10 - The minimum Python version needed to run Ganeti is now 2.6.
11 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
12 to always require authentication, ``--require-authentication``. It can
13 be specified in ``$sysconfdir/default/ganeti``.
14 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
15 be used to manage the disk templates to be used by instances in the cluster.
16 Initially, it will be set to a list that includes plain, drbd, if they were
17 enabled by specifying a volume group name, and file and sharedfile, if those
18 were enabled at configure time. Additionally, it will include all disk
19 templates that are currently used by instances. The order of disk templates
20 will be based on Ganeti's history of supporting them. In the future, the
21 first entry of the list will be used as a default disk template on instance
23 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
24 configuration back to the previous stable version.
25 - hail now honors network restrictions when allocating nodes. This led to an
26 update of the IAllocator protocol. See the IAllocator documentation for
33 *(Released Fri, 24 May 2013)*
35 Incompatible/important changes
36 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
38 - Instance policies for disk size were documented to be on a per-disk
39 basis, but hail applied them to the sum of all disks. This has been
41 - ``hbal`` will now exit with status 0 if, during job execution over
42 LUXI, early exit has been requested and all jobs are successful;
43 before, exit status 1 was used, which cannot be differentiated from
45 - Compatibility with newer versions of rbd has been fixed
46 - ``gnt-instance batch-create`` has been changed to use the bulk create
47 opcode from Ganeti. This lead to incompatible changes in the format of
48 the JSON file. It's now not a custom dict anymore but a dict
49 compatible with the ``OpInstanceCreate`` opcode.
50 - Parent directories for file storage need to be listed in
51 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
52 write the file automatically based on old configuration values, but it
53 can not distribute it across all nodes and the file contents should be
54 verified. Use ``gnt-cluster copyfile
55 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
56 upgraded. The reason for requiring this list of paths now is that
57 before it would have been possible to inject new paths via RPC,
58 allowing files to be created in arbitrary locations. The RPC protocol
59 is protected using SSL/X.509 certificates, but as a design principle
60 Ganeti does not permit arbitrary paths to be passed.
61 - The parsing of the variants file for OSes (see
62 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
63 lines and comment lines (starting with ``#``) are ignored for better
65 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
66 longer available. ``gnt-node add`` now invokes a new tool on the
67 destination node, named ``prepare-node-join``, to configure the SSH
68 daemon. Paramiko is no longer necessary to configure nodes' SSH
69 daemons via ``gnt-node add``.
70 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
71 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
72 the flag is not reset after a master failover.
73 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
74 is recommended. 2.6 will be mandatory from the 2.8 series.
80 - New network management functionality to support automatic allocation
81 of IP addresses and managing of network parameters. See
82 :manpage:`gnt-network(8)` for more details.
83 - New external storage backend, to allow managing arbitrary storage
84 systems external to the cluster. See
85 :manpage:`ganeti-extstorage-interface(7)`.
86 - New ``exclusive-storage`` node parameter added, restricted to
87 nodegroup level. When it's set to true, physical disks are assigned in
88 an exclusive fashion to instances, as documented in :doc:`Partitioned
89 Ganeti <design-partitioned>`. Currently, only instances using the
90 ``plain`` disk template are supported.
91 - The KVM hypervisor has been updated with many new hypervisor
92 parameters, including a generic one for passing arbitrary command line
93 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
94 compatible up to qemu 1.4.
95 - A new tool, called ``mon-collector``, is the stand-alone executor of
96 the data collectors for a monitoring system. As of this version, it
97 just includes the DRBD data collector, that can be executed by calling
98 ``mon-collector`` using the ``drbd`` parameter. See
99 :manpage:`mon-collector(7)`.
100 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
101 for RAPI users. It allows granting permissions to query for
102 information to a specific user without giving
103 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
104 - A new tool named ``node-cleanup`` has been added. It cleans remains of
105 a cluster from a machine by stopping all daemons, removing
106 certificates and ssconf files. Unless the ``--no-backup`` option is
107 given, copies of the certificates are made.
108 - Instance creations now support the use of opportunistic locking,
109 potentially speeding up the (parallel) creation of multiple instances.
110 This feature is currently only available via the :doc:`RAPI
111 <rapi>` interface and when an instance allocator is used. If the
112 ``opportunistic_locking`` parameter is set the opcode will try to
113 acquire as many locks as possible, but will not wait for any locks
114 held by other opcodes. If not enough resources can be found to
115 allocate the instance, the temporary error code
116 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
117 retried thereafter, with or without opportunistic locking.
118 - New experimental linux-ha resource scripts.
119 - Restricted-commands support: ganeti can now be asked (via command line
120 or rapi) to perform commands on a node. These are passed via ganeti
121 RPC rather than ssh. This functionality is restricted to commands
122 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
123 reasons. The file is not copied automatically.
129 - Diskless instances are now externally mirrored (Issue 237). This for
130 now has only been tested in conjunction with explicit target nodes for
132 - Queries not needing locks or RPC access to the node can now be
133 performed by the confd daemon, making them independent from jobs, and
134 thus faster to execute. This is selectable at configure time.
135 - The functionality for allocating multiple instances at once has been
136 overhauled and is now also available through :doc:`RAPI <rapi>`.
140 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
142 - Disks added through ``gnt-instance modify`` or created through
143 ``gnt-instance recreate-disks`` are wiped, if the
144 ``prealloc_wipe_disks`` flag is set.
145 - If wiping newly created disks fails, the disks are removed. Also,
146 partial failures in creating disks through ``gnt-instance modify``
147 triggers a cleanup of the partially-created disks.
148 - Removing the master IP address doesn't fail if the address has been
150 - Fix ownership of the OS log dir
151 - Workaround missing SO_PEERCRED constant (Issue 191)
157 *(Released Fri, 3 May 2013)*
159 This was the first release candidate of the 2.7 series. Since beta3:
161 - Fix kvm compatibility with qemu 1.4 (Issue 389)
162 - Documentation updates (admin guide, upgrade notes, install
163 instructions) (Issue 372)
164 - Fix gnt-group list nodes and instances count (Issue 436)
165 - Fix compilation without non-mandatory libraries (Issue 441)
166 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
167 - Make confd logging more verbose at INFO level (Issue 435)
168 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
169 - Fix failure path for instance storage type conversion (Issue 229)
170 - Update htools text backend documentation
171 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
172 - Disable inter-cluster instance move for file-based instances, because
173 it is dependant on instance export, which is not supported for
174 file-based instances. (Issue 414)
175 - Fix gnt-job crashes on non-ascii characters (Issue 427)
176 - Fix volume group checks on non-vm-capable nodes (Issue 432)
182 *(Released Mon, 22 Apr 2013)*
184 This was the third beta release of the 2.7 series. Since beta2:
186 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
187 - Fix data loss on wrong usage of ``gnt-instance move``
188 - Properly export errors in confd-based job queries
189 - Add ``users-setup`` tool
190 - Fix iallocator protocol to report 0 as a disk size for diskless
191 instances. This avoids hail breaking when a diskless instance is
193 - Fix job queue directory permission problem that made confd job queries
194 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
195 for access to archived jobs (Issue 406).
196 - Limit the sizes of networks supported by ``gnt-network`` to something
197 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
198 - Fix bugs in instance disk template conversion
199 - Fix GHC 7 compatibility
200 - Fix ``burnin`` install path (Issue 426).
201 - Allow very small disk grows (Issue 347).
202 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
203 that noded doesn't import masterd code (Issue 419).
204 - Make sure the default metavg at cluster init is the same as the vg, if
205 unspecified (Issue 358).
206 - Fix cleanup of partially created disks (part of Issue 416)
212 *(Released Tue, 2 Apr 2013)*
214 This was the second beta release of the 2.7 series. Since beta1:
216 - Networks no longer have a "type" slot, since this information was
217 unused in Ganeti: instead of it tags should be used.
218 - The rapi client now has a ``target_node`` option to MigrateInstance.
219 - Fix early exit return code for hbal (Issue 386).
220 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
221 - Fix ``rbd showmapped`` output parsing (Issue 312).
222 - Networks are now referenced indexed by UUID, rather than name. This
223 will require running cfgupgrade, from 2.7.0beta1, if networks are in
225 - The OS environment now includes network information.
226 - Deleting of a network is now disallowed if any instance nic is using
227 it, to prevent dangling references.
228 - External storage is now documented in man pages.
229 - The exclusive_storage flag can now only be set at nodegroup level.
230 - Hbal can now submit an explicit priority with its jobs.
231 - Many network related locking fixes.
232 - Bump up the required pylint version to 0.25.1.
233 - Fix the ``no_remember`` option in RAPI client.
234 - Many ipolicy related tests, qa, and fixes.
235 - Many documentation improvements and fixes.
236 - Fix building with ``--disable-file-storage``.
237 - Fix ``-q`` option in htools, which was broken if passed more than
239 - Some haskell/python interaction improvements and fixes.
240 - Fix iallocator in case of missing LVM storage.
241 - Fix confd config load in case of ``--no-lvm-storage``.
242 - The confd/query functionality is now mentioned in the security
249 *(Released Wed, 6 Feb 2013)*
251 This was the first beta release of the 2.7 series. All important changes
252 are listed in the latest 2.7 entry.
258 *(Released Fri, 21 Dec 2012)*
260 Important behaviour change: hbal won't rebalance anymore instances which
261 have the ``auto_balance`` attribute set to false. This was the intention
262 all along, but until now it only skipped those from the N+1 memory
263 reservation (DRBD-specific).
265 A significant number of bug fixes in this release:
267 - Fixed disk adoption interaction with ipolicy checks.
268 - Fixed networking issues when instances are started, stopped or
269 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
270 - Fixed the warning in cluster verify for shared storage instances not
272 - Fixed removal of storage directory on shared file storage (issue 262).
273 - Fixed validation of LVM volume group name in OpClusterSetParams
274 (``gnt-cluster modify``) (issue 285).
275 - Fixed runtime memory increases (``gnt-instance modify -m``).
276 - Fixed live migration under Xen's ``xl`` mode.
277 - Fixed ``gnt-instance console`` with ``xl``.
278 - Fixed building with newer Haskell compiler/libraries.
279 - Fixed PID file writing in Haskell daemons (confd); this prevents
280 restart issues if confd was launched manually (outside of
281 ``daemon-util``) while another copy of it was running
282 - Fixed a type error when doing live migrations with KVM (issue 297) and
283 the error messages for failing migrations have been improved.
