8 *(Released Mon, 24 Jun 2013)*
10 Incompatible/important changes
11 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
13 - Instance policy can contain multiple instance specs, as described in
14 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
15 <design-partitioned>`. As a consequence, it's not possible to partially change
16 or override instance specs. Bounding specs (min and max) can be specified as a
17 whole using the new option ``--ipolicy-bounds-specs``, while standard
18 specs use the new option ``--ipolicy-std-specs``.
19 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
20 gnt-instance is a valid YAML object.
21 - hail now honors network restrictions when allocating nodes. This led to an
22 update of the IAllocator protocol. See the IAllocator documentation for
28 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
29 to always require authentication, ``--require-authentication``. It can
30 be specified in ``$sysconfdir/default/ganeti``.
31 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
32 be used to manage the disk templates to be used by instances in the cluster.
33 Initially, it will be set to a list that includes plain, drbd, if they were
34 enabled by specifying a volume group name, and file and sharedfile, if those
35 were enabled at configure time. Additionally, it will include all disk
36 templates that are currently used by instances. The order of disk templates
37 will be based on Ganeti's history of supporting them. In the future, the
38 first entry of the list will be used as a default disk template on instance
40 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
41 configuration back to the previous stable version.
42 - Disk templates in group ipolicy can be restored to the default value.
43 - Initial support for diskless instances and virtual clusters in QA.
44 - More QA and unit tests for instance policies.
45 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
46 describing why the opcode itself was executed.
47 - The monitoring daemon is now available. It allows users to query the cluster
48 for obtaining information about the status of the system. The daemon is only
49 responsible for providing the information over the network: the actual data
50 gathering is performed by data collectors (currently, only the DRBD status
51 collector is available).
52 - In order to help developers work on Ganeti, a new script
53 (``devel/build_chroot``) is provided, for building a chroot that contains all
54 the required development libraries and tools for compiling Ganeti on a Debian
56 - A new tool, ``harep``, for performing self-repair and recreation of instances
57 in Ganeti has been added.
58 - Split queries are enabled for tags, network, exports, cluster info, groups,
60 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
61 It prints the command line to set the current policies, to ease
66 The following new dependencies have been added:
69 - The ``curl`` library is not optional anymore for compiling the Haskell code.
70 - ``snap-server`` library (if monitoring is enabled).
73 - The minimum Python version needed to run Ganeti is now 2.6.
74 - ``yaml`` library (only for running the QA).
80 *(Released Thu, 04 Jul 2013)*
82 Incompatible/important changes
83 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
85 - Instance policies for disk size were documented to be on a per-disk
86 basis, but hail applied them to the sum of all disks. This has been
88 - ``hbal`` will now exit with status 0 if, during job execution over
89 LUXI, early exit has been requested and all jobs are successful;
90 before, exit status 1 was used, which cannot be differentiated from
92 - Compatibility with newer versions of rbd has been fixed
93 - ``gnt-instance batch-create`` has been changed to use the bulk create
94 opcode from Ganeti. This lead to incompatible changes in the format of
95 the JSON file. It's now not a custom dict anymore but a dict
96 compatible with the ``OpInstanceCreate`` opcode.
97 - Parent directories for file storage need to be listed in
98 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
99 write the file automatically based on old configuration values, but it
100 can not distribute it across all nodes and the file contents should be
101 verified. Use ``gnt-cluster copyfile
102 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
103 upgraded. The reason for requiring this list of paths now is that
104 before it would have been possible to inject new paths via RPC,
105 allowing files to be created in arbitrary locations. The RPC protocol
106 is protected using SSL/X.509 certificates, but as a design principle
107 Ganeti does not permit arbitrary paths to be passed.
108 - The parsing of the variants file for OSes (see
109 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
110 lines and comment lines (starting with ``#``) are ignored for better
112 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
113 longer available. ``gnt-node add`` now invokes a new tool on the
114 destination node, named ``prepare-node-join``, to configure the SSH
115 daemon. Paramiko is no longer necessary to configure nodes' SSH
116 daemons via ``gnt-node add``.
117 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
118 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
119 the flag is not reset after a master failover.
120 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
121 is recommended. 2.6 will be mandatory from the 2.8 series.
127 - New network management functionality to support automatic allocation
128 of IP addresses and managing of network parameters. See
129 :manpage:`gnt-network(8)` for more details.
130 - New external storage backend, to allow managing arbitrary storage
131 systems external to the cluster. See
132 :manpage:`ganeti-extstorage-interface(7)`.
133 - New ``exclusive-storage`` node parameter added, restricted to
134 nodegroup level. When it's set to true, physical disks are assigned in
135 an exclusive fashion to instances, as documented in :doc:`Partitioned
136 Ganeti <design-partitioned>`. Currently, only instances using the
137 ``plain`` disk template are supported.
138 - The KVM hypervisor has been updated with many new hypervisor
139 parameters, including a generic one for passing arbitrary command line
140 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
141 compatible up to qemu 1.4.
142 - A new tool, called ``mon-collector``, is the stand-alone executor of
143 the data collectors for a monitoring system. As of this version, it
144 just includes the DRBD data collector, that can be executed by calling
145 ``mon-collector`` using the ``drbd`` parameter. See
146 :manpage:`mon-collector(7)`.
147 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
148 for RAPI users. It allows granting permissions to query for
149 information to a specific user without giving
150 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
151 - A new tool named ``node-cleanup`` has been added. It cleans remains of
152 a cluster from a machine by stopping all daemons, removing
153 certificates and ssconf files. Unless the ``--no-backup`` option is
154 given, copies of the certificates are made.
155 - Instance creations now support the use of opportunistic locking,
156 potentially speeding up the (parallel) creation of multiple instances.
157 This feature is currently only available via the :doc:`RAPI
158 <rapi>` interface and when an instance allocator is used. If the
159 ``opportunistic_locking`` parameter is set the opcode will try to
160 acquire as many locks as possible, but will not wait for any locks
161 held by other opcodes. If not enough resources can be found to
162 allocate the instance, the temporary error code
163 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
164 retried thereafter, with or without opportunistic locking.
165 - New experimental linux-ha resource scripts.
166 - Restricted-commands support: ganeti can now be asked (via command line
167 or rapi) to perform commands on a node. These are passed via ganeti
168 RPC rather than ssh. This functionality is restricted to commands
169 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
170 reasons. The file is not copied automatically.
176 - Diskless instances are now externally mirrored (Issue 237). This for
177 now has only been tested in conjunction with explicit target nodes for
179 - Queries not needing locks or RPC access to the node can now be
180 performed by the confd daemon, making them independent from jobs, and
181 thus faster to execute. This is selectable at configure time.
182 - The functionality for allocating multiple instances at once has been
183 overhauled and is now also available through :doc:`RAPI <rapi>`.
185 There are no significant changes from version 2.7.0~rc3.
191 *(Released Tue, 25 Jun 2013)*
193 - Fix permissions on the confd query socket (Issue 477)
194 - Fix permissions on the job archive dir (Issue 498)
195 - Fix handling of an internal exception in replace-disks (Issue 472)
196 - Fix gnt-node info handling of shortened names (Issue 497)
197 - Fix gnt-instance grow-disk when wiping is enabled
198 - Documentation improvements, and support for newer pandoc
199 - Fix hspace honoring ipolicy for disks (Issue 484)
200 - Improve handling of the ``kvm_extra`` HV parameter
206 *(Released Fri, 24 May 2013)*
208 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
210 - Disks added through ``gnt-instance modify`` or created through
211 ``gnt-instance recreate-disks`` are wiped, if the
212 ``prealloc_wipe_disks`` flag is set.
213 - If wiping newly created disks fails, the disks are removed. Also,
214 partial failures in creating disks through ``gnt-instance modify``
215 triggers a cleanup of the partially-created disks.
216 - Removing the master IP address doesn't fail if the address has been
218 - Fix ownership of the OS log dir
219 - Workaround missing SO_PEERCRED constant (Issue 191)
225 *(Released Fri, 3 May 2013)*
227 This was the first release candidate of the 2.7 series. Since beta3:
229 - Fix kvm compatibility with qemu 1.4 (Issue 389)
230 - Documentation updates (admin guide, upgrade notes, install
231 instructions) (Issue 372)
232 - Fix gnt-group list nodes and instances count (Issue 436)
233 - Fix compilation without non-mandatory libraries (Issue 441)
234 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
235 - Make confd logging more verbose at INFO level (Issue 435)
236 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
237 - Fix failure path for instance storage type conversion (Issue 229)
238 - Update htools text backend documentation
239 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
240 - Disable inter-cluster instance move for file-based instances, because
241 it is dependant on instance export, which is not supported for
242 file-based instances. (Issue 414)
243 - Fix gnt-job crashes on non-ascii characters (Issue 427)
244 - Fix volume group checks on non-vm-capable nodes (Issue 432)
250 *(Released Mon, 22 Apr 2013)*
252 This was the third beta release of the 2.7 series. Since beta2:
254 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
255 - Fix data loss on wrong usage of ``gnt-instance move``
256 - Properly export errors in confd-based job queries
257 - Add ``users-setup`` tool
258 - Fix iallocator protocol to report 0 as a disk size for diskless
259 instances. This avoids hail breaking when a diskless instance is
261 - Fix job queue directory permission problem that made confd job queries
262 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
263 for access to archived jobs (Issue 406).
264 - Limit the sizes of networks supported by ``gnt-network`` to something
265 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
266 - Fix bugs in instance disk template conversion
267 - Fix GHC 7 compatibility
268 - Fix ``burnin`` install path (Issue 426).
269 - Allow very small disk grows (Issue 347).
270 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
271 that noded doesn't import masterd code (Issue 419).
272 - Make sure the default metavg at cluster init is the same as the vg, if
273 unspecified (Issue 358).
274 - Fix cleanup of partially created disks (part of Issue 416)
280 *(Released Tue, 2 Apr 2013)*
282 This was the second beta release of the 2.7 series. Since beta1:
284 - Networks no longer have a "type" slot, since this information was
285 unused in Ganeti: instead of it tags should be used.
286 - The rapi client now has a ``target_node`` option to MigrateInstance.
287 - Fix early exit return code for hbal (Issue 386).
288 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
289 - Fix ``rbd showmapped`` output parsing (Issue 312).
290 - Networks are now referenced indexed by UUID, rather than name. This
291 will require running cfgupgrade, from 2.7.0beta1, if networks are in
293 - The OS environment now includes network information.
294 - Deleting of a network is now disallowed if any instance nic is using
295 it, to prevent dangling references.
296 - External storage is now documented in man pages.
297 - The exclusive_storage flag can now only be set at nodegroup level.
298 - Hbal can now submit an explicit priority with its jobs.
299 - Many network related locking fixes.
300 - Bump up the required pylint version to 0.25.1.
