, the machines are ready for a cluster creation; in case
1137 you want to remove Ganeti completely, you need to also undo some of the
1138 SSH changes and log directories:
1140 - ``rm -rf /var/log/ganeti /srv/ganeti`` (replace with the correct
1142 - remove from ``/root/.ssh`` the keys that Ganeti added (check the
1143 ``authorized_keys`` and ``id_dsa`` files)
1144 - regenerate the host's SSH keys (check the OpenSSH startup scripts)
1147 Otherwise, if you plan to re-create the cluster, you can just go ahead
1148 and rerun ``gnt-cluster init``.
1153 The tags handling (addition, removal, listing) is similar for all the
1154 objects that support it (instances, nodes, and the cluster).
1159 Note that the set of characters present in a tag and the maximum tag
1160 length are restricted. Currently the maximum length is 128 characters,
1161 there can be at most 4096 tags per object, and the set of characters is
1162 comprised by alphanumeric characters and additionally ``.+*/:@-``.
1167 Tags can be added via ``add-tags``::
1169 gnt-instance add-tags INSTANCE a b c
1170 gnt-node add-tags INSTANCE a b c
1171 gnt-cluster add-tags a b c
1174 The above commands add three tags to an instance, to a node and to the
1175 cluster. Note that the cluster command only takes tags as arguments,
1176 whereas the node and instance commands first required the node and
1179 Tags can also be added from a file, via the ``--from=FILENAME``
1180 argument. The file is expected to contain one tag per line.
1182 Tags can also be remove via a syntax very similar to the add one::
1184 gnt-instance remove-tags INSTANCE a b c
1188 gnt-instance list-tags
1190 gnt-cluster list-tags
1195 It is also possible to execute a global search on the all tags defined
1196 in the cluster configuration, via a cluster command::
1198 gnt-cluster search-tags REGEXP
1200 The parameter expected is a regular expression (see
1201 :manpage:`regex(7)`). This will return all tags that match the search,
1202 together with the object they are defined in (the names being show in a
1203 hierarchical kind of way)::
1205 node1# gnt-cluster search-tags o
1207 /instances/instance1 owner:bar
1213 The various jobs submitted by the instance/node/cluster commands can be
1214 examined, canceled and archived by various invocations of the
1215 ``gnt-job`` command.
1217 First is the job list command::
1220 17771 success INSTANCE_QUERY_DATA
1221 17773 success CLUSTER_VERIFY_DISKS
1222 17775 success CLUSTER_REPAIR_DISK_SIZES
1223 17776 error CLUSTER_RENAME(cluster.example.com)
1224 17780 success CLUSTER_REDIST_CONF
1225 17792 success INSTANCE_REBOOT(instance1.example.com)
1227 More detailed information about a job can be found via the ``info``
1230 node1# gnt-job info 17776
1233 Received: 2009-10-25 23:18:02.180569
1234 Processing start: 2009-10-25 23:18:02.200335 (delta 0.019766s)
1235 Processing end: 2009-10-25 23:18:02.279743 (delta 0.079408s)
1236 Total processing time: 0.099174 seconds
1240 Processing start: 2009-10-25 23:18:02.200335
1241 Processing end: 2009-10-25 23:18:02.252282
1243 name: cluster.example.com
1246 [Neither the name nor the IP address of the cluster has changed]