284 - Fixed opcode validation for the out-of-band commands (``gnt-node
286 - Fixed a type error when unsetting OS hypervisor parameters (issue
287 311); now it's possible to unset all OS-specific hypervisor
289 - Fixed the ``dry-run`` mode for many operations: verification of
290 results was over-zealous but didn't take into account the ``dry-run``
291 operation, resulting in "wrong" failures.
292 - Fixed bash completion in ``gnt-job list`` when the job queue has
293 hundreds of entries; especially with older ``bash`` versions, this
294 results in significant CPU usage.
296 And lastly, a few other improvements have been made:
298 - Added option to force master-failover without voting (issue 282).
299 - Clarified error message on lock conflict (issue 287).
300 - Logging of newly submitted jobs has been improved (issue 290).
301 - Hostname checks have been made uniform between instance rename and
303 - The ``--submit`` option is now supported by ``gnt-debug delay``.
304 - Shutting down the master daemon by sending SIGTERM now stops it from
305 processing jobs waiting for locks; instead, those jobs will be started
306 once again after the master daemon is started the next time (issue
308 - Support for Xen's ``xl`` program has been improved (besides the fixes
310 - Reduced logging noise in the Haskell confd daemon (only show one log
311 entry for each config reload, instead of two).
312 - Several man page updates and typo fixes.
318 *(Released Fri, 12 Oct 2012)*
320 A small bugfix release. Among the bugs fixed:
322 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
323 made the command unusable.
324 - Commands that issue many jobs don't fail anymore just because some jobs
325 take so long that other jobs are archived.
326 - Failures during ``gnt-instance reinstall`` are reflected by the exit
328 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
330 - When ``always_failover`` is set, ``--allow-failover`` is not required
331 in migrate commands anymore.
332 - ``bash_completion`` works even if extglob is disabled.
333 - Fixed bug with locks that made failover for RDB-based instances fail.
334 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
335 a random node instead of one based on the allocator metric.
336 - Support for newer versions of pylint and pep8.
337 - Hail doesn't fail anymore when trying to add an instance of type
338 ``file``, ``sharedfile`` or ``rbd``.
339 - Added new Makefile target to rebuild the whole distribution, so that
340 all files are included.
346 *(Released Fri, 27 Jul 2012)*
349 .. attention:: The ``LUXI`` protocol has been made more consistent
350 regarding its handling of command arguments. This, however, leads to
351 incompatibility issues with previous versions. Please ensure that you
352 restart Ganeti daemons soon after the upgrade, otherwise most
353 ``LUXI`` calls (job submission, setting/resetting the drain flag,
354 pausing/resuming the watcher, cancelling and archiving jobs, querying
355 the cluster configuration) will fail.
364 The current ``admin_up`` field, which used to denote whether an instance
365 should be running or not, has been removed. Instead, ``admin_state`` is
366 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
368 The rational behind this is that an instance being “down” can have
371 - it could be down during a reboot
372 - it could be temporarily be down for a reinstall
373 - or it could be down because it is deprecated and kept just for its
376 The previous Boolean state was making it difficult to do capacity
377 calculations: should Ganeti reserve memory for a down instance? Now, the
378 tri-state field makes it clear:
380 - in ``up`` and ``down`` state, all resources are reserved for the
381 instance, and it can be at any time brought up if it is down
382 - in ``offline`` state, only disk space is reserved for it, but not
385 The field can have an extra use: since the transition between ``up`` and
386 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
387 transition between ``offline`` and ``down`` is done via ``gnt-instance
388 modify``, it is possible to given different rights to users. For
389 example, owners of an instance could be allowed to start/stop it, but
390 not transition it out of the offline state.
392 Instance policies and specs
393 +++++++++++++++++++++++++++
395 In previous Ganeti versions, an instance creation request was not
396 limited on the minimum size and on the maximum size just by the cluster
397 resources. As such, any policy could be implemented only in third-party
398 clients (RAPI clients, or shell wrappers over ``gnt-*``
399 tools). Furthermore, calculating cluster capacity via ``hspace`` again
400 required external input with regards to instance sizes.
402 In order to improve these workflows and to allow for example better
403 per-node group differentiation, we introduced instance specs, which
406 - minimum instance disk size, disk count, memory size, cpu count
407 - maximum values for the above metrics
408 - and “standard” values (used in ``hspace`` to calculate the standard
411 The minimum/maximum values can be also customised at node-group level,
412 for example allowing more powerful hardware to support bigger instance
415 Beside the instance specs, there are a few other settings belonging to
416 the instance policy framework. It is possible now to customise, per
417 cluster and node-group:
419 - the list of allowed disk templates
420 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
421 - the maximum ratio of instance to spindles (see below for more
422 information) for local storage
424 All these together should allow all tools that talk to Ganeti to know
425 what are the ranges of allowed values for instances and the
426 over-subscription that is allowed.
428 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
429 instance configuration, and the physical CPU configuration from the
430 node. For the spindle ratios however, we didn't track before these
431 values, so new parameters have been added:
433 - a new node parameter ``spindle_count``, defaults to 1, customisable at
434 node group or node level
435 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
437 Note that spindles in this context doesn't need to mean actual
438 mechanical hard-drives; it's just a relative number for both the node
439 I/O capacity and instance I/O consumption.
441 Instance migration behaviour
442 ++++++++++++++++++++++++++++
444 While live-migration is in general desirable over failover, it is
445 possible that for some workloads it is actually worse, due to the
446 variable time of the “suspend” phase during live migration.
448 To allow the tools to work consistently over such instances (without
449 having to hard-code instance names), a new backend parameter
450 ``always_failover`` has been added to control the migration/failover
451 behaviour. When set to True, all migration requests for an instance will
452 instead fall-back to failover.
454 Instance memory ballooning
455 ++++++++++++++++++++++++++
457 Initial support for memory ballooning has been added. The memory for an
458 instance is no longer fixed (backend parameter ``memory``), but instead
459 can vary between minimum and maximum values (backend parameters
460 ``minmem`` and ``maxmem``). Currently we only change an instance's
463 - live migrating or failing over and instance and the target node
464 doesn't have enough memory
465 - user requests changing the memory via ``gnt-instance modify
471 In order to control the use of specific CPUs by instance, support for
472 controlling CPU pinning has been added for the Xen, HVM and LXC
473 hypervisors. This is controlled by a new hypervisor parameter
474 ``cpu_mask``; details about possible values for this are in the
475 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
476 VCPU-to-CPU mapping) form will work well only when all nodes in your
477 cluster have the same amount of CPUs.
482 Another area in which Ganeti was not customisable were the parameters
483 used for storage configuration, e.g. how many stripes to use for LVM,
484 DRBD resync configuration, etc.
486 To improve this area, we've added disks parameters, which are
487 customisable at cluster and node group level, and which allow to
488 specify various parameters for disks (DRBD has the most parameters
489 currently), for example:
491 - DRBD resync algorithm and parameters (e.g. speed)
492 - the default VG for meta-data volumes for DRBD
493 - number of stripes for LVM (plain disk template)
496 These parameters can be modified via ``gnt-cluster modify -D …`` and
497 ``gnt-group modify -D …``, and are used at either instance creation (in
498 case of LVM stripes, for example) or at disk “activation” time
501 Rados block device support
502 ++++++++++++++++++++++++++
504 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
505 denoted by the ``rbd`` disk template type. This is considered
506 experimental, feedback is welcome. For details on configuring it, see
507 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
512 The existing master IP functionality works well only in simple setups (a
513 single network shared by all nodes); however, if nodes belong to
514 different networks, then the ``/32`` setup and lack of routing
515 information is not enough.
517 To allow the master IP to function well in more complex cases, the
518 system was reworked as follows:
520 - a master IP netmask setting has been added
521 - the master IP activation/turn-down code was moved from the node daemon
523 - whether to run the Ganeti-supplied master IP script or a user-supplied
524 on is a ``gnt-cluster init`` setting
526 Details about the location of the standard and custom setup scripts are
527 in the man page :manpage:`gnt-cluster(8)`; for information about the
528 setup script protocol, look at the Ganeti-supplied script.
533 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
536 It is now possible to use TLS-protected connections, and when renewing
537 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
538 it is now possible to specify spice or spice CA certificates. Also, it
539 is possible to configure a password for SPICE sessions via the
540 hypervisor parameter ``spice_password_file``.
542 There are also new parameters to control the compression and streaming
543 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
544 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
545 for the spice parameters.
547 Lastly, it is now possible to see the SPICE connection information via
548 ``gnt-instance console``.
553 A new tool (``tools/ovfconverter``) has been added that supports
554 conversion between Ganeti and the `Open Virtualization Format
555 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
558 This relies on the ``qemu-img`` tool to convert the disk formats, so the
559 actual compatibility with other virtualization solutions depends on it.
564 The configuration query daemon (``ganeti-confd``) is now optional, and
565 has been rewritten in Haskell; whether to use the daemon at all, use the
566 Python (default) or the Haskell version is selectable at configure time
567 via the ``--enable-confd`` parameter, which can take one of the
568 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
569 daemon will result in a smaller footprint; for larger systems, we
570 welcome feedback on the Haskell version which might become the default
573 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
574 daemon running. The Python version doesn't implement the new call.
577 User interface changes
578 ~~~~~~~~~~~~~~~~~~~~~~
580 We have replaced the ``--disks`` option of ``gnt-instance
581 replace-disks`` with a more flexible ``--disk`` option, which allows
582 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
583 disk size and mode can be changed upon recreation (via ``gnt-instance
584 recreate-disks``, which accepts the same ``--disk`` option).
586 As many people are used to a ``show`` command, we have added that as an
587 alias to ``info`` on all ``gnt-*`` commands.
589 The ``gnt-instance grow-disk`` command has a new mode in which it can
590 accept the target size of the disk, instead of the delta; this can be
591 more safe since two runs in absolute mode will be idempotent, and
592 sometimes it's also easier to specify the desired size directly.
594 Also the handling of instances with regard to offline secondaries has
595 been improved. Instance operations should not fail because one of it's
596 secondary nodes is offline, even though it's safe to proceed.
598 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
599 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
600 minors to instance name.
605 RAPI coverage has improved, with (for example) new resources for
606 recreate-disks, node power-cycle, etc.
611 There is partial support for ``xl`` in the Xen hypervisor; feedback is
614 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
615 whether to still support Python 2.4 or move to Python 2.6 as minimum
618 Support for Fedora has been slightly improved; the provided example
619 init.d script should work better on it and the INSTALL file should
620 document the needed dependencies.
625 The deprecated ``QueryLocks`` LUXI request has been removed. Use
626 ``Query(what=QR_LOCK, ...)`` instead.