301 - Fix the ``no_remember`` option in RAPI client.
302 - Many ipolicy related tests, qa, and fixes.
303 - Many documentation improvements and fixes.
304 - Fix building with ``--disable-file-storage``.
305 - Fix ``-q`` option in htools, which was broken if passed more than
307 - Some haskell/python interaction improvements and fixes.
308 - Fix iallocator in case of missing LVM storage.
309 - Fix confd config load in case of ``--no-lvm-storage``.
310 - The confd/query functionality is now mentioned in the security
317 *(Released Wed, 6 Feb 2013)*
319 This was the first beta release of the 2.7 series. All important changes
320 are listed in the latest 2.7 entry.
326 *(Released Fri, 21 Dec 2012)*
328 Important behaviour change: hbal won't rebalance anymore instances which
329 have the ``auto_balance`` attribute set to false. This was the intention
330 all along, but until now it only skipped those from the N+1 memory
331 reservation (DRBD-specific).
333 A significant number of bug fixes in this release:
335 - Fixed disk adoption interaction with ipolicy checks.
336 - Fixed networking issues when instances are started, stopped or
337 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
338 - Fixed the warning in cluster verify for shared storage instances not
340 - Fixed removal of storage directory on shared file storage (issue 262).
341 - Fixed validation of LVM volume group name in OpClusterSetParams
342 (``gnt-cluster modify``) (issue 285).
343 - Fixed runtime memory increases (``gnt-instance modify -m``).
344 - Fixed live migration under Xen's ``xl`` mode.
345 - Fixed ``gnt-instance console`` with ``xl``.
346 - Fixed building with newer Haskell compiler/libraries.
347 - Fixed PID file writing in Haskell daemons (confd); this prevents
348 restart issues if confd was launched manually (outside of
349 ``daemon-util``) while another copy of it was running
350 - Fixed a type error when doing live migrations with KVM (issue 297) and
351 the error messages for failing migrations have been improved.
352 - Fixed opcode validation for the out-of-band commands (``gnt-node
354 - Fixed a type error when unsetting OS hypervisor parameters (issue
355 311); now it's possible to unset all OS-specific hypervisor
357 - Fixed the ``dry-run`` mode for many operations: verification of
358 results was over-zealous but didn't take into account the ``dry-run``
359 operation, resulting in "wrong" failures.
360 - Fixed bash completion in ``gnt-job list`` when the job queue has
361 hundreds of entries; especially with older ``bash`` versions, this
362 results in significant CPU usage.
364 And lastly, a few other improvements have been made:
366 - Added option to force master-failover without voting (issue 282).
367 - Clarified error message on lock conflict (issue 287).
368 - Logging of newly submitted jobs has been improved (issue 290).
369 - Hostname checks have been made uniform between instance rename and
371 - The ``--submit`` option is now supported by ``gnt-debug delay``.
372 - Shutting down the master daemon by sending SIGTERM now stops it from
373 processing jobs waiting for locks; instead, those jobs will be started
374 once again after the master daemon is started the next time (issue
376 - Support for Xen's ``xl`` program has been improved (besides the fixes
378 - Reduced logging noise in the Haskell confd daemon (only show one log
379 entry for each config reload, instead of two).
380 - Several man page updates and typo fixes.
386 *(Released Fri, 12 Oct 2012)*
388 A small bugfix release. Among the bugs fixed:
390 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
391 made the command unusable.
392 - Commands that issue many jobs don't fail anymore just because some jobs
393 take so long that other jobs are archived.
394 - Failures during ``gnt-instance reinstall`` are reflected by the exit
396 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
398 - When ``always_failover`` is set, ``--allow-failover`` is not required
399 in migrate commands anymore.
400 - ``bash_completion`` works even if extglob is disabled.
401 - Fixed bug with locks that made failover for RDB-based instances fail.
402 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
403 a random node instead of one based on the allocator metric.
404 - Support for newer versions of pylint and pep8.
405 - Hail doesn't fail anymore when trying to add an instance of type
406 ``file``, ``sharedfile`` or ``rbd``.
407 - Added new Makefile target to rebuild the whole distribution, so that
408 all files are included.
414 *(Released Fri, 27 Jul 2012)*
417 .. attention:: The ``LUXI`` protocol has been made more consistent
418 regarding its handling of command arguments. This, however, leads to
419 incompatibility issues with previous versions. Please ensure that you
420 restart Ganeti daemons soon after the upgrade, otherwise most
421 ``LUXI`` calls (job submission, setting/resetting the drain flag,
422 pausing/resuming the watcher, cancelling and archiving jobs, querying
423 the cluster configuration) will fail.
432 The current ``admin_up`` field, which used to denote whether an instance
433 should be running or not, has been removed. Instead, ``admin_state`` is
434 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
436 The rational behind this is that an instance being “down” can have
439 - it could be down during a reboot
440 - it could be temporarily be down for a reinstall
441 - or it could be down because it is deprecated and kept just for its
444 The previous Boolean state was making it difficult to do capacity
445 calculations: should Ganeti reserve memory for a down instance? Now, the
446 tri-state field makes it clear:
448 - in ``up`` and ``down`` state, all resources are reserved for the
449 instance, and it can be at any time brought up if it is down
450 - in ``offline`` state, only disk space is reserved for it, but not
453 The field can have an extra use: since the transition between ``up`` and
454 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
455 transition between ``offline`` and ``down`` is done via ``gnt-instance
456 modify``, it is possible to given different rights to users. For
457 example, owners of an instance could be allowed to start/stop it, but
458 not transition it out of the offline state.
460 Instance policies and specs
461 +++++++++++++++++++++++++++
463 In previous Ganeti versions, an instance creation request was not
464 limited on the minimum size and on the maximum size just by the cluster
465 resources. As such, any policy could be implemented only in third-party
466 clients (RAPI clients, or shell wrappers over ``gnt-*``
467 tools). Furthermore, calculating cluster capacity via ``hspace`` again
468 required external input with regards to instance sizes.
470 In order to improve these workflows and to allow for example better
471 per-node group differentiation, we introduced instance specs, which
474 - minimum instance disk size, disk count, memory size, cpu count
475 - maximum values for the above metrics
476 - and “standard” values (used in ``hspace`` to calculate the standard
479 The minimum/maximum values can be also customised at node-group level,
480 for example allowing more powerful hardware to support bigger instance
483 Beside the instance specs, there are a few other settings belonging to
484 the instance policy framework. It is possible now to customise, per
485 cluster and node-group:
487 - the list of allowed disk templates
488 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
489 - the maximum ratio of instance to spindles (see below for more
490 information) for local storage
492 All these together should allow all tools that talk to Ganeti to know
493 what are the ranges of allowed values for instances and the
494 over-subscription that is allowed.
496 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
497 instance configuration, and the physical CPU configuration from the
498 node. For the spindle ratios however, we didn't track before these
499 values, so new parameters have been added:
501 - a new node parameter ``spindle_count``, defaults to 1, customisable at
502 node group or node level
503 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
505 Note that spindles in this context doesn't need to mean actual
506 mechanical hard-drives; it's just a relative number for both the node
507 I/O capacity and instance I/O consumption.
509 Instance migration behaviour
510 ++++++++++++++++++++++++++++
512 While live-migration is in general desirable over failover, it is
513 possible that for some workloads it is actually worse, due to the
514 variable time of the “suspend” phase during live migration.
516 To allow the tools to work consistently over such instances (without
517 having to hard-code instance names), a new backend parameter
518 ``always_failover`` has been added to control the migration/failover
519 behaviour. When set to True, all migration requests for an instance will
520 instead fall-back to failover.
522 Instance memory ballooning
523 ++++++++++++++++++++++++++
525 Initial support for memory ballooning has been added. The memory for an
526 instance is no longer fixed (backend parameter ``memory``), but instead
527 can vary between minimum and maximum values (backend parameters
528 ``minmem`` and ``maxmem``). Currently we only change an instance's
531 - live migrating or failing over and instance and the target node
532 doesn't have enough memory
533 - user requests changing the memory via ``gnt-instance modify
539 In order to control the use of specific CPUs by instance, support for
540 controlling CPU pinning has been added for the Xen, HVM and LXC
541 hypervisors. This is controlled by a new hypervisor parameter
542 ``cpu_mask``; details about possible values for this are in the
543 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
544 VCPU-to-CPU mapping) form will work well only when all nodes in your
545 cluster have the same amount of CPUs.
550 Another area in which Ganeti was not customisable were the parameters
551 used for storage configuration, e.g. how many stripes to use for LVM,
552 DRBD resync configuration, etc.
554 To improve this area, we've added disks parameters, which are
555 customisable at cluster and node group level, and which allow to
556 specify various parameters for disks (DRBD has the most parameters
557 currently), for example:
559 - DRBD resync algorithm and parameters (e.g. speed)
560 - the default VG for meta-data volumes for DRBD
561 - number of stripes for LVM (plain disk template)
564 These parameters can be modified via ``gnt-cluster modify -D …`` and
565 ``gnt-group modify -D …``, and are used at either instance creation (in
566 case of LVM stripes, for example) or at disk “activation” time
569 Rados block device support
570 ++++++++++++++++++++++++++
572 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
573 denoted by the ``rbd`` disk template type. This is considered
574 experimental, feedback is welcome. For details on configuring it, see
575 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
580 The existing master IP functionality works well only in simple setups (a
581 single network shared by all nodes); however, if nodes belong to
582 different networks, then the ``/32`` setup and lack of routing
583 information is not enough.
585 To allow the master IP to function well in more complex cases, the
586 system was reworked as follows:
588 - a master IP netmask setting has been added
589 - the master IP activation/turn-down code was moved from the node daemon
591 - whether to run the Ganeti-supplied master IP script or a user-supplied
592 on is a ``gnt-cluster init`` setting
594 Details about the location of the standard and custom setup scripts are
595 in the man page :manpage:`gnt-cluster(8)`; for information about the
596 setup script protocol, look at the Ganeti-supplied script.
601 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
604 It is now possible to use TLS-protected connections, and when renewing
605 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
606 it is now possible to specify spice or spice CA certificates. Also, it
607 is possible to configure a password for SPICE sessions via the
608 hypervisor parameter ``spice_password_file``.
610 There are also new parameters to control the compression and streaming
611 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
612 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
613 for the spice parameters.
615 Lastly, it is now possible to see the SPICE connection information via
616 ``gnt-instance console``.
621 A new tool (``tools/ovfconverter``) has been added that supports
622 conversion between Ganeti and the `Open Virtualization Format
623 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
626 This relies on the ``qemu-img`` tool to convert the disk formats, so the
627 actual compatibility with other virtualization solutions depends on it.