1249 During the execution of a job, it's possible to follow the output of a
1250 job, similar to the log that one get from the ``gnt-`` commands, via the
1253 node1# gnt-instance add --submit … instance1
1255 node1# gnt-job watch 17818
1256 Output from job 17818 follows
1257 -----------------------------
1258 Mon Oct 26 00:22:48 2009 - INFO: Selected nodes for instance instance1 via iallocator dumb: node1, node2
1259 Mon Oct 26 00:22:49 2009 * creating instance disks...
1260 Mon Oct 26 00:22:52 2009 adding instance instance1 to cluster config
1261 Mon Oct 26 00:22:52 2009 - INFO: Waiting for instance instance1 to sync disks.
1263 Mon Oct 26 00:23:03 2009 creating os for instance instance1 on node node1
1264 Mon Oct 26 00:23:03 2009 * running the instance OS create scripts...
1265 Mon Oct 26 00:23:13 2009 * starting instance...
1268 This is useful if you need to follow a job's progress from multiple
1271 A job that has not yet started to run can be canceled::
1273 node1# gnt-job cancel 17810
1275 But not one that has already started execution::
1277 node1# gnt-job cancel 17805
1278 Job 17805 is no longer waiting in the queue
1280 There are two queues for jobs: the *current* and the *archive*
1281 queue. Jobs are initially submitted to the current queue, and they stay
1282 in that queue until they have finished execution (either successfully or
1283 not). At that point, they can be moved into the archive queue, and the
1284 ganeti-watcher script will do this automatically after 6 hours. The
1285 ganeti-cleaner script will remove the jobs from the archive directory
1288 Note that only jobs in the current queue can be viewed via the list and
1289 info commands; Ganeti itself doesn't examine the archive directory. If
1290 you need to see an older job, either move the file manually in the
1291 top-level queue directory, or look at its contents (it's a
1292 JSON-formatted file).
1294 Special Ganeti deployments
1295 --------------------------
1297 Since Ganeti 2.4, it is possible to extend the Ganeti deployment with
1298 two custom scenarios: Ganeti inside Ganeti and multi-site model.
1300 Running Ganeti under Ganeti
1301 +++++++++++++++++++++++++++
1303 It is sometimes useful to be able to use a Ganeti instance as a Ganeti
1304 node (part of another cluster, usually). One example scenario is two
1305 small clusters, where we want to have an additional master candidate
1306 that holds the cluster configuration and can be used for helping with
1307 the master voting process.
1309 However, these Ganeti instance should not host instances themselves, and
1310 should not be considered in the normal capacity planning, evacuation
1311 strategies, etc. In order to accomplish this, mark these nodes as
1312 non-``vm_capable``::
1314 node1# gnt-node modify --vm-capable=no node3
1316 The vm_capable status can be listed as usual via ``gnt-node list``::
1318 node1# gnt-node list -oname,vm_capable
1324 When this flag is set, the cluster will not do any operations that
1325 relate to instances on such nodes, e.g. hypervisor operations,
1326 disk-related operations, etc. Basically they will just keep the ssconf
1327 files, and if master candidates the full configuration.
1332 If Ganeti is deployed in multi-site model, with each site being a node
1333 group (so that instances are not relocated across the WAN by mistake),
1334 it is conceivable that either the WAN latency is high or that some sites
1335 have a lower reliability than others. In this case, it doesn't make
1336 sense to replicate the job information across all sites (or even outside
1337 of a “central” node group), so it should be possible to restrict which
1338 nodes can become master candidates via the auto-promotion algorithm.
1340 Ganeti 2.4 introduces for this purpose a new ``master_capable`` flag,
1341 which (when unset) prevents nodes from being marked as master
1342 candidates, either manually or automatically.
1344 As usual, the node modify operation can change this flag::
1346 node1# gnt-node modify --auto-promote --master-capable=no node3
1347 Fri Jan 7 06:23:07 2011 - INFO: Demoting from master candidate
1348 Fri Jan 7 06:23:08 2011 - INFO: Promoted nodes to master candidate role: node4
1350 - master_capable -> False
1351 - master_candidate -> False
1353 And the node list operation will list this flag::
1355 node1# gnt-node list -oname,master_capable node1 node2 node3
1361 Note that marking a node both not ``vm_capable`` and not
1362 ``master_capable`` makes the node practically unusable from Ganeti's
1363 point of view. Hence these two flags should be used probably in
1364 contrast: some nodes will be only master candidates (master_capable but
1365 not vm_capable), and other nodes will only hold instances (vm_capable
1366 but not master_capable).