628 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
629 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
630 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
631 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
632 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
634 RAPI client: ``CertificateError`` now derives from
635 ``GanetiApiError``. This should make it more easy to handle Ganeti
638 Deprecation warnings due to PyCrypto/paramiko import in
639 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
640 make sure to run an up-to-date paramiko version, if you use this tool.
642 The QA scripts now depend on Python 2.5 or above (the main code base
643 still works with Python 2.4).
645 The configuration file (``config.data``) is now written without
646 indentation for performance reasons; if you want to edit it, it can be
647 re-formatted via ``tools/fmtjson``.
649 A number of bugs has been fixed in the cluster merge tool.
651 ``x509`` certification verification (used in import-export) has been
652 changed to allow the same clock skew as permitted by the cluster
653 verification. This will remove some rare but hard to diagnose errors in
660 *(Released Thu, 19 Jul 2012)*
662 Very few changes from rc4 to the final release, only bugfixes:
664 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
665 instance, fix CDROM booting for KVM instances)
666 - fixed node group modification of node parameters
667 - fixed issue in LUClusterVerifyGroup with multi-group clusters
668 - fixed generation of bash completion to ensure a stable ordering
675 *(Released Fri, 13 Jul 2012)*
677 Third release candidate for 2.6. The following changes were done from
680 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
681 - Fixed an inconsistency in the LUXI protocol with the provided
682 arguments (NOT backwards compatible)
683 - Fixed a bug with node groups ipolicy where ``min`` was greater than
684 the cluster ``std`` value
685 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
686 easier instance debugging on nodes (requires ``hconfd`` to work)
692 *(Released Tue, 03 Jul 2012)*
694 Second release candidate for 2.6. The following changes were done from
697 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
699 - Fixed a RAPI regression on missing beparams/memory
700 - Fixed redistribution of files on offline nodes
701 - Added possibility to run activate-disks even though secondaries are
702 offline. With this change it relaxes also the strictness on some other
703 commands which use activate disks internally:
704 * ``gnt-instance start|reboot|rename|backup|export``
705 - Made it possible to remove safely an instance if its secondaries are
707 - Made it possible to reinstall even though secondaries are offline
713 *(Released Mon, 25 Jun 2012)*
715 First release candidate for 2.6. The following changes were done from
718 - Fixed bugs with disk parameters and ``rbd`` templates as well as
720 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
721 behaviour. It supports now the modify operation
722 - ``hcheck`` implemented to analyze cluster health and possibility of
723 improving health by rebalance
724 - ``hbal`` has been improved in dealing with split instances
730 *(Released Mon, 11 Jun 2012)*
732 Second beta release of 2.6. The following changes were done from beta2
735 - Fixed ``daemon-util`` with non-root user models
736 - Fixed creation of plain instances with ``--no-wait-for-sync``
737 - Fix wrong iv_names when running ``cfgupgrade``
738 - Export more information in RAPI group queries
739 - Fixed bug when changing instance network interfaces
740 - Extended burnin to do NIC changes
741 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
742 - Changed default for DRBD barriers
743 - Fixed DRBD error reporting for syncer rate
744 - Verify the options on disk parameters
746 And of course various fixes to documentation and improved unittests and
753 *(Released Wed, 23 May 2012)*
755 First beta release of 2.6. The following changes were done from beta1 to
758 - integrated patch for distributions without ``start-stop-daemon``
759 - adapted example init.d script to work on Fedora
760 - fixed log handling in Haskell daemons
761 - adapted checks in the watcher for pycurl linked against libnss
762 - add partial support for ``xl`` instead of ``xm`` for Xen
763 - fixed a type issue in cluster verification
764 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
767 Plus integrated fixes from the 2.5 branch:
769 - fixed ``kvm-ifup`` to use ``/bin/bash``
770 - fixed parallel build failures
771 - KVM live migration when using a custom keymap
777 *(Released Tue, 24 Jul 2012)*
779 A small bugfix release, with no new features:
781 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
782 different default shell (e.g. Debian, Ubuntu)
783 - fixed KVM startup and live migration with a custom keymap (fixes Issue
784 243 and Debian bug #650664)
785 - fixed compatibility with KVM versions that don't support multiple boot
786 devices (fixes Issue 230 and Debian bug #624256)
788 Additionally, a few fixes were done to the build system (fixed parallel
789 build failures) and to the unittests (fixed race condition in test for
790 FileID functions, and the default enable/disable mode for QA test is now
797 *(Released Fri, 11 May 2012)*
799 A small bugfix release.
801 The main issues solved are on the topic of compatibility with newer LVM
804 - fixed parsing of ``lv_attr`` field
805 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
806 the ``--force`` flag is needed
808 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
809 to accept kernel 3.x too (was hardcoded to 2.6.*).
811 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
812 of hook results and that also made display of other errors suboptimal
813 was fixed; the code behaves now like 2.4 and earlier.
815 Another change in 2.5, the cleanup of the OS scripts environment, is too
816 aggressive: it removed even the ``PATH`` variable, which requires the OS
817 scripts to *always* need to export it. Since this is a bit too strict,
818 we now export a minimal PATH, the same that we export for hooks.
820 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
821 integrated into this release.
823 Finally, a few other miscellaneous changes were done (no new features,
824 just small improvements):
826 - Fix ``gnt-group --help`` display
827 - Fix hardcoded Xen kernel path
828 - Fix grow-disk handling of invalid units
829 - Update synopsis for ``gnt-cluster repair-disk-sizes``
830 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
836 *(Released Thu, 12 Apr 2012)*
838 Incompatible/important changes and bugfixes
839 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
841 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
842 resource's ``ip_check`` parameter changed from ``True`` to ``False``
843 to match the underlying LUXI interface.
844 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
845 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
846 not maintain backwards-compatibility as the underlying operation
847 changed in an incompatible way. The RAPI client can talk to old
848 servers, but it needs to be told so as the return value changed.
849 - When creating file-based instances via RAPI, the ``file_driver``
850 parameter no longer defaults to ``loop`` and must be specified.
851 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
853 - Support for the undocumented and deprecated RAPI instance creation
854 request format version 0 has been dropped. Use version 1, supported
855 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
856 - Pyparsing 1.4.6 or above is required, see :doc:`installation
857 documentation <install>`.
858 - The "cluster-verify" hooks are now executed per group by the
859 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
860 if you just run ``gnt-cluster verify``, which generates one opcode per
862 - The environment as passed to the OS scripts is cleared, and thus no
863 environment variables defined in the node daemon's environment will be
864 inherited by the scripts.
865 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
867 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
868 support operations involving multiple node groups.
869 - Offline nodes are ignored when failing over an instance.
870 - Support for KVM version 1.0, which changed the version reporting format
872 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
874 - ``Makefile`` is now compatible with Automake 1.11.2
875 - Includes all bugfixes made in the 2.4 series
880 - The ganeti-htools project has been merged into the ganeti-core source
881 tree and will be built as part of Ganeti (see :doc:`install-quick`).
882 - Implemented support for :doc:`shared storage <design-shared-storage>`.
883 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
884 GPT-style partition tables (requires `parted
885 <http://www.gnu.org/s/parted/>`_).
886 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
887 - Allowed adding tags on instance creation.
888 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
889 - Allow instances to be started in a paused state, enabling the user to
890 see the complete console output on boot using the console.
891 - Added new hypervisor flag to control default reboot behaviour
892 (``reboot_behavior``).
893 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
894 - Improved out-of-band management support:
896 - Added ``gnt-node health`` command reporting the health status of
898 - Added ``gnt-node power`` command to manage power status of nodes.
899 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
901 - Instance migration can fall back to failover if instance is not
903 - Filters can be used when listing nodes, instances, groups and locks;
904 see :manpage:`ganeti(7)` manpage.
905 - Added post-execution status as variables to :doc:`hooks <hooks>`
907 - Instance tags are exported/imported together with the instance.
908 - When given an explicit job ID, ``gnt-job info`` will work for archived
910 - Jobs can define dependencies on other jobs (not yet supported via
911 RAPI or command line, but used by internal commands and usable via
914 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
917 - Instance failover is now available as a RAPI resource
918 (``/2/instances/[instance_name]/failover``).
919 - ``gnt-instance info`` defaults to static information if primary node
921 - Opcodes have a new ``comment`` attribute.
922 - Added basic SPICE support to KVM hypervisor.
923 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
925 Node group improvements
926 ~~~~~~~~~~~~~~~~~~~~~~~
928 - ``gnt-cluster verify`` has been modified to check groups separately,
929 thereby improving performance.
930 - Node group support has been added to ``gnt-cluster verify-disks``,
931 which now operates per node group.
932 - Watcher has been changed to work better with node groups.
934 - One process and state file per node group.
935 - Slow watcher in one group doesn't block other group's watcher.
937 - Added new command, ``gnt-group evacuate``, to move all instances in a
938 node group to other groups.
939 - Added ``gnt-instance change-group`` to move an instance to another
941 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
942 per-group operations.
943 - Node groups can be tagged.
944 - Some operations switch from an exclusive to a shared lock as soon as
946 - Instance's primary and secondary nodes' groups are now available as
947 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
948 and ``snodes.group.uuid``).
953 - Numerous updates to documentation and manpages.
955 - :doc:`RAPI <rapi>` documentation now has detailed parameter
957 - Some opcode/job results are now also documented, see :doc:`RAPI
960 - A lockset's internal lock is now also visible in lock monitor.
961 - Log messages from job queue workers now contain information about the
962 opcode they're processing.
963 - ``gnt-instance console`` no longer requires the instance lock.
964 - A short delay when waiting for job changes reduces the number of LUXI
965 requests significantly.
966 - DRBD metadata volumes are overwritten with zeros during disk creation.
967 - Out-of-band commands no longer acquire the cluster lock in exclusive
969 - ``devel/upload`` now uses correct permissions for directories.