632 The configuration query daemon (``ganeti-confd``) is now optional, and
633 has been rewritten in Haskell; whether to use the daemon at all, use the
634 Python (default) or the Haskell version is selectable at configure time
635 via the ``--enable-confd`` parameter, which can take one of the
636 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
637 daemon will result in a smaller footprint; for larger systems, we
638 welcome feedback on the Haskell version which might become the default
641 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
642 daemon running. The Python version doesn't implement the new call.
645 User interface changes
646 ~~~~~~~~~~~~~~~~~~~~~~
648 We have replaced the ``--disks`` option of ``gnt-instance
649 replace-disks`` with a more flexible ``--disk`` option, which allows
650 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
651 disk size and mode can be changed upon recreation (via ``gnt-instance
652 recreate-disks``, which accepts the same ``--disk`` option).
654 As many people are used to a ``show`` command, we have added that as an
655 alias to ``info`` on all ``gnt-*`` commands.
657 The ``gnt-instance grow-disk`` command has a new mode in which it can
658 accept the target size of the disk, instead of the delta; this can be
659 more safe since two runs in absolute mode will be idempotent, and
660 sometimes it's also easier to specify the desired size directly.
662 Also the handling of instances with regard to offline secondaries has
663 been improved. Instance operations should not fail because one of it's
664 secondary nodes is offline, even though it's safe to proceed.
666 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
667 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
668 minors to instance name.
673 RAPI coverage has improved, with (for example) new resources for
674 recreate-disks, node power-cycle, etc.
679 There is partial support for ``xl`` in the Xen hypervisor; feedback is
682 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
683 whether to still support Python 2.4 or move to Python 2.6 as minimum
686 Support for Fedora has been slightly improved; the provided example
687 init.d script should work better on it and the INSTALL file should
688 document the needed dependencies.
693 The deprecated ``QueryLocks`` LUXI request has been removed. Use
694 ``Query(what=QR_LOCK, ...)`` instead.
696 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
697 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
698 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
699 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
700 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
702 RAPI client: ``CertificateError`` now derives from
703 ``GanetiApiError``. This should make it more easy to handle Ganeti
706 Deprecation warnings due to PyCrypto/paramiko import in
707 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
708 make sure to run an up-to-date paramiko version, if you use this tool.
710 The QA scripts now depend on Python 2.5 or above (the main code base
711 still works with Python 2.4).
713 The configuration file (``config.data``) is now written without
714 indentation for performance reasons; if you want to edit it, it can be
715 re-formatted via ``tools/fmtjson``.
717 A number of bugs has been fixed in the cluster merge tool.
719 ``x509`` certification verification (used in import-export) has been
720 changed to allow the same clock skew as permitted by the cluster
721 verification. This will remove some rare but hard to diagnose errors in
728 *(Released Thu, 19 Jul 2012)*
730 Very few changes from rc4 to the final release, only bugfixes:
732 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
733 instance, fix CDROM booting for KVM instances)
734 - fixed node group modification of node parameters
735 - fixed issue in LUClusterVerifyGroup with multi-group clusters
736 - fixed generation of bash completion to ensure a stable ordering
743 *(Released Fri, 13 Jul 2012)*
745 Third release candidate for 2.6. The following changes were done from
748 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
749 - Fixed an inconsistency in the LUXI protocol with the provided
750 arguments (NOT backwards compatible)
751 - Fixed a bug with node groups ipolicy where ``min`` was greater than
752 the cluster ``std`` value
753 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
754 easier instance debugging on nodes (requires ``hconfd`` to work)
760 *(Released Tue, 03 Jul 2012)*
762 Second release candidate for 2.6. The following changes were done from
765 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
767 - Fixed a RAPI regression on missing beparams/memory
768 - Fixed redistribution of files on offline nodes
769 - Added possibility to run activate-disks even though secondaries are
770 offline. With this change it relaxes also the strictness on some other
771 commands which use activate disks internally:
772 * ``gnt-instance start|reboot|rename|backup|export``
773 - Made it possible to remove safely an instance if its secondaries are
775 - Made it possible to reinstall even though secondaries are offline
781 *(Released Mon, 25 Jun 2012)*
783 First release candidate for 2.6. The following changes were done from
786 - Fixed bugs with disk parameters and ``rbd`` templates as well as
788 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
789 behaviour. It supports now the modify operation
790 - ``hcheck`` implemented to analyze cluster health and possibility of
791 improving health by rebalance
792 - ``hbal`` has been improved in dealing with split instances
798 *(Released Mon, 11 Jun 2012)*
800 Second beta release of 2.6. The following changes were done from beta2
803 - Fixed ``daemon-util`` with non-root user models
804 - Fixed creation of plain instances with ``--no-wait-for-sync``
805 - Fix wrong iv_names when running ``cfgupgrade``
806 - Export more information in RAPI group queries
807 - Fixed bug when changing instance network interfaces
808 - Extended burnin to do NIC changes
809 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
810 - Changed default for DRBD barriers
811 - Fixed DRBD error reporting for syncer rate
812 - Verify the options on disk parameters
814 And of course various fixes to documentation and improved unittests and
821 *(Released Wed, 23 May 2012)*
823 First beta release of 2.6. The following changes were done from beta1 to
826 - integrated patch for distributions without ``start-stop-daemon``
827 - adapted example init.d script to work on Fedora
828 - fixed log handling in Haskell daemons
829 - adapted checks in the watcher for pycurl linked against libnss
830 - add partial support for ``xl`` instead of ``xm`` for Xen
831 - fixed a type issue in cluster verification
832 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
835 Plus integrated fixes from the 2.5 branch:
837 - fixed ``kvm-ifup`` to use ``/bin/bash``
838 - fixed parallel build failures
839 - KVM live migration when using a custom keymap
845 *(Released Tue, 24 Jul 2012)*
847 A small bugfix release, with no new features:
849 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
850 different default shell (e.g. Debian, Ubuntu)
851 - fixed KVM startup and live migration with a custom keymap (fixes Issue
852 243 and Debian bug #650664)
853 - fixed compatibility with KVM versions that don't support multiple boot
854 devices (fixes Issue 230 and Debian bug #624256)
856 Additionally, a few fixes were done to the build system (fixed parallel
857 build failures) and to the unittests (fixed race condition in test for
858 FileID functions, and the default enable/disable mode for QA test is now
865 *(Released Fri, 11 May 2012)*
867 A small bugfix release.
869 The main issues solved are on the topic of compatibility with newer LVM
872 - fixed parsing of ``lv_attr`` field
873 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
874 the ``--force`` flag is needed
876 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
877 to accept kernel 3.x too (was hardcoded to 2.6.*).
879 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
880 of hook results and that also made display of other errors suboptimal
881 was fixed; the code behaves now like 2.4 and earlier.
883 Another change in 2.5, the cleanup of the OS scripts environment, is too
884 aggressive: it removed even the ``PATH`` variable, which requires the OS
885 scripts to *always* need to export it. Since this is a bit too strict,
886 we now export a minimal PATH, the same that we export for hooks.
888 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
889 integrated into this release.
891 Finally, a few other miscellaneous changes were done (no new features,
892 just small improvements):
894 - Fix ``gnt-group --help`` display
895 - Fix hardcoded Xen kernel path
896 - Fix grow-disk handling of invalid units
897 - Update synopsis for ``gnt-cluster repair-disk-sizes``
898 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
904 *(Released Thu, 12 Apr 2012)*
906 Incompatible/important changes and bugfixes
907 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
909 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
910 resource's ``ip_check`` parameter changed from ``True`` to ``False``
911 to match the underlying LUXI interface.
912 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
913 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
914 not maintain backwards-compatibility as the underlying operation
915 changed in an incompatible way. The RAPI client can talk to old
916 servers, but it needs to be told so as the return value changed.
917 - When creating file-based instances via RAPI, the ``file_driver``
918 parameter no longer defaults to ``loop`` and must be specified.
919 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
921 - Support for the undocumented and deprecated RAPI instance creation
922 request format version 0 has been dropped. Use version 1, supported
923 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
924 - Pyparsing 1.4.6 or above is required, see :doc:`installation
925 documentation <install>`.
926 - The "cluster-verify" hooks are now executed per group by the
927 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
928 if you just run ``gnt-cluster verify``, which generates one opcode per
930 - The environment as passed to the OS scripts is cleared, and thus no
931 environment variables defined in the node daemon's environment will be
932 inherited by the scripts.
933 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
935 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
936 support operations involving multiple node groups.
937 - Offline nodes are ignored when failing over an instance.
938 - Support for KVM version 1.0, which changed the version reporting format
940 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
942 - ``Makefile`` is now compatible with Automake 1.11.2
943 - Includes all bugfixes made in the 2.4 series
948 - The ganeti-htools project has been merged into the ganeti-core source
949 tree and will be built as part of Ganeti (see :doc:`install-quick`).
950 - Implemented support for :doc:`shared storage <design-shared-storage>`.
951 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
952 GPT-style partition tables (requires `parted
953 <http://www.gnu.org/s/parted/>`_).
954 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
955 - Allowed adding tags on instance creation.
956 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
957 - Allow instances to be started in a paused state, enabling the user to
958 see the complete console output on boot using the console.
959 - Added new hypervisor flag to control default reboot behaviour
960 (``reboot_behavior``).
961 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
962 - Improved out-of-band management support:
964 - Added ``gnt-node health`` command reporting the health status of
966 - Added ``gnt-node power`` command to manage power status of nodes.
967 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
969 - Instance migration can fall back to failover if instance is not
971 - Filters can be used when listing nodes, instances, groups and locks;
972 see :manpage:`ganeti(7)` manpage.
973 - Added post-execution status as variables to :doc:`hooks <hooks>`
975 - Instance tags are exported/imported together with the instance.
976 - When given an explicit job ID, ``gnt-job info`` will work for archived
978 - Jobs can define dependencies on other jobs (not yet supported via
979 RAPI or command line, but used by internal commands and usable via
982 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
985 - Instance failover is now available as a RAPI resource
986 (``/2/instances/[instance_name]/failover``).
987 - ``gnt-instance info`` defaults to static information if primary node
989 - Opcodes have a new ``comment`` attribute.
990 - Added basic SPICE support to KVM hypervisor.
991 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
993 Node group improvements
994 ~~~~~~~~~~~~~~~~~~~~~~~
996 - ``gnt-cluster verify`` has been modified to check groups separately,
997 thereby improving performance.
998 - Node group support has been added to ``gnt-cluster verify-disks``,
999 which now operates per node group.
1000 - Watcher has been changed to work better with node groups.
1002 - One process and state file per node group.
1003 - Slow watcher in one group doesn't block other group's watcher.
1005 - Added new command, ``gnt-group evacuate``, to move all instances in a
1006 node group to other groups.
1007 - Added ``gnt-instance change-group`` to move an instance to another
1009 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1010 per-group operations.
1011 - Node groups can be tagged.