1372 Beside the usual ``gnt-`` and ``ganeti-`` commands which are provided
1373 and installed in ``$prefix/sbin`` at install time, there are a couple of
1374 other tools installed which are used seldom but can be helpful in some
1380 The ``lvmstrap`` tool, introduced in :ref:`configure-lvm-label` section,
1381 has two modes of operation:
1383 - ``diskinfo`` shows the discovered disks on the system and their status
1384 - ``create`` takes all not-in-use disks and creates a volume group out
1387 .. warning:: The ``create`` argument to this command causes data-loss!
1392 The ``cfgupgrade`` tools is used to upgrade between major (and minor)
1393 Ganeti versions. Point-releases are usually transparent for the admin.
1395 More information about the upgrade procedure is listed on the wiki at
1396 http://code.google.com/p/ganeti/wiki/UpgradeNotes.
1398 There is also a script designed to upgrade from Ganeti 1.2 to 2.0,
1399 called ``cfgupgrade12``.
1404 .. note:: This command is not actively maintained; make sure you backup
1405 your configuration before using it
1407 This can be used as an alternative to direct editing of the
1408 main configuration file if Ganeti has a bug and prevents you, for
1409 example, from removing an instance or a node from the configuration
1417 .. warning:: This command will erase existing instances if given as
1420 This tool is used to exercise either the hardware of machines or
1421 alternatively the Ganeti software. It is safe to run on an existing
1422 cluster **as long as you don't pass it existing instance names**.
1424 The command will, by default, execute a comprehensive set of operations
1425 against a list of instances, these being:
1428 - disk replacement (for redundant instances)
1429 - failover and migration (for redundant instances)
1430 - move (for non-redundant instances)
1432 - add disks, remove disk
1433 - add NICs, remove NICs
1434 - export and then import
1438 - and finally removal of the test instances
1440 Executing all these operations will test that the hardware performs
1441 well: the creation, disk replace, disk add and disk growth will exercise
1442 the storage and network; the migrate command will test the memory of the
1443 systems. Depending on the passed options, it can also test that the
1444 instance OS definitions are executing properly the rename, import and
1450 This tool takes the Ganeti configuration and outputs a "sanitized"
1451 version, by randomizing or clearing:
1453 - DRBD secrets and cluster public key (always)
1454 - host names (optional)
1456 - OS names (optional)
1457 - LV names (optional, only useful for very old clusters which still have
1458 instances whose LVs are based on the instance name)
1460 By default, all optional items are activated except the LV name
1461 randomization. When passing ``--no-randomization``, which disables the
1462 optional items (i.e. just the DRBD secrets and cluster public keys are
1463 randomized), the resulting file can be used as a safety copy of the
1464 cluster config - while not trivial, the layout of the cluster can be
1465 recreated from it and if the instance disks have not been lost it
1466 permits recovery from the loss of all master candidates.
1471 See :doc:`separate documentation for move-instance <move-instance>`.
1473 .. TODO: document cluster-merge tool
1476 Other Ganeti projects
1477 ---------------------
1479 There are two other Ganeti-related projects that can be useful in a
1480 Ganeti deployment. These can be downloaded from the project site
1481 (http://code.google.com/p/ganeti/) and the repositories are also on the
1482 project git site (http://git.ganeti.org).
1487 The ``ganeti-nbma`` software is designed to allow instances to live on a
1488 separate, virtual network from the nodes, and in an environment where
1489 nodes are not guaranteed to be able to reach each other via multicasting
1490 or broadcasting. For more information see the README in the source
1496 The ``ganeti-htools`` software consists of a set of tools:
1498 - ``hail``: an advanced iallocator script compared to Ganeti's builtin
1500 - ``hbal``: a tool for rebalancing the cluster, i.e. moving instances
1501 around in order to better use the resources on the nodes
1502 - ``hspace``: a tool for estimating the available capacity of a cluster,
1503 so that capacity planning can be done efficiently
1505 For more information and installation instructions, see the README file
1506 in the source archive.
1508 .. vim: set textwidth=72 :