975 *(Released Fri, 23 Mar 2012)*
977 This was the sixth release candidate of the 2.5 series.
983 *(Released Mon, 9 Jan 2012)*
985 This was the fifth release candidate of the 2.5 series.
991 *(Released Thu, 27 Oct 2011)*
993 This was the fourth release candidate of the 2.5 series.
999 *(Released Wed, 26 Oct 2011)*
1001 This was the third release candidate of the 2.5 series.
1007 *(Released Tue, 18 Oct 2011)*
1009 This was the second release candidate of the 2.5 series.
1015 *(Released Tue, 4 Oct 2011)*
1017 This was the first release candidate of the 2.5 series.
1023 *(Released Wed, 31 Aug 2011)*
1025 This was the third beta release of the 2.5 series.
1031 *(Released Mon, 22 Aug 2011)*
1033 This was the second beta release of the 2.5 series.
1039 *(Released Fri, 12 Aug 2011)*
1041 This was the first beta release of the 2.5 series.
1047 *(Released Thu, 27 Oct 2011)*
1049 - Fixed bug when parsing command line parameter values ending in
1051 - Fixed assertion error after unclean master shutdown
1052 - Disable HTTP client pool for RPC, significantly reducing memory usage
1054 - Fixed queue archive creation with wrong permissions
1060 *(Released Tue, 23 Aug 2011)*
1064 - Fixed documentation for importing with ``--src-dir`` option
1065 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1066 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1072 *(Released Fri, 5 Aug 2011)*
1074 Many bug-fixes and a few small features:
1076 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1077 issues when you tried to add an instance with two MAC addresses in one
1079 - KVM: fixed per-instance stored UID value
1080 - KVM: configure bridged NICs at migration start
1081 - KVM: Fix a bug where instance will not start with never KVM versions
1083 - Added OS search path to ``gnt-cluster info``
1084 - Fixed an issue with ``file_storage_dir`` where you were forced to
1085 provide an absolute path, but the documentation states it is a
1086 relative path, the documentation was right
1087 - Added a new parameter to instance stop/start called ``--no-remember``
1088 that will make the state change to not be remembered
1089 - Implemented ``no_remember`` at RAPI level
1090 - Improved the documentation
1091 - Node evacuation: don't call IAllocator if node is already empty
1092 - Fixed bug in DRBD8 replace disks on current nodes
1093 - Fixed bug in recreate-disks for DRBD instances
1094 - Moved assertion checking locks in ``gnt-instance replace-disks``
1095 causing it to abort with not owning the right locks for some situation
1096 - Job queue: Fixed potential race condition when cancelling queued jobs
1097 - Fixed off-by-one bug in job serial generation
1098 - ``gnt-node volumes``: Fix instance names
1099 - Fixed aliases in bash completion
1100 - Fixed a bug in reopening log files after being sent a SIGHUP
1101 - Added a flag to burnin to allow specifying VCPU count
1102 - Bugfixes to non-root Ganeti configuration
1108 *(Released Thu, 12 May 2011)*
1110 Many bug-fixes and a few new small features:
1112 - Fixed a bug related to log opening failures
1113 - Fixed a bug in instance listing with orphan instances
1114 - Fixed a bug which prevented resetting the cluster-level node parameter
1115 ``oob_program`` to the default
1116 - Many fixes related to the ``cluster-merge`` tool
1117 - Fixed a race condition in the lock monitor, which caused failures
1118 during (at least) creation of many instances in parallel
1119 - Improved output for gnt-job info
1120 - Removed the quiet flag on some ssh calls which prevented debugging
1122 - Improved the N+1 failure messages in cluster verify by actually
1123 showing the memory values (needed and available)
1124 - Increased lock attempt timeouts so that when executing long operations
1125 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1126 too early and thus prevent the use of the 'fair' mechanism
1127 - Changed instance query data (``gnt-instance info``) to not acquire
1128 locks unless needed, thus allowing its use on locked instance if only
1129 static information is asked for
1130 - Improved behaviour with filesystems that do not support rename on an
1132 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1133 kept locks on all nodes during the wipe, which is unneeded
1134 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1135 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1136 10GiB) which caused the wipe to fail right at the end in some cases
1137 - Fixed master IP activation when doing master failover with no-voting
1138 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1139 the master node itself
1140 - Fixed potential data-loss in under disk full conditions, where Ganeti
1141 wouldn't check correctly the return code and would consider
1142 partially-written files 'correct'
1143 - Fixed bug related to multiple VGs and DRBD disk replacing
1144 - Added new disk parameter ``metavg`` that allows placement of the meta
1145 device for DRBD in a different volume group
1146 - Fixed error handling in the node daemon when the system libc doesn't
1147 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1148 - Fixed lock release during replace-disks, which kept cluster-wide locks
1149 when doing disk replaces with an iallocator script
1150 - Added check for missing bridges in cluster verify
1151 - Handle EPIPE errors while writing to the terminal better, so that
1152 piping the output to e.g. ``less`` doesn't cause a backtrace
1153 - Fixed rare case where a ^C during Luxi calls could have been
1154 interpreted as server errors, instead of simply terminating
1155 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1157 - Added a few more parameters to the KVM hypervisor, allowing a second
1158 CDROM, custom disk type for CDROMs and a floppy image
1159 - Removed redundant message in instance rename when the name is given
1161 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1162 disks on new nodes, allowing recreation when the original instance
1163 nodes are completely gone
1164 - Added option when converting disk templates to DRBD to skip waiting
1165 for the resync, in order to make the instance available sooner
1166 - Added two new variables to the OS scripts environment (containing the
1168 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1169 to allow use of ``pvgrub`` as bootloader
1170 - Changed the instance memory modifications to only check out-of-memory
1171 conditions on memory increases, and turned the secondary node warnings
1172 into errors (they can still be overridden via ``--force``)
1173 - Fixed the handling of a corner case when the Python installation gets
1174 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1175 to execute a command that doesn't exist
1176 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1177 node of the instance returned failures during instance shutdown; this
1178 adds the option ``--ignore-consistency`` to gnt-instance move
1180 And as usual, various improvements to the error messages, documentation
1187 *(Released Wed, 09 Mar 2011)*
1189 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1190 the RAPI users file if run twice (even with ``--dry-run``).
1192 The release fixes that bug (nothing else changed).
1198 *(Released Mon, 07 Mar 2011)*
1200 Final 2.4.0 release. Just a few small fixes:
1202 - Fixed RAPI node evacuate
1203 - Fixed the kvm-ifup script
1204 - Fixed internal error handling for special job cases
1205 - Updated man page to specify the escaping feature for options
1211 *(Released Mon, 28 Feb 2011)*
1213 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1214 that this feature wiped the disks of the wrong instance, leading to loss
1219 - Fixed title of query field containing instance name
1220 - Expanded the glossary in the documentation
1221 - Fixed one unittest (internal issue)
1227 *(Released Mon, 21 Feb 2011)*
1229 A number of bug fixes plus just a couple functionality changes.
1231 On the user-visible side, the ``gnt-* list`` command output has changed
1232 with respect to "special" field states. The current rc1 style of display
1233 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1234 the default output mode special fields states are displayed as follows:
1236 - Offline resource: ``*``
1237 - Unavailable/not applicable: ``-``
1238 - Data missing (RPC failure): ``?``
1239 - Unknown field: ``??``
1241 Another user-visible change is the addition of ``--force-join`` to
1246 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1247 - Fixed regression in RAPI/instance reinstall where all parameters were
1248 required (instead of optional)
1249 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1250 - Fixed iallocator usage (offline nodes were not considered offline)
1251 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1252 - Fixed hypervisor and OS parameter validation with respect to
1253 non-vm_capable nodes
1254 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1256 - Fixed ``tools/listrunner`` with respect to agent-based usage
1262 *(Released Fri, 4 Feb 2011)*
1264 Many changes and fixes since the beta1 release. While there were some
1265 internal changes, the code has been mostly stabilised for the RC
1268 Note: the dumb allocator was removed in this release, as it was not kept
1269 up-to-date with the IAllocator protocol changes. It is recommended to
1270 use the ``hail`` command from the ganeti-htools package.
1272 Note: the 2.4 and up versions of Ganeti are not compatible with the
1273 0.2.x branch of ganeti-htools. You need to upgrade to
1274 ganeti-htools-0.3.0 (or later).
1276 Regressions fixed from 2.3
1277 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1279 - Fixed the ``gnt-cluster verify-disks`` command
1280 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1282 - Fixed disk adoption breakage
1283 - Fixed wrong headers in instance listing for field aliases
1288 - Fixed corner case in KVM handling of NICs
1289 - Fixed many cases of wrong handling of non-vm_capable nodes
1290 - Fixed a bug where a missing instance symlink was not possible to
1291 recreate with any ``gnt-*`` command (now ``gnt-instance
1292 activate-disks`` does it)
1293 - Fixed the volume group name as reported by ``gnt-cluster
1295 - Increased timeouts for the import-export code, hopefully leading to
1296 fewer aborts due network or instance timeouts
1297 - Fixed bug in ``gnt-node list-storage``
1298 - Fixed bug where not all daemons were started on cluster
1299 initialisation, but only at the first watcher run
1300 - Fixed many bugs in the OOB implementation
1301 - Fixed watcher behaviour in presence of instances with offline
1303 - Fixed instance list output for instances running on the wrong node
1304 - a few fixes to the cluster-merge tool, but it still cannot merge
1305 multi-node groups (currently it is not recommended to use this tool)
1311 - Improved network configuration for the KVM hypervisor
1312 - Added e1000 as a supported NIC for Xen-HVM
1313 - Improved the lvmstrap tool to also be able to use partitions, as
1314 opposed to full disks
1315 - Improved speed of disk wiping (the cluster parameter
1316 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1317 of instance creations
1318 - Added documentation for the OS parameters
1319 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1320 hypervisor is not responding
1321 - Added display of blacklisted and hidden OS information in
1322 ``gnt-cluster info``
1323 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1324 NIC and node parameters, which might create problems with currently
1325 invalid (but undetected) configuration files, but prevents validation
1326 failures when unrelated parameters are modified
1327 - Changed cluster initialisation to wait for the master daemon to become
1329 - Expanded the RAPI interface:
1331 - Added config redistribution resource
1332 - Added activation/deactivation of instance disks
1333 - Added export of console information
1335 - Implemented log file reopening on SIGHUP, which allows using
1336 logrotate(8) for the Ganeti log files
1337 - Added a basic OOB helper script as an example
1343 *(Released Fri, 14 Jan 2011)*
1348 - Fixed timezone issues when formatting timestamps
1349 - Added support for node groups, available via ``gnt-group`` and other
1351 - Added out-of-band framework and management, see :doc:`design
1352 document <design-oob>`
1353 - Removed support for roman numbers from ``gnt-node list`` and
1354 ``gnt-instance list``.
1355 - Allowed modification of master network interface via ``gnt-cluster
1356 modify --master-netdev``
1357 - Accept offline secondaries while shutting down instance disks
1358 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1359 - Added support for multiple LVM volume groups
1360 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1362 - Added commands to list available fields:
1364 - ``gnt-node list-fields``
1365 - ``gnt-group list-fields``
1366 - ``gnt-instance list-fields``
1368 - Updated documentation and man pages
1373 - Moved ``rapi_users`` file into separate directory, now named
1374 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1376 - Added new tool for running commands on many machines,
1377 ``tools/ganeti-listrunner``
1378 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1379 improving the ``gnt-instance console`` output
1380 - Allowed customisation of disk index separator at ``configure`` time
1381 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1382 - Added support for non-partitioned md disks in ``lvmstrap``
1383 - Added script to gracefully power off KVM instances
1384 - Split ``utils`` module into smaller parts
1385 - Changed query operations to return more detailed information, e.g.