1012 - Some operations switch from an exclusive to a shared lock as soon as
1014 - Instance's primary and secondary nodes' groups are now available as
1015 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1016 and ``snodes.group.uuid``).
1021 - Numerous updates to documentation and manpages.
1023 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1025 - Some opcode/job results are now also documented, see :doc:`RAPI
1028 - A lockset's internal lock is now also visible in lock monitor.
1029 - Log messages from job queue workers now contain information about the
1030 opcode they're processing.
1031 - ``gnt-instance console`` no longer requires the instance lock.
1032 - A short delay when waiting for job changes reduces the number of LUXI
1033 requests significantly.
1034 - DRBD metadata volumes are overwritten with zeros during disk creation.
1035 - Out-of-band commands no longer acquire the cluster lock in exclusive
1037 - ``devel/upload`` now uses correct permissions for directories.
1043 *(Released Fri, 23 Mar 2012)*
1045 This was the sixth release candidate of the 2.5 series.
1051 *(Released Mon, 9 Jan 2012)*
1053 This was the fifth release candidate of the 2.5 series.
1059 *(Released Thu, 27 Oct 2011)*
1061 This was the fourth release candidate of the 2.5 series.
1067 *(Released Wed, 26 Oct 2011)*
1069 This was the third release candidate of the 2.5 series.
1075 *(Released Tue, 18 Oct 2011)*
1077 This was the second release candidate of the 2.5 series.
1083 *(Released Tue, 4 Oct 2011)*
1085 This was the first release candidate of the 2.5 series.
1091 *(Released Wed, 31 Aug 2011)*
1093 This was the third beta release of the 2.5 series.
1099 *(Released Mon, 22 Aug 2011)*
1101 This was the second beta release of the 2.5 series.
1107 *(Released Fri, 12 Aug 2011)*
1109 This was the first beta release of the 2.5 series.
1115 *(Released Thu, 27 Oct 2011)*
1117 - Fixed bug when parsing command line parameter values ending in
1119 - Fixed assertion error after unclean master shutdown
1120 - Disable HTTP client pool for RPC, significantly reducing memory usage
1122 - Fixed queue archive creation with wrong permissions
1128 *(Released Tue, 23 Aug 2011)*
1132 - Fixed documentation for importing with ``--src-dir`` option
1133 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1134 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1140 *(Released Fri, 5 Aug 2011)*
1142 Many bug-fixes and a few small features:
1144 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1145 issues when you tried to add an instance with two MAC addresses in one
1147 - KVM: fixed per-instance stored UID value
1148 - KVM: configure bridged NICs at migration start
1149 - KVM: Fix a bug where instance will not start with never KVM versions
1151 - Added OS search path to ``gnt-cluster info``
1152 - Fixed an issue with ``file_storage_dir`` where you were forced to
1153 provide an absolute path, but the documentation states it is a
1154 relative path, the documentation was right
1155 - Added a new parameter to instance stop/start called ``--no-remember``
1156 that will make the state change to not be remembered
1157 - Implemented ``no_remember`` at RAPI level
1158 - Improved the documentation
1159 - Node evacuation: don't call IAllocator if node is already empty
1160 - Fixed bug in DRBD8 replace disks on current nodes
1161 - Fixed bug in recreate-disks for DRBD instances
1162 - Moved assertion checking locks in ``gnt-instance replace-disks``
1163 causing it to abort with not owning the right locks for some situation
1164 - Job queue: Fixed potential race condition when cancelling queued jobs
1165 - Fixed off-by-one bug in job serial generation
1166 - ``gnt-node volumes``: Fix instance names
1167 - Fixed aliases in bash completion
1168 - Fixed a bug in reopening log files after being sent a SIGHUP
1169 - Added a flag to burnin to allow specifying VCPU count
1170 - Bugfixes to non-root Ganeti configuration
1176 *(Released Thu, 12 May 2011)*
1178 Many bug-fixes and a few new small features:
1180 - Fixed a bug related to log opening failures
1181 - Fixed a bug in instance listing with orphan instances
1182 - Fixed a bug which prevented resetting the cluster-level node parameter
1183 ``oob_program`` to the default
1184 - Many fixes related to the ``cluster-merge`` tool
1185 - Fixed a race condition in the lock monitor, which caused failures
1186 during (at least) creation of many instances in parallel
1187 - Improved output for gnt-job info
1188 - Removed the quiet flag on some ssh calls which prevented debugging
1190 - Improved the N+1 failure messages in cluster verify by actually
1191 showing the memory values (needed and available)
1192 - Increased lock attempt timeouts so that when executing long operations
1193 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1194 too early and thus prevent the use of the 'fair' mechanism
1195 - Changed instance query data (``gnt-instance info``) to not acquire
1196 locks unless needed, thus allowing its use on locked instance if only
1197 static information is asked for
1198 - Improved behaviour with filesystems that do not support rename on an
1200 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1201 kept locks on all nodes during the wipe, which is unneeded
1202 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1203 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1204 10GiB) which caused the wipe to fail right at the end in some cases
1205 - Fixed master IP activation when doing master failover with no-voting
1206 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1207 the master node itself
1208 - Fixed potential data-loss in under disk full conditions, where Ganeti
1209 wouldn't check correctly the return code and would consider
1210 partially-written files 'correct'
1211 - Fixed bug related to multiple VGs and DRBD disk replacing
1212 - Added new disk parameter ``metavg`` that allows placement of the meta
1213 device for DRBD in a different volume group
1214 - Fixed error handling in the node daemon when the system libc doesn't
1215 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1216 - Fixed lock release during replace-disks, which kept cluster-wide locks
1217 when doing disk replaces with an iallocator script
1218 - Added check for missing bridges in cluster verify
1219 - Handle EPIPE errors while writing to the terminal better, so that
1220 piping the output to e.g. ``less`` doesn't cause a backtrace
1221 - Fixed rare case where a ^C during Luxi calls could have been
1222 interpreted as server errors, instead of simply terminating
1223 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1225 - Added a few more parameters to the KVM hypervisor, allowing a second
1226 CDROM, custom disk type for CDROMs and a floppy image
1227 - Removed redundant message in instance rename when the name is given
1229 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1230 disks on new nodes, allowing recreation when the original instance
1231 nodes are completely gone
1232 - Added option when converting disk templates to DRBD to skip waiting
1233 for the resync, in order to make the instance available sooner
1234 - Added two new variables to the OS scripts environment (containing the
1236 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1237 to allow use of ``pvgrub`` as bootloader
1238 - Changed the instance memory modifications to only check out-of-memory
1239 conditions on memory increases, and turned the secondary node warnings
1240 into errors (they can still be overridden via ``--force``)
1241 - Fixed the handling of a corner case when the Python installation gets
1242 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1243 to execute a command that doesn't exist
1244 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1245 node of the instance returned failures during instance shutdown; this
1246 adds the option ``--ignore-consistency`` to gnt-instance move
1248 And as usual, various improvements to the error messages, documentation
1255 *(Released Wed, 09 Mar 2011)*
1257 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1258 the RAPI users file if run twice (even with ``--dry-run``).
1260 The release fixes that bug (nothing else changed).
1266 *(Released Mon, 07 Mar 2011)*
1268 Final 2.4.0 release. Just a few small fixes:
1270 - Fixed RAPI node evacuate
1271 - Fixed the kvm-ifup script
1272 - Fixed internal error handling for special job cases
1273 - Updated man page to specify the escaping feature for options
1279 *(Released Mon, 28 Feb 2011)*
1281 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1282 that this feature wiped the disks of the wrong instance, leading to loss
1287 - Fixed title of query field containing instance name
1288 - Expanded the glossary in the documentation
1289 - Fixed one unittest (internal issue)
1295 *(Released Mon, 21 Feb 2011)*
1297 A number of bug fixes plus just a couple functionality changes.
1299 On the user-visible side, the ``gnt-* list`` command output has changed
1300 with respect to "special" field states. The current rc1 style of display
1301 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1302 the default output mode special fields states are displayed as follows:
1304 - Offline resource: ``*``
1305 - Unavailable/not applicable: ``-``
1306 - Data missing (RPC failure): ``?``
1307 - Unknown field: ``??``
1309 Another user-visible change is the addition of ``--force-join`` to
1314 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1315 - Fixed regression in RAPI/instance reinstall where all parameters were
1316 required (instead of optional)
1317 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1318 - Fixed iallocator usage (offline nodes were not considered offline)
1319 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1320 - Fixed hypervisor and OS parameter validation with respect to
1321 non-vm_capable nodes
1322 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1324 - Fixed ``tools/listrunner`` with respect to agent-based usage
1330 *(Released Fri, 4 Feb 2011)*
1332 Many changes and fixes since the beta1 release. While there were some
1333 internal changes, the code has been mostly stabilised for the RC
1336 Note: the dumb allocator was removed in this release, as it was not kept
1337 up-to-date with the IAllocator protocol changes. It is recommended to
1338 use the ``hail`` command from the ganeti-htools package.
1340 Note: the 2.4 and up versions of Ganeti are not compatible with the
1341 0.2.x branch of ganeti-htools. You need to upgrade to
1342 ganeti-htools-0.3.0 (or later).
1344 Regressions fixed from 2.3
1345 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1347 - Fixed the ``gnt-cluster verify-disks`` command
1348 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1350 - Fixed disk adoption breakage
1351 - Fixed wrong headers in instance listing for field aliases
1356 - Fixed corner case in KVM handling of NICs
1357 - Fixed many cases of wrong handling of non-vm_capable nodes
1358 - Fixed a bug where a missing instance symlink was not possible to
1359 recreate with any ``gnt-*`` command (now ``gnt-instance
1360 activate-disks`` does it)
1361 - Fixed the volume group name as reported by ``gnt-cluster
1363 - Increased timeouts for the import-export code, hopefully leading to
1364 fewer aborts due network or instance timeouts
1365 - Fixed bug in ``gnt-node list-storage``
1366 - Fixed bug where not all daemons were started on cluster
1367 initialisation, but only at the first watcher run
1368 - Fixed many bugs in the OOB implementation
1369 - Fixed watcher behaviour in presence of instances with offline
1371 - Fixed instance list output for instances running on the wrong node
1372 - a few fixes to the cluster-merge tool, but it still cannot merge
1373 multi-node groups (currently it is not recommended to use this tool)
1379 - Improved network configuration for the KVM hypervisor
1380 - Added e1000 as a supported NIC for Xen-HVM
1381 - Improved the lvmstrap tool to also be able to use partitions, as
1382 opposed to full disks
1383 - Improved speed of disk wiping (the cluster parameter
1384 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1385 of instance creations
1386 - Added documentation for the OS parameters
1387 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1388 hypervisor is not responding
1389 - Added display of blacklisted and hidden OS information in
1390 ``gnt-cluster info``
1391 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1392 NIC and node parameters, which might create problems with currently
1393 invalid (but undetected) configuration files, but prevents validation
1394 failures when unrelated parameters are modified
1395 - Changed cluster initialisation to wait for the master daemon to become
1397 - Expanded the RAPI interface:
1399 - Added config redistribution resource
1400 - Added activation/deactivation of instance disks
1401 - Added export of console information
1403 - Implemented log file reopening on SIGHUP, which allows using
1404 logrotate(8) for the Ganeti log files
1405 - Added a basic OOB helper script as an example
1411 *(Released Fri, 14 Jan 2011)*
1416 - Fixed timezone issues when formatting timestamps
1417 - Added support for node groups, available via ``gnt-group`` and other
1419 - Added out-of-band framework and management, see :doc:`design
1420 document <design-oob>`
1421 - Removed support for roman numbers from ``gnt-node list`` and
1422 ``gnt-instance list``.