1386 whether an information is unavailable due to an offline node. To use
1387 this new functionality, the LUXI call ``Query`` must be used. Field
1388 information is now stored by the master daemon and can be retrieved
1389 using ``QueryFields``. Instances, nodes and groups can also be queried
1390 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1391 exposed via RAPI). The following commands make use of this
1392 infrastructure change:
1394 - ``gnt-group list``
1395 - ``gnt-group list-fields``
1397 - ``gnt-node list-fields``
1398 - ``gnt-instance list``
1399 - ``gnt-instance list-fields``
1400 - ``gnt-debug locks``
1405 - New RAPI resources (see :doc:`rapi`):
1409 - ``/2/groups/[group_name]``
1410 - ``/2/groups/[group_name]/assign-nodes``
1411 - ``/2/groups/[group_name]/modify``
1412 - ``/2/groups/[group_name]/rename``
1413 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1417 - Implemented ``no_install`` for instance creation
1418 - Implemented OS parameters for instance reinstallation, allowing
1419 use of special settings on reinstallation (e.g. for preserving data)
1424 - Added IPv6 support in import/export
1425 - Pause DRBD synchronization while wiping disks on instance creation
1426 - Updated unittests and QA scripts
1427 - Improved network parameters passed to KVM
1428 - Converted man pages from docbook to reStructuredText
1434 *(Released Mon, 20 Dec 2010)*
1436 Released version 2.3.1~rc1 without any changes.
1442 *(Released Wed, 1 Dec 2010)*
1444 - impexpd: Disable OpenSSL compression in socat if possible (backport
1445 from master, commit e90739d625b, see :doc:`installation guide
1446 <install-quick>` for details)
1447 - Changed unittest coverage report to exclude test scripts
1448 - Added script to check version format
1454 *(Released Wed, 1 Dec 2010)*
1456 Released version 2.3.0~rc1 without any changes.
1462 *(Released Fri, 19 Nov 2010)*
1464 A number of bugfixes and documentation updates:
1466 - Update ganeti-os-interface documentation
1467 - Fixed a bug related to duplicate MACs or similar items which should be
1469 - Fix breakage in OS state modify
1470 - Reinstall instance: disallow offline secondaries (fixes bug related to
1471 OS changing but reinstall failing)
1472 - plus all the other fixes between 2.2.1 and 2.2.2
1478 *(Released Tue, 2 Nov 2010)*
1480 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1481 - Fixed master failover race with watcher
1482 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1484 - Accept previously stopped instance for export with instance removal
1485 - Simplify and extend the environment variables for instance OS scripts
1486 - Added new node flags, ``master_capable`` and ``vm_capable``
1487 - Added optional instance disk wiping prior during allocation. This is a
1488 cluster-wide option and can be set/modified using
1489 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1490 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1491 :doc:`install-quick`
1492 - Added a new watcher option (``--ignore-pause``)
1493 - Added option to ignore offline node on instance start/stop
1494 (``--ignore-offline``)
1495 - Allow overriding OS parameters with ``gnt-instance reinstall``
1496 - Added ability to change node's secondary IP address using ``gnt-node
1498 - Implemented privilege separation for all daemons except
1499 ``ganeti-noded``, see ``configure`` options
1500 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1502 - Implemented job priorities (see ``ganeti(7)`` manpage)
1503 - Ignore failures while shutting down instances during failover from
1505 - Exit daemon's bootstrap process only once daemon is ready
1506 - Export more information via ``LUInstanceQuery``/remote API
1507 - Improved documentation, QA and unittests
1508 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1509 restart if the file was created or changed
1510 - Added LUXI protocol version sent with each request and response,
1511 allowing detection of server/client mismatches
1512 - Moved the Python scripts among gnt-* and ganeti-* into modules
1513 - Moved all code related to setting up SSH to an external script,
1515 - Infrastructure changes for node group support in future versions
1521 *(Released Fri, 19 Nov 2010)*
1523 A few small bugs fixed, and some improvements to the build system:
1525 - Fix documentation regarding conversion to drbd
1526 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1528 - Fix error handling in node modify with multiple changes
1529 - Allow remote imports without checked names
1535 *(Released Tue, 19 Oct 2010)*
1537 - Disable SSL session ID cache in RPC client
1543 *(Released Thu, 14 Oct 2010)*
1545 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1546 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1548 - Fix problems with interaction between Curl and Python's HTTP server,
1549 resulting in increased speed in many RPC calls
1550 - Improve our release script to prevent breakage with older aclocal and
1557 *(Released Thu, 7 Oct 2010)*
1559 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1560 value retrieved from master
1561 - Added support for blacklisted or hidden OS definitions
1562 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1563 - Added support for -mem-path in KVM hypervisor abstraction layer
1564 - Allow overriding instance parameters in tool for inter-cluster
1565 instance moves (``tools/move-instance``)
1566 - Improved opcode summaries (e.g. in ``gnt-job list``)
1567 - Improve consistency of OS listing by sorting it
1568 - Documentation updates
1574 *(Released Fri, 8 Oct 2010)*
1576 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1582 *(Released Mon, 4 Oct 2010)*
1584 - Fixed regression in ``gnt-instance rename``
1590 *(Released Wed, 22 Sep 2010)*
1592 - Fixed OS_VARIANT variable for OS scripts
1593 - Fixed cluster tag operations via RAPI
1594 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1595 - Disabled RAPI CA checks in watcher
1601 *(Released Mon, 23 Aug 2010)*
1603 - Support DRBD versions of the format "a.b.c.d"
1605 - Re-introduce support for usage from multiple threads in RAPI client
1606 - Instance renames and modify via RAPI
1607 - Work around race condition between processing and archival in job
1609 - Mark opcodes following failed one as failed, too
1610 - Job field ``lock_status`` was removed due to difficulties making it
1611 work with the changed job queue in Ganeti 2.2; a better way to monitor
1612 locks is expected for a later 2.2.x release
1613 - Fixed dry-run behaviour with many commands
1614 - Support ``ssh-agent`` again when adding nodes
1615 - Many additional bugfixes
1621 *(Released Fri, 30 Jul 2010)*
1623 Important change: the internal RPC mechanism between Ganeti nodes has
1624 changed from using a home-grown http library (based on the Python base
1625 libraries) to use the PycURL library. This requires that PycURL is
1626 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1627 against GnuTLS by default. cURL's support for GnuTLS had known issues
1628 before cURL 7.21.0 and we recommend using the latest cURL release or
1629 linking against OpenSSL. Most other distributions already link PycURL
1630 and cURL against OpenSSL. The command::
1632 python -c 'import pycurl; print pycurl.version'
1634 can be used to determine the libraries PycURL and cURL are linked
1637 Other significant changes:
1639 - Rewrote much of the internals of the job queue, in order to achieve
1640 better parallelism; this decouples job query operations from the job
1641 processing, and it should allow much nicer behaviour of the master
1642 daemon under load, and it also has uncovered some long-standing bugs
1643 related to the job serialisation (now fixed)
1644 - Added a default iallocator setting to the cluster parameters,
1645 eliminating the need to always pass nodes or an iallocator for
1646 operations that require selection of new node(s)
1647 - Added experimental support for the LXC virtualization method
1648 - Added support for OS parameters, which allows the installation of
1649 instances to pass parameter to OS scripts in order to customise the
1651 - Added a hypervisor parameter controlling the migration type (live or
1652 non-live), since hypervisors have various levels of reliability; this
1653 has renamed the 'live' parameter to 'mode'
1654 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1655 logical volumes, meaning that cluster verify will ignore them and not
1656 flag their presence as errors
1657 - The watcher will now reset the error count for failed instances after
1658 8 hours, thus allowing self-healing if the problem that caused the
1659 instances to be down/fail to start has cleared in the meantime
1660 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1661 check for, and warn, if the drbd module parameter ``usermode_helper``
1662 is not consistent with the cluster-wide setting; this is needed to
1663 make diagnose easier of failed drbd creations
1664 - Started adding base IPv6 support, but this is not yet
1665 enabled/available for use
1666 - Rename operations (cluster, instance) will now return the new name,
1667 which is especially useful if a short name was passed in
1668 - Added support for instance migration in RAPI
1669 - Added a tool to pre-configure nodes for the SSH setup, before joining
1670 them to the cluster; this will allow in the future a simplified model
1671 for node joining (but not yet fully enabled in 2.2); this needs the
1672 paramiko python library
1673 - Fixed handling of name-resolving errors
1674 - Fixed consistency of job results on the error path
1675 - Fixed master-failover race condition when executed multiple times in
1677 - Fixed many bugs related to the job queue (mostly introduced during the
1678 2.2 development cycle, so not all are impacting 2.1)
1679 - Fixed instance migration with missing disk symlinks
1680 - Fixed handling of unknown jobs in ``gnt-job archive``
1681 - And many other small fixes/improvements
1685 - Enhanced both the unittest and the QA coverage
1686 - Switched the opcode validation to a generic model, and extended the
1687 validation to all opcode parameters
1688 - Changed more parts of the code that write shell scripts to use the
1690 - Switched the master daemon to use the asyncore library for the Luxi
1697 *(Released Thu, 17 Jun 2010)*
1699 - Added tool (``move-instance``) and infrastructure to move instances
1700 between separate clusters (see :doc:`separate documentation
1701 <move-instance>` and :doc:`design document <design-2.2>`)
1702 - Added per-request RPC timeout
1703 - RAPI now requires a Content-Type header for requests with a body (e.g.
1704 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1705 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1706 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1708 - Implemented initial support for running Ganeti daemons as separate
1709 users, see configure-time flags ``--with-user-prefix`` and
1710 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1712 - Instances can be removed after export (``gnt-backup export
1713 --remove-instance``)
1714 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1715 key (instead of 1024 bit)
1716 - Added new cluster configuration file for cluster domain secret
1717 - Import/export now use SSL instead of SSH
1718 - Added support for showing estimated time when exporting an instance,
1719 see the ``ganeti-os-interface(7)`` manpage and look for
1726 *(Released Tue, 16 Nov 2010)*
1728 Some more bugfixes. Unless critical bugs occur, this will be the last
1731 - Fix case of MAC special-values
1732 - Fix mac checker regex
1733 - backend: Fix typo causing "out of range" error
1734 - Add missing --units in gnt-instance list man page
1740 *(Released Tue, 24 Aug 2010)*
1743 - Don't ignore secondary node silently on non-mirrored disk templates
1745 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1746 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1747 - Properly document the usb_mouse parameter
1748 - Fix path in ganeti-rapi(8) (issue 116)
1749 - Adjust error message when the ganeti user's .ssh directory is
1751 - Add same-node-check when changing the disk template to drbd
1757 *(Released Fri, 16 Jul 2010)*
1760 - Add an option to only select some reboot types during qa/burnin.