1423 - Allowed modification of master network interface via ``gnt-cluster
1424 modify --master-netdev``
1425 - Accept offline secondaries while shutting down instance disks
1426 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1427 - Added support for multiple LVM volume groups
1428 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1430 - Added commands to list available fields:
1432 - ``gnt-node list-fields``
1433 - ``gnt-group list-fields``
1434 - ``gnt-instance list-fields``
1436 - Updated documentation and man pages
1441 - Moved ``rapi_users`` file into separate directory, now named
1442 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1444 - Added new tool for running commands on many machines,
1445 ``tools/ganeti-listrunner``
1446 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1447 improving the ``gnt-instance console`` output
1448 - Allowed customisation of disk index separator at ``configure`` time
1449 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1450 - Added support for non-partitioned md disks in ``lvmstrap``
1451 - Added script to gracefully power off KVM instances
1452 - Split ``utils`` module into smaller parts
1453 - Changed query operations to return more detailed information, e.g.
1454 whether an information is unavailable due to an offline node. To use
1455 this new functionality, the LUXI call ``Query`` must be used. Field
1456 information is now stored by the master daemon and can be retrieved
1457 using ``QueryFields``. Instances, nodes and groups can also be queried
1458 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1459 exposed via RAPI). The following commands make use of this
1460 infrastructure change:
1462 - ``gnt-group list``
1463 - ``gnt-group list-fields``
1465 - ``gnt-node list-fields``
1466 - ``gnt-instance list``
1467 - ``gnt-instance list-fields``
1468 - ``gnt-debug locks``
1473 - New RAPI resources (see :doc:`rapi`):
1477 - ``/2/groups/[group_name]``
1478 - ``/2/groups/[group_name]/assign-nodes``
1479 - ``/2/groups/[group_name]/modify``
1480 - ``/2/groups/[group_name]/rename``
1481 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1485 - Implemented ``no_install`` for instance creation
1486 - Implemented OS parameters for instance reinstallation, allowing
1487 use of special settings on reinstallation (e.g. for preserving data)
1492 - Added IPv6 support in import/export
1493 - Pause DRBD synchronization while wiping disks on instance creation
1494 - Updated unittests and QA scripts
1495 - Improved network parameters passed to KVM
1496 - Converted man pages from docbook to reStructuredText
1502 *(Released Mon, 20 Dec 2010)*
1504 Released version 2.3.1~rc1 without any changes.
1510 *(Released Wed, 1 Dec 2010)*
1512 - impexpd: Disable OpenSSL compression in socat if possible (backport
1513 from master, commit e90739d625b, see :doc:`installation guide
1514 <install-quick>` for details)
1515 - Changed unittest coverage report to exclude test scripts
1516 - Added script to check version format
1522 *(Released Wed, 1 Dec 2010)*
1524 Released version 2.3.0~rc1 without any changes.
1530 *(Released Fri, 19 Nov 2010)*
1532 A number of bugfixes and documentation updates:
1534 - Update ganeti-os-interface documentation
1535 - Fixed a bug related to duplicate MACs or similar items which should be
1537 - Fix breakage in OS state modify
1538 - Reinstall instance: disallow offline secondaries (fixes bug related to
1539 OS changing but reinstall failing)
1540 - plus all the other fixes between 2.2.1 and 2.2.2
1546 *(Released Tue, 2 Nov 2010)*
1548 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1549 - Fixed master failover race with watcher
1550 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1552 - Accept previously stopped instance for export with instance removal
1553 - Simplify and extend the environment variables for instance OS scripts
1554 - Added new node flags, ``master_capable`` and ``vm_capable``
1555 - Added optional instance disk wiping prior during allocation. This is a
1556 cluster-wide option and can be set/modified using
1557 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1558 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1559 :doc:`install-quick`
1560 - Added a new watcher option (``--ignore-pause``)
1561 - Added option to ignore offline node on instance start/stop
1562 (``--ignore-offline``)
1563 - Allow overriding OS parameters with ``gnt-instance reinstall``
1564 - Added ability to change node's secondary IP address using ``gnt-node
1566 - Implemented privilege separation for all daemons except
1567 ``ganeti-noded``, see ``configure`` options
1568 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1570 - Implemented job priorities (see ``ganeti(7)`` manpage)
1571 - Ignore failures while shutting down instances during failover from
1573 - Exit daemon's bootstrap process only once daemon is ready
1574 - Export more information via ``LUInstanceQuery``/remote API
1575 - Improved documentation, QA and unittests
1576 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1577 restart if the file was created or changed
1578 - Added LUXI protocol version sent with each request and response,
1579 allowing detection of server/client mismatches
1580 - Moved the Python scripts among gnt-* and ganeti-* into modules
1581 - Moved all code related to setting up SSH to an external script,
1583 - Infrastructure changes for node group support in future versions
1589 *(Released Fri, 19 Nov 2010)*
1591 A few small bugs fixed, and some improvements to the build system:
1593 - Fix documentation regarding conversion to drbd
1594 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1596 - Fix error handling in node modify with multiple changes
1597 - Allow remote imports without checked names
1603 *(Released Tue, 19 Oct 2010)*
1605 - Disable SSL session ID cache in RPC client
1611 *(Released Thu, 14 Oct 2010)*
1613 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1614 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1616 - Fix problems with interaction between Curl and Python's HTTP server,
1617 resulting in increased speed in many RPC calls
1618 - Improve our release script to prevent breakage with older aclocal and
1625 *(Released Thu, 7 Oct 2010)*
1627 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1628 value retrieved from master
1629 - Added support for blacklisted or hidden OS definitions
1630 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1631 - Added support for -mem-path in KVM hypervisor abstraction layer
1632 - Allow overriding instance parameters in tool for inter-cluster
1633 instance moves (``tools/move-instance``)
1634 - Improved opcode summaries (e.g. in ``gnt-job list``)
1635 - Improve consistency of OS listing by sorting it
1636 - Documentation updates
1642 *(Released Fri, 8 Oct 2010)*
1644 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1650 *(Released Mon, 4 Oct 2010)*
1652 - Fixed regression in ``gnt-instance rename``
1658 *(Released Wed, 22 Sep 2010)*
1660 - Fixed OS_VARIANT variable for OS scripts
1661 - Fixed cluster tag operations via RAPI
1662 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1663 - Disabled RAPI CA checks in watcher
1669 *(Released Mon, 23 Aug 2010)*
1671 - Support DRBD versions of the format "a.b.c.d"
1673 - Re-introduce support for usage from multiple threads in RAPI client
1674 - Instance renames and modify via RAPI
1675 - Work around race condition between processing and archival in job
1677 - Mark opcodes following failed one as failed, too
1678 - Job field ``lock_status`` was removed due to difficulties making it
1679 work with the changed job queue in Ganeti 2.2; a better way to monitor
1680 locks is expected for a later 2.2.x release
1681 - Fixed dry-run behaviour with many commands
1682 - Support ``ssh-agent`` again when adding nodes
1683 - Many additional bugfixes
1689 *(Released Fri, 30 Jul 2010)*
1691 Important change: the internal RPC mechanism between Ganeti nodes has
1692 changed from using a home-grown http library (based on the Python base
1693 libraries) to use the PycURL library. This requires that PycURL is
1694 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1695 against GnuTLS by default. cURL's support for GnuTLS had known issues
1696 before cURL 7.21.0 and we recommend using the latest cURL release or
1697 linking against OpenSSL. Most other distributions already link PycURL
1698 and cURL against OpenSSL. The command::
1700 python -c 'import pycurl; print pycurl.version'
1702 can be used to determine the libraries PycURL and cURL are linked
1705 Other significant changes:
1707 - Rewrote much of the internals of the job queue, in order to achieve
1708 better parallelism; this decouples job query operations from the job
1709 processing, and it should allow much nicer behaviour of the master
1710 daemon under load, and it also has uncovered some long-standing bugs
1711 related to the job serialisation (now fixed)
1712 - Added a default iallocator setting to the cluster parameters,
1713 eliminating the need to always pass nodes or an iallocator for
1714 operations that require selection of new node(s)
1715 - Added experimental support for the LXC virtualization method
1716 - Added support for OS parameters, which allows the installation of
1717 instances to pass parameter to OS scripts in order to customise the
1719 - Added a hypervisor parameter controlling the migration type (live or
1720 non-live), since hypervisors have various levels of reliability; this
1721 has renamed the 'live' parameter to 'mode'
1722 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1723 logical volumes, meaning that cluster verify will ignore them and not
1724 flag their presence as errors
1725 - The watcher will now reset the error count for failed instances after
1726 8 hours, thus allowing self-healing if the problem that caused the
1727 instances to be down/fail to start has cleared in the meantime
1728 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1729 check for, and warn, if the drbd module parameter ``usermode_helper``
1730 is not consistent with the cluster-wide setting; this is needed to
1731 make diagnose easier of failed drbd creations
1732 - Started adding base IPv6 support, but this is not yet
1733 enabled/available for use
1734 - Rename operations (cluster, instance) will now return the new name,
1735 which is especially useful if a short name was passed in
1736 - Added support for instance migration in RAPI
1737 - Added a tool to pre-configure nodes for the SSH setup, before joining
1738 them to the cluster; this will allow in the future a simplified model
1739 for node joining (but not yet fully enabled in 2.2); this needs the
1740 paramiko python library
1741 - Fixed handling of name-resolving errors
1742 - Fixed consistency of job results on the error path
1743 - Fixed master-failover race condition when executed multiple times in
1745 - Fixed many bugs related to the job queue (mostly introduced during the
1746 2.2 development cycle, so not all are impacting 2.1)
1747 - Fixed instance migration with missing disk symlinks
1748 - Fixed handling of unknown jobs in ``gnt-job archive``
1749 - And many other small fixes/improvements
1753 - Enhanced both the unittest and the QA coverage
1754 - Switched the opcode validation to a generic model, and extended the
1755 validation to all opcode parameters
1756 - Changed more parts of the code that write shell scripts to use the
1758 - Switched the master daemon to use the asyncore library for the Luxi
1765 *(Released Thu, 17 Jun 2010)*
1767 - Added tool (``move-instance``) and infrastructure to move instances
1768 between separate clusters (see :doc:`separate documentation
1769 <move-instance>` and :doc:`design document <design-2.2>`)
1770 - Added per-request RPC timeout
1771 - RAPI now requires a Content-Type header for requests with a body (e.g.
1772 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1773 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1774 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1776 - Implemented initial support for running Ganeti daemons as separate
1777 users, see configure-time flags ``--with-user-prefix`` and
1778 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1780 - Instances can be removed after export (``gnt-backup export
1781 --remove-instance``)
1782 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1783 key (instead of 1024 bit)
1784 - Added new cluster configuration file for cluster domain secret
1785 - Import/export now use SSL instead of SSH
1786 - Added support for showing estimated time when exporting an instance,
1787 see the ``ganeti-os-interface(7)`` manpage and look for
1794 *(Released Tue, 16 Nov 2010)*
1796 Some more bugfixes. Unless critical bugs occur, this will be the last
1799 - Fix case of MAC special-values
1800 - Fix mac checker regex
1801 - backend: Fix typo causing "out of range" error
1802 - Add missing --units in gnt-instance list man page
1808 *(Released Tue, 24 Aug 2010)*
1811 - Don't ignore secondary node silently on non-mirrored disk templates
1813 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1814 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1815 - Properly document the usb_mouse parameter
1816 - Fix path in ganeti-rapi(8) (issue 116)
1817 - Adjust error message when the ganeti user's .ssh directory is
1819 - Add same-node-check when changing the disk template to drbd
1825 *(Released Fri, 16 Jul 2010)*
1828 - Add an option to only select some reboot types during qa/burnin.
1829 (on some hypervisors consequent reboots are not supported)
1830 - Fix infrequent race condition in master failover. Sometimes the old
1831 master ip address would be still detected as up for a short time
1832 after it was removed, causing failover to fail.
1833 - Decrease mlockall warnings when the ctypes module is missing. On
1834 Python 2.4 we support running even if no ctypes module is installed,
1835 but we were too verbose about this issue.
1836 - Fix building on old distributions, on which man doesn't have a
1838 - Fix RAPI not to ignore the MAC address on instance creation
1839 - Implement the old instance creation format in the RAPI client.
1845 *(Released Thu, 01 Jul 2010)*
1847 A small bugfix release:
1848 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1849 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1850 a non-existing option
1851 - Fix instance create: the --force-variant option was ignored
1852 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1853 - Fix modify node storage with non-FQDN arguments
1854 - Fix RAPI client to authenticate under Python 2.6 when used
1855 for more than 5 requests needing authentication
1856 - Fix gnt-instance modify -t (storage) giving a wrong error message
1857 when converting a non-shutdown drbd instance to plain
1863 *(Released Fri, 18 Jun 2010)*
1865 A small bugfix release:
1867 - Fix live migration of KVM instances started with older Ganeti
1868 versions which had fewer hypervisor parameters
1869 - Fix gnt-instance grow-disk on down instances
1870 - Fix an error-reporting bug during instance migration
1871 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1872 silently ignoring broken ones
1873 - Fix an RPC error reporting bug affecting, for example, RAPI client
1875 - Fix bug triggered by different API version os-es on different nodes
1876 - Fix a bug in instance startup with custom hvparams: OS level
1877 parameters would fail to be applied.
1878 - Fix the RAPI client under Python 2.6 (but more work is needed to
1879 make it work completely well with OpenSSL)
1880 - Fix handling of errors when resolving names from DNS
1886 *(Released Thu, 3 Jun 2010)*
1888 A medium sized development cycle. Some new features, and some
1889 fixes/small improvements/cleanups.
1891 Significant features
1892 ~~~~~~~~~~~~~~~~~~~~
1894 The node deamon now tries to mlock itself into memory, unless the
1895 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1896 its logs, and falls back to console logging. This allows emergency
1897 features such as ``gnt-node powercycle`` to work even in the event of a
1898 broken node disk (tested offlining the disk hosting the node's
1899 filesystem and dropping its memory caches; don't try this at home)
1901 KVM: add vhost-net acceleration support. It can be tested with a new
1902 enough version of the kernel and of qemu-kvm.
1904 KVM: Add instance chrooting feature. If you use privilege dropping for
1905 your VMs you can also now force them to chroot to an empty directory,
1906 before starting the emulated guest.
1908 KVM: Add maximum migration bandwith and maximum downtime tweaking
1909 support (requires a new-enough version of qemu-kvm).
1911 Cluster verify will now warn if the master node doesn't have the master
1912 ip configured on it.
1914 Add a new (incompatible) instance creation request format to RAPI which
1915 supports all parameters (previously only a subset was supported, and it
1916 wasn't possible to extend the old format to accomodate all the new
1917 features. The old format is still supported, and a client can check for
1918 this feature, before using it, by checking for its presence in the
1919 ``features`` RAPI resource.
1921 Now with ancient latin support. Try it passing the ``--roman`` option to
1922 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1923 (requires the python-roman module to be installed, in order to work).
1928 As usual many internal code refactorings, documentation updates, and
1931 - Lots of improvements and cleanups to the experimental Remote API
1932 (RAPI) client library.
1933 - A new unit test suite for the core daemon libraries.
1934 - A fix to creating missing directories makes sure the umask is not
1935 applied anymore. This enforces the same directory permissions
1937 - Better handling terminating daemons with ctrl+c (used when running
1938 them in debugging mode).
1939 - Fix a race condition in live migrating a KVM instance, when stat()
1940 on the old proc status file returned EINVAL, which is an unexpected
1942 - Fixed manpage checking with newer man and utf-8 charachters. But now
1943 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1949 *(Released Fri, 7 May 2010)*
1951 Fix a bug which prevented untagged KVM instances from starting.
1957 *(Released Fri, 7 May 2010)*
1959 Another release with a long development cycle, during which many
1960 different features were added.
1962 Significant features
1963 ~~~~~~~~~~~~~~~~~~~~
1965 The KVM hypervisor now can run the individual instances as non-root, to
1966 reduce the impact of a VM being hijacked due to bugs in the
1967 hypervisor. It is possible to run all instances as a single (non-root)
1968 user, to manually specify a user for each instance, or to dynamically
1969 allocate a user out of a cluster-wide pool to each instance, with the
1970 guarantee that no two instances will run under the same user ID on any
1973 An experimental RAPI client library, that can be used standalone
1974 (without the other Ganeti libraries), is provided in the source tree as
1975 ``lib/rapi/client.py``. Note this client might change its interface in
1976 the future, as we iterate on its capabilities.
1978 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1979 replace the cluster's certificates and crypto keys. This might help in
1980 case they have been compromised, or have simply expired.
1982 A new disk option for instance creation has been added that allows one
1983 to "adopt" currently existing logical volumes, with data
1984 preservation. This should allow easier migration to Ganeti from
1985 unmanaged (or managed via other software) instances.
1987 Another disk improvement is the possibility to convert between redundant
1988 (DRBD) and plain (LVM) disk configuration for an instance. This should
1989 allow better scalability (starting with one node and growing the
1990 cluster, or shrinking a two-node cluster to one node).
1992 A new feature that could help with automated node failovers has been
1993 implemented: if a node sees itself as offline (by querying the master
1994 candidates), it will try to shutdown (hard) all instances and any active
1995 DRBD devices. This reduces the risk of duplicate instances if an
1996 external script automatically failovers the instances on such nodes. To
1997 enable this, the cluster parameter ``maintain_node_health`` should be
1998 enabled; in the future this option (per the name) will enable other
1999 automatic maintenance features.
2001 Instance export/import now will reuse the original instance
2002 specifications for all parameters; that means exporting an instance,
2003 deleting it and the importing it back should give an almost identical
2004 instance. Note that the default import behaviour has changed from
2005 before, where it created only one NIC; now it recreates the original
2008 Cluster verify has added a few new checks: SSL certificates validity,
2009 /etc/hosts consistency across the cluster, etc.
2014 As usual, many internal changes were done, documentation fixes,
2017 - Fixed cluster initialization with disabled cluster storage (regression
2018 introduced in 2.1.1)
2019 - File-based storage supports growing the disks
2020 - Fixed behaviour of node role changes
2021 - Fixed cluster verify for some corner cases, plus a general rewrite of
2022 cluster verify to allow future extension with more checks
2023 - Fixed log spamming by watcher and node daemon (regression introduced
2025 - Fixed possible validation issues when changing the list of enabled
2027 - Fixed cleanup of /etc/hosts during node removal
2028 - Fixed RAPI response for invalid methods
2029 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2030 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2031 from ide disks, etc.)
2032 - Allow OS changes without re-installation (to record a changed OS
2033 outside of Ganeti, or to allow OS renames)
2034 - Allow instance creation without OS installation (useful for example if
2035 the OS will be installed manually, or restored from a backup not in
2037 - Implemented option to make cluster ``copyfile`` use the replication
2039 - Added list of enabled hypervisors to ssconf (possibly useful for
2041 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2043 - A partial form of node re-IP is possible via node readd, which now
2044 allows changed node primary IP
2045 - Command line utilities now show an informational message if the job is
2047 - The logs of the master daemon now show the PID/UID/GID of the
2054 *(Released Fri, 12 Mar 2010)*
2056 During the 2.1.0 long release candidate cycle, a lot of improvements and
2057 changes have accumulated with were released later as 2.1.1.
2062 The node evacuate command (``gnt-node evacuate``) was significantly
2063 rewritten, and as such the IAllocator protocol was changed - a new
2064 request type has been added. This unfortunate change during a stable
2065 series is designed to improve performance of node evacuations; on
2066 clusters with more than about five nodes and which are well-balanced,
2067 evacuation should proceed in parallel for all instances of the node
2068 being evacuated. As such, any existing IAllocator scripts need to be
2069 updated, otherwise the above command will fail due to the unknown
2070 request. The provided "dumb" allocator has not been updated; but the
2071 ganeti-htools package supports the new protocol since version 0.2.4.
2073 Another important change is increased validation of node and instance
2074 names. This might create problems in special cases, if invalid host
2075 names are being used.
2077 Also, a new layer of hypervisor parameters has been added, that sits at
2078 OS level between the cluster defaults and the instance ones. This allows
2079 customisation of virtualization parameters depending on the installed
2080 OS. For example instances with OS 'X' may have a different KVM kernel
2081 (or any other parameter) than the cluster defaults. This is intended to
2082 help managing a multiple OSes on the same cluster, without manual
2083 modification of each instance's parameters.