1761 (on some hypervisors consequent reboots are not supported)
1762 - Fix infrequent race condition in master failover. Sometimes the old
1763 master ip address would be still detected as up for a short time
1764 after it was removed, causing failover to fail.
1765 - Decrease mlockall warnings when the ctypes module is missing. On
1766 Python 2.4 we support running even if no ctypes module is installed,
1767 but we were too verbose about this issue.
1768 - Fix building on old distributions, on which man doesn't have a
1770 - Fix RAPI not to ignore the MAC address on instance creation
1771 - Implement the old instance creation format in the RAPI client.
1777 *(Released Thu, 01 Jul 2010)*
1779 A small bugfix release:
1780 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1781 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1782 a non-existing option
1783 - Fix instance create: the --force-variant option was ignored
1784 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1785 - Fix modify node storage with non-FQDN arguments
1786 - Fix RAPI client to authenticate under Python 2.6 when used
1787 for more than 5 requests needing authentication
1788 - Fix gnt-instance modify -t (storage) giving a wrong error message
1789 when converting a non-shutdown drbd instance to plain
1795 *(Released Fri, 18 Jun 2010)*
1797 A small bugfix release:
1799 - Fix live migration of KVM instances started with older Ganeti
1800 versions which had fewer hypervisor parameters
1801 - Fix gnt-instance grow-disk on down instances
1802 - Fix an error-reporting bug during instance migration
1803 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1804 silently ignoring broken ones
1805 - Fix an RPC error reporting bug affecting, for example, RAPI client
1807 - Fix bug triggered by different API version os-es on different nodes
1808 - Fix a bug in instance startup with custom hvparams: OS level
1809 parameters would fail to be applied.
1810 - Fix the RAPI client under Python 2.6 (but more work is needed to
1811 make it work completely well with OpenSSL)
1812 - Fix handling of errors when resolving names from DNS
1818 *(Released Thu, 3 Jun 2010)*
1820 A medium sized development cycle. Some new features, and some
1821 fixes/small improvements/cleanups.
1823 Significant features
1824 ~~~~~~~~~~~~~~~~~~~~
1826 The node deamon now tries to mlock itself into memory, unless the
1827 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1828 its logs, and falls back to console logging. This allows emergency
1829 features such as ``gnt-node powercycle`` to work even in the event of a
1830 broken node disk (tested offlining the disk hosting the node's
1831 filesystem and dropping its memory caches; don't try this at home)
1833 KVM: add vhost-net acceleration support. It can be tested with a new
1834 enough version of the kernel and of qemu-kvm.
1836 KVM: Add instance chrooting feature. If you use privilege dropping for
1837 your VMs you can also now force them to chroot to an empty directory,
1838 before starting the emulated guest.
1840 KVM: Add maximum migration bandwith and maximum downtime tweaking
1841 support (requires a new-enough version of qemu-kvm).
1843 Cluster verify will now warn if the master node doesn't have the master
1844 ip configured on it.
1846 Add a new (incompatible) instance creation request format to RAPI which
1847 supports all parameters (previously only a subset was supported, and it
1848 wasn't possible to extend the old format to accomodate all the new
1849 features. The old format is still supported, and a client can check for
1850 this feature, before using it, by checking for its presence in the
1851 ``features`` RAPI resource.
1853 Now with ancient latin support. Try it passing the ``--roman`` option to
1854 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1855 (requires the python-roman module to be installed, in order to work).
1860 As usual many internal code refactorings, documentation updates, and
1863 - Lots of improvements and cleanups to the experimental Remote API
1864 (RAPI) client library.
1865 - A new unit test suite for the core daemon libraries.
1866 - A fix to creating missing directories makes sure the umask is not
1867 applied anymore. This enforces the same directory permissions
1869 - Better handling terminating daemons with ctrl+c (used when running
1870 them in debugging mode).
1871 - Fix a race condition in live migrating a KVM instance, when stat()
1872 on the old proc status file returned EINVAL, which is an unexpected
1874 - Fixed manpage checking with newer man and utf-8 charachters. But now
1875 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1881 *(Released Fri, 7 May 2010)*
1883 Fix a bug which prevented untagged KVM instances from starting.
1889 *(Released Fri, 7 May 2010)*
1891 Another release with a long development cycle, during which many
1892 different features were added.
1894 Significant features
1895 ~~~~~~~~~~~~~~~~~~~~
1897 The KVM hypervisor now can run the individual instances as non-root, to
1898 reduce the impact of a VM being hijacked due to bugs in the
1899 hypervisor. It is possible to run all instances as a single (non-root)
1900 user, to manually specify a user for each instance, or to dynamically
1901 allocate a user out of a cluster-wide pool to each instance, with the
1902 guarantee that no two instances will run under the same user ID on any
1905 An experimental RAPI client library, that can be used standalone
1906 (without the other Ganeti libraries), is provided in the source tree as
1907 ``lib/rapi/client.py``. Note this client might change its interface in
1908 the future, as we iterate on its capabilities.
1910 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1911 replace the cluster's certificates and crypto keys. This might help in
1912 case they have been compromised, or have simply expired.
1914 A new disk option for instance creation has been added that allows one
1915 to "adopt" currently existing logical volumes, with data
1916 preservation. This should allow easier migration to Ganeti from
1917 unmanaged (or managed via other software) instances.
1919 Another disk improvement is the possibility to convert between redundant
1920 (DRBD) and plain (LVM) disk configuration for an instance. This should
1921 allow better scalability (starting with one node and growing the
1922 cluster, or shrinking a two-node cluster to one node).
1924 A new feature that could help with automated node failovers has been
1925 implemented: if a node sees itself as offline (by querying the master
1926 candidates), it will try to shutdown (hard) all instances and any active
1927 DRBD devices. This reduces the risk of duplicate instances if an
1928 external script automatically failovers the instances on such nodes. To
1929 enable this, the cluster parameter ``maintain_node_health`` should be
1930 enabled; in the future this option (per the name) will enable other
1931 automatic maintenance features.
1933 Instance export/import now will reuse the original instance
1934 specifications for all parameters; that means exporting an instance,
1935 deleting it and the importing it back should give an almost identical
1936 instance. Note that the default import behaviour has changed from
1937 before, where it created only one NIC; now it recreates the original
1940 Cluster verify has added a few new checks: SSL certificates validity,
1941 /etc/hosts consistency across the cluster, etc.
1946 As usual, many internal changes were done, documentation fixes,
1949 - Fixed cluster initialization with disabled cluster storage (regression
1950 introduced in 2.1.1)
1951 - File-based storage supports growing the disks
1952 - Fixed behaviour of node role changes
1953 - Fixed cluster verify for some corner cases, plus a general rewrite of
1954 cluster verify to allow future extension with more checks
1955 - Fixed log spamming by watcher and node daemon (regression introduced
1957 - Fixed possible validation issues when changing the list of enabled
1959 - Fixed cleanup of /etc/hosts during node removal
1960 - Fixed RAPI response for invalid methods
1961 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1962 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1963 from ide disks, etc.)
1964 - Allow OS changes without re-installation (to record a changed OS
1965 outside of Ganeti, or to allow OS renames)
1966 - Allow instance creation without OS installation (useful for example if
1967 the OS will be installed manually, or restored from a backup not in
1969 - Implemented option to make cluster ``copyfile`` use the replication
1971 - Added list of enabled hypervisors to ssconf (possibly useful for
1973 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1975 - A partial form of node re-IP is possible via node readd, which now
1976 allows changed node primary IP
1977 - Command line utilities now show an informational message if the job is
1979 - The logs of the master daemon now show the PID/UID/GID of the
1986 *(Released Fri, 12 Mar 2010)*
1988 During the 2.1.0 long release candidate cycle, a lot of improvements and
1989 changes have accumulated with were released later as 2.1.1.
1994 The node evacuate command (``gnt-node evacuate``) was significantly
1995 rewritten, and as such the IAllocator protocol was changed - a new
1996 request type has been added. This unfortunate change during a stable
1997 series is designed to improve performance of node evacuations; on
1998 clusters with more than about five nodes and which are well-balanced,
1999 evacuation should proceed in parallel for all instances of the node
2000 being evacuated. As such, any existing IAllocator scripts need to be
2001 updated, otherwise the above command will fail due to the unknown
2002 request. The provided "dumb" allocator has not been updated; but the
2003 ganeti-htools package supports the new protocol since version 0.2.4.
2005 Another important change is increased validation of node and instance
2006 names. This might create problems in special cases, if invalid host
2007 names are being used.
2009 Also, a new layer of hypervisor parameters has been added, that sits at
2010 OS level between the cluster defaults and the instance ones. This allows
2011 customisation of virtualization parameters depending on the installed
2012 OS. For example instances with OS 'X' may have a different KVM kernel
2013 (or any other parameter) than the cluster defaults. This is intended to
2014 help managing a multiple OSes on the same cluster, without manual
2015 modification of each instance's parameters.
2017 A tool for merging clusters, ``cluster-merge``, has been added in the
2018 tools sub-directory.