2085 A tool for merging clusters, ``cluster-merge``, has been added in the
2086 tools sub-directory.
2091 - Improved the int/float conversions that should make the code more
2092 robust in face of errors from the node daemons
2093 - Fixed the remove node code in case of internal configuration errors
2094 - Fixed the node daemon behaviour in face of inconsistent queue
2095 directory (e.g. read-only file-system where we can't open the files
2097 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2098 now it either aborts cleanly or, if given the new "auto_promote"
2099 parameter, will automatically promote other nodes as needed
2100 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2101 completely prevent Ganeti from working
2102 - Fixed bug for instance export when not all disks were successfully
2104 - Fixed behaviour of node add when the new node is slow in starting up
2106 - Fixed handling of signals in the LUXI client, which should improve
2107 behaviour of command-line scripts
2108 - Added checks for invalid node/instance names in the configuration (now
2109 flagged during cluster verify)
2110 - Fixed watcher behaviour for disk activation errors
2111 - Fixed two potentially endless loops in http library, which led to the
2112 RAPI daemon hanging and consuming 100% CPU in some cases
2113 - Fixed bug in RAPI daemon related to hashed passwords
2114 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2116 - Enhanced compatibility with non-Debian OSes, but not using absolute
2117 path in some commands and allowing customisation of the ssh
2118 configuration directory
2119 - Fixed possible future issue with new Python versions by abiding to the
2120 proper use of ``__slots__`` attribute on classes
2121 - Added checks that should prevent directory traversal attacks
2122 - Many documentation fixes based on feedback from users
2127 - Added an "early_release" more for instance replace disks and node
2128 evacuate, where we release locks earlier and thus allow higher
2129 parallelism within the cluster
2130 - Added watcher hooks, intended to allow the watcher to restart other
2131 daemons (e.g. from the ganeti-nbma project), but they can be used of
2132 course for any other purpose
2133 - Added a compile-time disable for DRBD barriers, to increase
2134 performance if the administrator trusts the power supply or the
2135 storage system to not lose writes
2136 - Added the option of using syslog for logging instead of, or in
2137 addition to, Ganeti's own log files
2138 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2139 can indeed boot from a paravirtual NIC
2140 - Added a generic debug level for many operations; while this is not
2141 used widely yet, it allows one to pass the debug value all the way to
2143 - Enhanced the hooks environment for instance moves (failovers,
2144 migrations) where the primary/secondary nodes changed during the
2145 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2146 - Enhanced data validations for many user-supplied values; one important
2147 item is the restrictions imposed on instance and node names, which
2148 might reject some (invalid) host names
2149 - Add a configure-time option to disable file-based storage, if it's not
2150 needed; this allows greater security separation between the master
2151 node and the other nodes from the point of view of the inter-node RPC
2153 - Added user notification in interactive tools if job is waiting in the
2154 job queue or trying to acquire locks
2155 - Added log messages when a job is waiting for locks
2156 - Added filtering by node tags in instance operations which admit
2157 multiple instances (start, stop, reboot, reinstall)
2158 - Added a new tool for cluster mergers, ``cluster-merge``
2159 - Parameters from command line which are of the form ``a=b,c=d`` can now
2160 use backslash escapes to pass in values which contain commas,
2161 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2163 - For KVM, the instance name is the first parameter passed to KVM, so
2164 that it's more visible in the process list
2170 *(Released Tue, 2 Mar 2010)*
2172 Ganeti 2.1 brings many improvements with it. Major changes:
2174 - Added infrastructure to ease automated disk repairs
2175 - Added new daemon to export configuration data in a cheaper way than
2176 using the remote API
2177 - Instance NICs can now be routed instead of being associated with a
2179 - Improved job locking logic to reduce impact of jobs acquiring multiple
2180 locks waiting for other long-running jobs
2182 In-depth implementation details can be found in the Ganeti 2.1 design
2188 - Added chroot hypervisor
2189 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2191 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2192 ``bootloader_path`` and ``bootloader_args``)
2193 - Added the ``use_localtime`` option for the xen-hvm and kvm
2194 hypervisors, and the default value for this has changed to false (in
2195 2.0 xen-hvm always enabled it)
2196 - Added luxi call to submit multiple jobs in one go
2197 - Added cluster initialization option to not modify ``/etc/hosts``
2199 - Added network interface parameters
2200 - Added dry run mode to some LUs
2201 - Added RAPI resources:
2203 - ``/2/instances/[instance_name]/info``
2204 - ``/2/instances/[instance_name]/replace-disks``
2205 - ``/2/nodes/[node_name]/evacuate``
2206 - ``/2/nodes/[node_name]/migrate``
2207 - ``/2/nodes/[node_name]/role``
2208 - ``/2/nodes/[node_name]/storage``
2209 - ``/2/nodes/[node_name]/storage/modify``
2210 - ``/2/nodes/[node_name]/storage/repair``
2212 - Added OpCodes to evacuate or migrate all instances on a node
2213 - Added new command to list storage elements on nodes (``gnt-node
2214 list-storage``) and modify them (``gnt-node modify-storage``)
2215 - Added new ssconf files with master candidate IP address
2216 (``ssconf_master_candidates_ips``), node primary IP address
2217 (``ssconf_node_primary_ips``) and node secondary IP address
2218 (``ssconf_node_secondary_ips``)
2219 - Added ``ganeti-confd`` and a client library to query the Ganeti
2220 configuration via UDP
2221 - Added ability to run hooks after cluster initialization and before
2223 - Added automatic mode for disk replace (``gnt-instance replace-disks
2225 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2226 after catastrophic data-loss
2227 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2229 - Added ``gnt-instance move`` command to move instances
2230 - Added ``gnt-cluster watcher`` command to control watcher
2231 - Added ``gnt-node powercycle`` command to powercycle nodes
2232 - Added new job status field ``lock_status``
2233 - Added parseable error codes to cluster verification (``gnt-cluster
2234 verify --error-codes``) and made output less verbose (use
2235 ``--verbose`` to restore previous behaviour)
2236 - Added UUIDs to the main config entities (cluster, nodes, instances)
2237 - Added support for OS variants
2238 - Added support for hashed passwords in the Ganeti remote API users file
2240 - Added option to specify maximum timeout on instance shutdown
2241 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2242 - Added new helper script to start and stop Ganeti daemons
2243 (``daemon-util``), with the intent to reduce the work necessary to
2244 adjust Ganeti for non-Debian distributions and to start/stop daemons
2246 - Added more unittests
2247 - Fixed critical bug in ganeti-masterd startup
2248 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2249 now customisable at the cluster level for both the KVM and Xen
2250 hypervisors using the new ``migration_port`` parameter
2251 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2252 reinstalling an instance
2253 - Allowed ``@`` in tag names
2254 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2255 - Many documentation updates
2256 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2257 - ``gnt-instance reinstall`` can now reinstall multiple instances
2258 - Updated many command line parameters
2259 - Introduced new OS API version 15
2260 - No longer support a default hypervisor
2261 - Treat virtual LVs as inexistent
2262 - Improved job locking logic to reduce lock contention
2263 - Match instance and node names case insensitively
2264 - Reimplemented bash completion script to be more complete
2271 *(Released Thu, 4 Feb 2010)*
2273 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2274 - Fix a string formatting bug
2275 - Improve safety of the code in some error paths
2276 - Improve data validation in the master of values returned from nodes
2282 *(Released Thu, 17 Dec 2009)*
2284 - Fix security issue due to missing validation of iallocator names; this
2285 allows local and remote execution of arbitrary executables
2286 - Fix failure of gnt-node list during instance removal
2287 - Ship the RAPI documentation in the archive
2293 *(Released Wed, 30 Sep 2009)*
2295 - Fixed many wrong messages
2296 - Fixed a few bugs related to the locking library
2297 - Fixed MAC checking at instance creation time
2298 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2299 - Fixed a few issues related to signal handling in both daemons and
2301 - Fixed the example startup script provided
2302 - Fixed insserv dependencies in the example startup script (patch from
2304 - Fixed handling of drained nodes in the iallocator framework
2305 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2307 - Fixed error related to invalid job IDs in job polling
2308 - Fixed job/opcode persistence on unclean master shutdown
2309 - Fixed handling of partial job processing after unclean master
2311 - Fixed error reporting from LUs, previously all errors were converted
2312 into execution errors
2313 - Fixed error reporting from burnin
2314 - Decreased significantly the memory usage of the job queue
2315 - Optimised slightly multi-job submission
2316 - Optimised slightly opcode loading
2317 - Backported the multi-job submit framework from the development
2318 branch; multi-instance start and stop should be faster
2319 - Added script to clean archived jobs after 21 days; this will reduce
2320 the size of the queue directory
2321 - Added some extra checks in disk size tracking
2322 - Added an example ethers hook script
2323 - Added a cluster parameter that prevents Ganeti from modifying of
2325 - Added more node information to RAPI responses
2326 - Added a ``gnt-job watch`` command that allows following the ouput of a
2328 - Added a bind-address option to ganeti-rapi
2329 - Added more checks to the configuration verify
2330 - Enhanced the burnin script such that some operations can be retried
2332 - Converted instance reinstall to multi-instance model
2338 *(Released Fri, 7 Aug 2009)*
2340 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2341 to allow using the pre-2.0.2 behaviour in activation, if any existing
2342 instances have mismatched disk sizes in the configuration
2343 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2344 any configuration mismatches for disk sizes
2345 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2346 failover to work on two-node clusters
2347 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2349 - Fixed detection of OS script errors in ``gnt-backup export``
2350 - Fixed exit code of ``gnt-backup export``
2356 *(Released Fri, 17 Jul 2009)*
2358 - Added experimental support for stripped logical volumes; this should
2359 enhance performance but comes with a higher complexity in the block
2360 device handling; stripping is only enabled when passing
2361 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2362 affected even in the non-stripped mode
2363 - Improved resiliency against transient failures at the end of DRBD
2364 resyncs, and in general of DRBD resync checks
2365 - Fixed a couple of issues with exports and snapshot errors
2366 - Fixed a couple of issues in instance listing
2367 - Added display of the disk size in ``gnt-instance info``
2368 - Fixed checking for valid OSes in instance creation
2369 - Fixed handling of the "vcpus" parameter in instance listing and in
2370 general of invalid parameters
2371 - Fixed http server library, and thus RAPI, to handle invalid
2372 username/password combinations correctly; this means that now they
2373 report unauthorized for queries too, not only for modifications,
2374 allowing earlier detect of configuration problems
2375 - Added a new "role" node list field, equivalent to the master/master
2376 candidate/drained/offline flags combinations
2377 - Fixed cluster modify and changes of candidate pool size
2378 - Fixed cluster verify error messages for wrong files on regular nodes
2379 - Fixed a couple of issues with node demotion from master candidate role
2380 - Fixed node readd issues
2381 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2382 - Added a new ``--no-voting`` option for masterfailover to fix failover
2383 on two-nodes clusters when the former master node is unreachable
2384 - Added instance reinstall over RAPI