2023 - Improved the int/float conversions that should make the code more
2024 robust in face of errors from the node daemons
2025 - Fixed the remove node code in case of internal configuration errors
2026 - Fixed the node daemon behaviour in face of inconsistent queue
2027 directory (e.g. read-only file-system where we can't open the files
2029 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2030 now it either aborts cleanly or, if given the new "auto_promote"
2031 parameter, will automatically promote other nodes as needed
2032 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2033 completely prevent Ganeti from working
2034 - Fixed bug for instance export when not all disks were successfully
2036 - Fixed behaviour of node add when the new node is slow in starting up
2038 - Fixed handling of signals in the LUXI client, which should improve
2039 behaviour of command-line scripts
2040 - Added checks for invalid node/instance names in the configuration (now
2041 flagged during cluster verify)
2042 - Fixed watcher behaviour for disk activation errors
2043 - Fixed two potentially endless loops in http library, which led to the
2044 RAPI daemon hanging and consuming 100% CPU in some cases
2045 - Fixed bug in RAPI daemon related to hashed passwords
2046 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2048 - Enhanced compatibility with non-Debian OSes, but not using absolute
2049 path in some commands and allowing customisation of the ssh
2050 configuration directory
2051 - Fixed possible future issue with new Python versions by abiding to the
2052 proper use of ``__slots__`` attribute on classes
2053 - Added checks that should prevent directory traversal attacks
2054 - Many documentation fixes based on feedback from users
2059 - Added an "early_release" more for instance replace disks and node
2060 evacuate, where we release locks earlier and thus allow higher
2061 parallelism within the cluster
2062 - Added watcher hooks, intended to allow the watcher to restart other
2063 daemons (e.g. from the ganeti-nbma project), but they can be used of
2064 course for any other purpose
2065 - Added a compile-time disable for DRBD barriers, to increase
2066 performance if the administrator trusts the power supply or the
2067 storage system to not lose writes
2068 - Added the option of using syslog for logging instead of, or in
2069 addition to, Ganeti's own log files
2070 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2071 can indeed boot from a paravirtual NIC
2072 - Added a generic debug level for many operations; while this is not
2073 used widely yet, it allows one to pass the debug value all the way to
2075 - Enhanced the hooks environment for instance moves (failovers,
2076 migrations) where the primary/secondary nodes changed during the
2077 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2078 - Enhanced data validations for many user-supplied values; one important
2079 item is the restrictions imposed on instance and node names, which
2080 might reject some (invalid) host names
2081 - Add a configure-time option to disable file-based storage, if it's not
2082 needed; this allows greater security separation between the master
2083 node and the other nodes from the point of view of the inter-node RPC
2085 - Added user notification in interactive tools if job is waiting in the
2086 job queue or trying to acquire locks
2087 - Added log messages when a job is waiting for locks
2088 - Added filtering by node tags in instance operations which admit
2089 multiple instances (start, stop, reboot, reinstall)
2090 - Added a new tool for cluster mergers, ``cluster-merge``
2091 - Parameters from command line which are of the form ``a=b,c=d`` can now
2092 use backslash escapes to pass in values which contain commas,
2093 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2095 - For KVM, the instance name is the first parameter passed to KVM, so
2096 that it's more visible in the process list
2102 *(Released Tue, 2 Mar 2010)*
2104 Ganeti 2.1 brings many improvements with it. Major changes:
2106 - Added infrastructure to ease automated disk repairs
2107 - Added new daemon to export configuration data in a cheaper way than
2108 using the remote API
2109 - Instance NICs can now be routed instead of being associated with a
2111 - Improved job locking logic to reduce impact of jobs acquiring multiple
2112 locks waiting for other long-running jobs
2114 In-depth implementation details can be found in the Ganeti 2.1 design
2120 - Added chroot hypervisor
2121 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2123 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2124 ``bootloader_path`` and ``bootloader_args``)
2125 - Added the ``use_localtime`` option for the xen-hvm and kvm
2126 hypervisors, and the default value for this has changed to false (in
2127 2.0 xen-hvm always enabled it)
2128 - Added luxi call to submit multiple jobs in one go
2129 - Added cluster initialization option to not modify ``/etc/hosts``
2131 - Added network interface parameters
2132 - Added dry run mode to some LUs
2133 - Added RAPI resources:
2135 - ``/2/instances/[instance_name]/info``
2136 - ``/2/instances/[instance_name]/replace-disks``
2137 - ``/2/nodes/[node_name]/evacuate``
2138 - ``/2/nodes/[node_name]/migrate``
2139 - ``/2/nodes/[node_name]/role``
2140 - ``/2/nodes/[node_name]/storage``
2141 - ``/2/nodes/[node_name]/storage/modify``
2142 - ``/2/nodes/[node_name]/storage/repair``
2144 - Added OpCodes to evacuate or migrate all instances on a node
2145 - Added new command to list storage elements on nodes (``gnt-node
2146 list-storage``) and modify them (``gnt-node modify-storage``)
2147 - Added new ssconf files with master candidate IP address
2148 (``ssconf_master_candidates_ips``), node primary IP address
2149 (``ssconf_node_primary_ips``) and node secondary IP address
2150 (``ssconf_node_secondary_ips``)
2151 - Added ``ganeti-confd`` and a client library to query the Ganeti
2152 configuration via UDP
2153 - Added ability to run hooks after cluster initialization and before
2155 - Added automatic mode for disk replace (``gnt-instance replace-disks
2157 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2158 after catastrophic data-loss
2159 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2161 - Added ``gnt-instance move`` command to move instances
2162 - Added ``gnt-cluster watcher`` command to control watcher
2163 - Added ``gnt-node powercycle`` command to powercycle nodes
2164 - Added new job status field ``lock_status``
2165 - Added parseable error codes to cluster verification (``gnt-cluster
2166 verify --error-codes``) and made output less verbose (use
2167 ``--verbose`` to restore previous behaviour)
2168 - Added UUIDs to the main config entities (cluster, nodes, instances)
2169 - Added support for OS variants
2170 - Added support for hashed passwords in the Ganeti remote API users file
2172 - Added option to specify maximum timeout on instance shutdown
2173 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2174 - Added new helper script to start and stop Ganeti daemons
2175 (``daemon-util``), with the intent to reduce the work necessary to
2176 adjust Ganeti for non-Debian distributions and to start/stop daemons
2178 - Added more unittests
2179 - Fixed critical bug in ganeti-masterd startup
2180 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2181 now customisable at the cluster level for both the KVM and Xen
2182 hypervisors using the new ``migration_port`` parameter
2183 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2184 reinstalling an instance
2185 - Allowed ``@`` in tag names
2186 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2187 - Many documentation updates
2188 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2189 - ``gnt-instance reinstall`` can now reinstall multiple instances
2190 - Updated many command line parameters
2191 - Introduced new OS API version 15
2192 - No longer support a default hypervisor
2193 - Treat virtual LVs as inexistent
2194 - Improved job locking logic to reduce lock contention
2195 - Match instance and node names case insensitively
2196 - Reimplemented bash completion script to be more complete
2203 *(Released Thu, 4 Feb 2010)*
2205 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2206 - Fix a string formatting bug
2207 - Improve safety of the code in some error paths
2208 - Improve data validation in the master of values returned from nodes
2214 *(Released Thu, 17 Dec 2009)*
2216 - Fix security issue due to missing validation of iallocator names; this
2217 allows local and remote execution of arbitrary executables
2218 - Fix failure of gnt-node list during instance removal
2219 - Ship the RAPI documentation in the archive
2225 *(Released Wed, 30 Sep 2009)*
2227 - Fixed many wrong messages
2228 - Fixed a few bugs related to the locking library
2229 - Fixed MAC checking at instance creation time
2230 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2231 - Fixed a few issues related to signal handling in both daemons and
2233 - Fixed the example startup script provided
2234 - Fixed insserv dependencies in the example startup script (patch from
2236 - Fixed handling of drained nodes in the iallocator framework
2237 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2239 - Fixed error related to invalid job IDs in job polling
2240 - Fixed job/opcode persistence on unclean master shutdown
2241 - Fixed handling of partial job processing after unclean master
2243 - Fixed error reporting from LUs, previously all errors were converted
2244 into execution errors
2245 - Fixed error reporting from burnin
2246 - Decreased significantly the memory usage of the job queue
2247 - Optimised slightly multi-job submission
2248 - Optimised slightly opcode loading
2249 - Backported the multi-job submit framework from the development
2250 branch; multi-instance start and stop should be faster
2251 - Added script to clean archived jobs after 21 days; this will reduce
2252 the size of the queue directory
2253 - Added some extra checks in disk size tracking
2254 - Added an example ethers hook script
2255 - Added a cluster parameter that prevents Ganeti from modifying of
2257 - Added more node information to RAPI responses
2258 - Added a ``gnt-job watch`` command that allows following the ouput of a
2260 - Added a bind-address option to ganeti-rapi
2261 - Added more checks to the configuration verify
2262 - Enhanced the burnin script such that some operations can be retried
2264 - Converted instance reinstall to multi-instance model
2270 *(Released Fri, 7 Aug 2009)*
2272 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2273 to allow using the pre-2.0.2 behaviour in activation, if any existing
2274 instances have mismatched disk sizes in the configuration
2275 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2276 any configuration mismatches for disk sizes
2277 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2278 failover to work on two-node clusters
2279 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2281 - Fixed detection of OS script errors in ``gnt-backup export``
2282 - Fixed exit code of ``gnt-backup export``
2288 *(Released Fri, 17 Jul 2009)*
2290 - Added experimental support for stripped logical volumes; this should
2291 enhance performance but comes with a higher complexity in the block
2292 device handling; stripping is only enabled when passing
2293 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2294 affected even in the non-stripped mode
2295 - Improved resiliency against transient failures at the end of DRBD
2296 resyncs, and in general of DRBD resync checks
2297 - Fixed a couple of issues with exports and snapshot errors
2298 - Fixed a couple of issues in instance listing
2299 - Added display of the disk size in ``gnt-instance info``
2300 - Fixed checking for valid OSes in instance creation
2301 - Fixed handling of the "vcpus" parameter in instance listing and in
2302 general of invalid parameters
2303 - Fixed http server library, and thus RAPI, to handle invalid
2304 username/password combinations correctly; this means that now they
2305 report unauthorized for queries too, not only for modifications,
2306 allowing earlier detect of configuration problems
2307 - Added a new "role" node list field, equivalent to the master/master
2308 candidate/drained/offline flags combinations
2309 - Fixed cluster modify and changes of candidate pool size
2310 - Fixed cluster verify error messages for wrong files on regular nodes
2311 - Fixed a couple of issues with node demotion from master candidate role
2312 - Fixed node readd issues
2313 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2314 - Added a new ``--no-voting`` option for masterfailover to fix failover
2315 on two-nodes clusters when the former master node is unreachable
2316 - Added instance reinstall over RAPI
2322 *(Released Tue, 16 Jun 2009)*
2324 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2325 allow re-adding the start in single-user option (regression from 1.2)
2326 - the watcher writes the instance status to a file, to allow monitoring
2327 to report the instance status (from the master) based on cached
2328 results of the watcher's queries; while this can get stale if the
2329 watcher is being locked due to other work on the cluster, this is
2330 still an improvement
2331 - the watcher now also restarts the node daemon and the rapi daemon if
2333 - fixed the watcher to handle full and drained queue cases
2334 - hooks export more instance data in the environment, which helps if
2335 hook scripts need to take action based on the instance's properties
2336 (no longer need to query back into ganeti)
2337 - instance failovers when the instance is stopped do not check for free
2338 RAM, so that failing over a stopped instance is possible in low memory
2340 - rapi uses queries for tags instead of jobs (for less job traffic), and
2341 for cluster tags it won't talk to masterd at all but read them from
2343 - a couple of error handling fixes in RAPI
2344 - drbd handling: improved the error handling of inconsistent disks after
2345 resync to reduce the frequency of "there are some degraded disks for
2346 this instance" messages
2347 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2348 error handling path called a wrong function name)
2354 *(Released Wed, 27 May 2009)*
2356 - no changes from rc5
2362 *(Released Wed, 20 May 2009)*
2364 - fix a couple of bugs (validation, argument checks)
2365 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2366 - some small improvements to RAPI and IAllocator
2367 - make watcher automatically start the master daemon if down
2373 *(Released Mon, 27 Apr 2009)*
2375 - change the OS list to not require locks; this helps with big clusters
2376 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2377 volume group is broken
2378 - ``gnt-instance info``, without any arguments, doesn't run for all
2379 instances anymore; either pass ``--all`` or pass the desired
2380 instances; this helps against mistakes on big clusters where listing
2381 the information for all instances takes a long time
2382 - miscellaneous doc and man pages fixes
2388 *(Released Wed, 8 Apr 2009)*
2390 - Change the internal locking model of some ``gnt-node`` commands, in
2391 order to reduce contention (and blocking of master daemon) when
2392 batching many creation/reinstall jobs
2393 - Fixes to Xen soft reboot
2394 - No longer build documentation at build time, instead distribute it in
2395 the archive, in order to reduce the need for the whole docbook/rst
2402 *(Released Fri, 27 Mar 2009)*
2404 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2405 - Fix watcher startup sequence, improves the behaviour of busy clusters
2406 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2407 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2408 bind address and other places
2409 - Some documentation fixes and updates
2415 *(Released Mon, 2 Mar 2009)*
2417 - More documentation updates, now all docs should be more-or-less
2419 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2421 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2423 - Fix an issue related to $libdir/run/ganeti and cluster creation
2429 *(Released Thu, 19 Feb 2009)*
2431 - Xen PVM and KVM have switched the default value for the instance root
2432 disk to the first partition on the first drive, instead of the whole
2433 drive; this means that the OS installation scripts must be changed
2435 - Man pages have been updated
2436 - RAPI has been switched by default to HTTPS, and the exported functions
2437 should all work correctly
2438 - RAPI v1 has been removed
2439 - Many improvements to the KVM hypervisor
2440 - Block device errors are now better reported
2441 - Many other bugfixes and small improvements
2447 *(Released Mon, 26 Jan 2009)*
2449 - Version 2 is a general rewrite of the code and therefore the
2450 differences are too many to list, see the design document for 2.0 in
2451 the ``doc/`` subdirectory for more details
2452 - In this beta version there is not yet a migration path from 1.2 (there
2453 will be one in the final 2.0 release)
2454 - A few significant changes are:
2456 - all commands are executed by a daemon (``ganeti-masterd``) and the
2457 various ``gnt-*`` commands are just front-ends to it
2458 - all the commands are entered into, and executed from a job queue,
2459 see the ``gnt-job(8)`` manpage
2460 - the RAPI daemon supports read-write operations, secured by basic
2461 HTTP authentication on top of HTTPS
2462 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2463 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2464 to migrate to DRBD 8 first while still running Ganeti 1.2)
2465 - DRBD devices are using statically allocated minor numbers, which
2466 will be assigned to existing instances during the migration process
2467 - there is support for both Xen PVM and Xen HVM instances running on
2469 - KVM virtualization is supported too
2470 - file-based storage has been implemented, which means that it is
2471 possible to run the cluster without LVM and DRBD storage, for
2472 example using a shared filesystem exported from shared storage (and
2473 still have live migration)
2479 *(Released Tue, 13 Jan 2009)*
2481 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2482 - Reuse the old instance mac address by default on instance import, if
2483 the instance name is the same.
2484 - Handle situations in which the node info rpc returns incomplete
2486 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2487 - Improved version of batcher:
2489 - state file support
2490 - instance mac address support
2491 - support for HVM clusters/instances
2493 - Add an option to show the number of cpu sockets and nodes in
2495 - Support OSes that handle more than one version of the OS api (but do
2496 not change the current API in any other way)
2497 - Fix ``gnt-node migrate``
2498 - ``gnt-debug`` man page
2499 - Fixes various more typos and small issues
2500 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2506 *(Released Wed, 24 Sep 2008)*
2508 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2509 type of disk exported to fully virtualized instances.
2510 - provide access to the serial console of HVM instances
2511 - instance auto_balance flag, set by default. If turned off it will
2512 avoid warnings on cluster verify if there is not enough memory to fail
2513 over an instance. in the future it will prevent automatically failing
2514 it over when we will support that.
2515 - batcher tool for instance creation, see ``tools/README.batcher``
2516 - ``gnt-instance reinstall --select-os`` to interactively select a new
2517 operating system when reinstalling an instance.
2518 - when changing the memory amount on instance modify a check has been
2519 added that the instance will be able to start. also warnings are
2520 emitted if the instance will not be able to fail over, if auto_balance
2522 - documentation fixes
2523 - sync fields between ``gnt-instance list/modify/add/import``
2524 - fix a race condition in drbd when the sync speed was set after giving
2525 the device a remote peer.
2531 *(Released Tue, 22 Jul 2008)*
2533 - note: the allowed size and number of tags per object were reduced
2534 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2535 - fixed twisted 8.x compatibility
2536 - fixed ``gnt-instance replace-disks`` with iallocator
2537 - add TCP keepalives on twisted connections to detect restarted nodes
2538 - disk increase support, see ``gnt-instance grow-disk``
2539 - implement bulk node/instance query for RAPI
2540 - add tags in node/instance listing (optional)
2541 - experimental migration (and live migration) support, read the man page
2542 for ``gnt-instance migrate``
2543 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2544 has some small improvements in handling its state file
2550 *(Released Fri, 13 Jun 2008)*
2552 - Experimental readonly, REST-based remote API implementation;
2553 automatically started on master node, TCP port 5080, if enabled by
2554 ``--enable-rapi`` parameter to configure script.
2555 - Instance allocator support. Add and import instance accept a
2556 ``--iallocator`` parameter, and call that instance allocator to decide
2557 which node to use for the instance. The iallocator document describes
2558 what's expected from an allocator script.
2559 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2560 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2561 node is lost there is still enough memory to fail over the instances
2563 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2564 ``gnt-cluster verify``, to check for site-specific compliance. All the
2565 hooks will run, and their output, if any, will be displayed. Any
2566 failing hook will make the verification return an error value.
2567 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2568 primary and secondary interfaces
2569 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2570 that is still declared as part of the cluster and has failed.
2571 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2572 specifying output fields, that just adds the chosen fields to the
2574 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2575 export from the filesystem.
2576 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2577 have been added. Using them you can enable/disable acpi and pae
2578 support, and specify a path for a cd image to be exported to the
2579 instance. These parameters as the name suggest only work on HVM
2581 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2582 PAE support will be set to the previously hardcoded values, but the
2583 (previously hardcoded) path to the CDROM ISO image will be unset and
2584 if required, needs to be set manually with ``gnt-instance modify``
2586 - The address to which an instance's VNC console is bound is now
2587 selectable per-instance, rather than being cluster wide. Of course
2588 this only applies to instances controlled via VNC, so currently just
2589 applies to HVM clusters.
2595 *(Released Mon, 18 Feb 2008)*
2597 - more tweaks to the disk activation code (especially helpful for DRBD)
2598 - change the default ``gnt-instance list`` output format, now there is
2599 one combined status field (see the manpage for the exact values this
2601 - some more fixes for the mac export to hooks change
2602 - make Ganeti not break with DRBD 8.2.x (which changed the version
2603 format in ``/proc/drbd``) (issue 24)
2604 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2605 template, allowing migration from DRBD0.7+MD to DRBD8
2611 *(Released Wed, 30 Jan 2008)*
2613 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2615 - add command aliases infrastructure and a few aliases
2616 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2617 man pages and the ``--help`` option of ``gnt-node
2618 list``/``gnt-instance list``
2619 - fix ``gnt-backup list`` with down nodes (issue 21)
2620 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2621 - fix the dist archive and add a check for including svn/git files in
2623 - some developer-related changes: improve the burnin and the QA suite,
2624 add an upload script for testing during development
2630 *(Released Wed, 16 Jan 2008)*
2632 - experimental HVM support, read the install document, section
2633 "Initializing the cluster"
2634 - allow for the PVM hypervisor per-instance kernel and initrd paths
2635 - add a new command ``gnt-cluster verify-disks`` which uses a new
2636 algorithm to improve the reconnection of the DRBD pairs if the device
2637 on the secondary node has gone away
2638 - make logical volume code auto-activate LVs at disk activation time
2639 - slightly improve the speed of activating disks
2640 - allow specification of the MAC address at instance creation time, and
2641 changing it later via ``gnt-instance modify``
2642 - fix handling of external commands that generate lots of output on
2644 - update documentation with regard to minimum version of DRBD8 supported
2650 *(Released Tue, 4 Dec 2007)*
2652 - Log the ``xm create`` output to the node daemon log on failure (to
2653 help diagnosing the error)
2654 - In debug mode, log all external commands output if failed to the logs
2655 - Change parsing of lvm commands to ignore stderr
2661 *(Released Wed, 28 Nov 2007)*
2663 - Another round of updates to the DRBD 8 code to deal with more failures
2664 in the replace secondary node operation
2665 - Some more logging of failures in disk operations (lvm, drbd)
2666 - A few documentation updates
2673 *(Released Tue, 13 Nov 2007)*
2675 - Change configuration file format from Python's Pickle to JSON.
2676 Upgrading is possible using the cfgupgrade utility.
2677 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2678 faster replace disks and is more stable (DRBD 8 has many improvements
2679 compared to DRBD 0.7)
2680 - Added command line tags support (see man pages for ``gnt-instance``,
2681 ``gnt-node``, ``gnt-cluster``)
2682 - Added instance rename support
2683 - Added multi-instance startup/shutdown
2684 - Added cluster rename support
2685 - Added ``gnt-node evacuate`` to simplify some node operations
2686 - Added instance reboot operation that can speedup reboot as compared to
2688 - Soften the requirement that hostnames are in FQDN format
2689 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2691 - Removed dependency on debian's patched fping that uses the
2692 non-standard ``-S`` option
2693 - Now the OS definitions are searched for in multiple, configurable
2694 paths (easier for distros to package)
2695 - Some changes to the hooks infrastructure (especially the new
2696 post-configuration update hook)
2697 - Other small bugfixes
2699 .. vim: set textwidth=72 syntax=rst :