2390 *(Released Tue, 16 Jun 2009)*
2392 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2393 allow re-adding the start in single-user option (regression from 1.2)
2394 - the watcher writes the instance status to a file, to allow monitoring
2395 to report the instance status (from the master) based on cached
2396 results of the watcher's queries; while this can get stale if the
2397 watcher is being locked due to other work on the cluster, this is
2398 still an improvement
2399 - the watcher now also restarts the node daemon and the rapi daemon if
2401 - fixed the watcher to handle full and drained queue cases
2402 - hooks export more instance data in the environment, which helps if
2403 hook scripts need to take action based on the instance's properties
2404 (no longer need to query back into ganeti)
2405 - instance failovers when the instance is stopped do not check for free
2406 RAM, so that failing over a stopped instance is possible in low memory
2408 - rapi uses queries for tags instead of jobs (for less job traffic), and
2409 for cluster tags it won't talk to masterd at all but read them from
2411 - a couple of error handling fixes in RAPI
2412 - drbd handling: improved the error handling of inconsistent disks after
2413 resync to reduce the frequency of "there are some degraded disks for
2414 this instance" messages
2415 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2416 error handling path called a wrong function name)
2422 *(Released Wed, 27 May 2009)*
2424 - no changes from rc5
2430 *(Released Wed, 20 May 2009)*
2432 - fix a couple of bugs (validation, argument checks)
2433 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2434 - some small improvements to RAPI and IAllocator
2435 - make watcher automatically start the master daemon if down
2441 *(Released Mon, 27 Apr 2009)*
2443 - change the OS list to not require locks; this helps with big clusters
2444 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2445 volume group is broken
2446 - ``gnt-instance info``, without any arguments, doesn't run for all
2447 instances anymore; either pass ``--all`` or pass the desired
2448 instances; this helps against mistakes on big clusters where listing
2449 the information for all instances takes a long time
2450 - miscellaneous doc and man pages fixes
2456 *(Released Wed, 8 Apr 2009)*
2458 - Change the internal locking model of some ``gnt-node`` commands, in
2459 order to reduce contention (and blocking of master daemon) when
2460 batching many creation/reinstall jobs
2461 - Fixes to Xen soft reboot
2462 - No longer build documentation at build time, instead distribute it in
2463 the archive, in order to reduce the need for the whole docbook/rst
2470 *(Released Fri, 27 Mar 2009)*
2472 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2473 - Fix watcher startup sequence, improves the behaviour of busy clusters
2474 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2475 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2476 bind address and other places
2477 - Some documentation fixes and updates
2483 *(Released Mon, 2 Mar 2009)*
2485 - More documentation updates, now all docs should be more-or-less
2487 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2489 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2491 - Fix an issue related to $libdir/run/ganeti and cluster creation
2497 *(Released Thu, 19 Feb 2009)*
2499 - Xen PVM and KVM have switched the default value for the instance root
2500 disk to the first partition on the first drive, instead of the whole
2501 drive; this means that the OS installation scripts must be changed
2503 - Man pages have been updated
2504 - RAPI has been switched by default to HTTPS, and the exported functions
2505 should all work correctly
2506 - RAPI v1 has been removed
2507 - Many improvements to the KVM hypervisor
2508 - Block device errors are now better reported
2509 - Many other bugfixes and small improvements
2515 *(Released Mon, 26 Jan 2009)*
2517 - Version 2 is a general rewrite of the code and therefore the
2518 differences are too many to list, see the design document for 2.0 in
2519 the ``doc/`` subdirectory for more details
2520 - In this beta version there is not yet a migration path from 1.2 (there
2521 will be one in the final 2.0 release)
2522 - A few significant changes are:
2524 - all commands are executed by a daemon (``ganeti-masterd``) and the
2525 various ``gnt-*`` commands are just front-ends to it
2526 - all the commands are entered into, and executed from a job queue,
2527 see the ``gnt-job(8)`` manpage
2528 - the RAPI daemon supports read-write operations, secured by basic
2529 HTTP authentication on top of HTTPS
2530 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2531 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2532 to migrate to DRBD 8 first while still running Ganeti 1.2)
2533 - DRBD devices are using statically allocated minor numbers, which
2534 will be assigned to existing instances during the migration process
2535 - there is support for both Xen PVM and Xen HVM instances running on
2537 - KVM virtualization is supported too
2538 - file-based storage has been implemented, which means that it is
2539 possible to run the cluster without LVM and DRBD storage, for
2540 example using a shared filesystem exported from shared storage (and
2541 still have live migration)
2547 *(Released Tue, 13 Jan 2009)*
2549 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2550 - Reuse the old instance mac address by default on instance import, if
2551 the instance name is the same.
2552 - Handle situations in which the node info rpc returns incomplete
2554 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2555 - Improved version of batcher:
2557 - state file support
2558 - instance mac address support
2559 - support for HVM clusters/instances
2561 - Add an option to show the number of cpu sockets and nodes in
2563 - Support OSes that handle more than one version of the OS api (but do
2564 not change the current API in any other way)
2565 - Fix ``gnt-node migrate``
2566 - ``gnt-debug`` man page
2567 - Fixes various more typos and small issues
2568 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2574 *(Released Wed, 24 Sep 2008)*
2576 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2577 type of disk exported to fully virtualized instances.
2578 - provide access to the serial console of HVM instances
2579 - instance auto_balance flag, set by default. If turned off it will
2580 avoid warnings on cluster verify if there is not enough memory to fail
2581 over an instance. in the future it will prevent automatically failing
2582 it over when we will support that.
2583 - batcher tool for instance creation, see ``tools/README.batcher``
2584 - ``gnt-instance reinstall --select-os`` to interactively select a new
2585 operating system when reinstalling an instance.
2586 - when changing the memory amount on instance modify a check has been
2587 added that the instance will be able to start. also warnings are
2588 emitted if the instance will not be able to fail over, if auto_balance
2590 - documentation fixes
2591 - sync fields between ``gnt-instance list/modify/add/import``
2592 - fix a race condition in drbd when the sync speed was set after giving
2593 the device a remote peer.
2599 *(Released Tue, 22 Jul 2008)*
2601 - note: the allowed size and number of tags per object were reduced
2602 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2603 - fixed twisted 8.x compatibility
2604 - fixed ``gnt-instance replace-disks`` with iallocator
2605 - add TCP keepalives on twisted connections to detect restarted nodes
2606 - disk increase support, see ``gnt-instance grow-disk``
2607 - implement bulk node/instance query for RAPI
2608 - add tags in node/instance listing (optional)
2609 - experimental migration (and live migration) support, read the man page
2610 for ``gnt-instance migrate``
2611 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2612 has some small improvements in handling its state file
2618 *(Released Fri, 13 Jun 2008)*
2620 - Experimental readonly, REST-based remote API implementation;
2621 automatically started on master node, TCP port 5080, if enabled by
2622 ``--enable-rapi`` parameter to configure script.
2623 - Instance allocator support. Add and import instance accept a
2624 ``--iallocator`` parameter, and call that instance allocator to decide
2625 which node to use for the instance. The iallocator document describes
2626 what's expected from an allocator script.
2627 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2628 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2629 node is lost there is still enough memory to fail over the instances
2631 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2632 ``gnt-cluster verify``, to check for site-specific compliance. All the
2633 hooks will run, and their output, if any, will be displayed. Any
2634 failing hook will make the verification return an error value.
2635 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2636 primary and secondary interfaces
2637 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2638 that is still declared as part of the cluster and has failed.
2639 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2640 specifying output fields, that just adds the chosen fields to the
2642 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2643 export from the filesystem.
2644 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2645 have been added. Using them you can enable/disable acpi and pae
2646 support, and specify a path for a cd image to be exported to the
2647 instance. These parameters as the name suggest only work on HVM
2649 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2650 PAE support will be set to the previously hardcoded values, but the
2651 (previously hardcoded) path to the CDROM ISO image will be unset and
2652 if required, needs to be set manually with ``gnt-instance modify``
2654 - The address to which an instance's VNC console is bound is now
2655 selectable per-instance, rather than being cluster wide. Of course
2656 this only applies to instances controlled via VNC, so currently just
2657 applies to HVM clusters.
2663 *(Released Mon, 18 Feb 2008)*
2665 - more tweaks to the disk activation code (especially helpful for DRBD)
2666 - change the default ``gnt-instance list`` output format, now there is
2667 one combined status field (see the manpage for the exact values this
2669 - some more fixes for the mac export to hooks change
2670 - make Ganeti not break with DRBD 8.2.x (which changed the version
2671 format in ``/proc/drbd``) (issue 24)
2672 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2673 template, allowing migration from DRBD0.7+MD to DRBD8
2679 *(Released Wed, 30 Jan 2008)*
2681 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2683 - add command aliases infrastructure and a few aliases
2684 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2685 man pages and the ``--help`` option of ``gnt-node
2686 list``/``gnt-instance list``
2687 - fix ``gnt-backup list`` with down nodes (issue 21)
2688 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2689 - fix the dist archive and add a check for including svn/git files in
2691 - some developer-related changes: improve the burnin and the QA suite,
2692 add an upload script for testing during development
2698 *(Released Wed, 16 Jan 2008)*
2700 - experimental HVM support, read the install document, section
2701 "Initializing the cluster"
2702 - allow for the PVM hypervisor per-instance kernel and initrd paths
2703 - add a new command ``gnt-cluster verify-disks`` which uses a new
2704 algorithm to improve the reconnection of the DRBD pairs if the device
2705 on the secondary node has gone away
2706 - make logical volume code auto-activate LVs at disk activation time
2707 - slightly improve the speed of activating disks
2708 - allow specification of the MAC address at instance creation time, and
2709 changing it later via ``gnt-instance modify``
2710 - fix handling of external commands that generate lots of output on
2712 - update documentation with regard to minimum version of DRBD8 supported
2718 *(Released Tue, 4 Dec 2007)*
2720 - Log the ``xm create`` output to the node daemon log on failure (to
2721 help diagnosing the error)
2722 - In debug mode, log all external commands output if failed to the logs
2723 - Change parsing of lvm commands to ignore stderr
2729 *(Released Wed, 28 Nov 2007)*
2731 - Another round of updates to the DRBD 8 code to deal with more failures
2732 in the replace secondary node operation
2733 - Some more logging of failures in disk operations (lvm, drbd)
2734 - A few documentation updates
2741 *(Released Tue, 13 Nov 2007)*
2743 - Change configuration file format from Python's Pickle to JSON.
2744 Upgrading is possible using the cfgupgrade utility.
2745 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2746 faster replace disks and is more stable (DRBD 8 has many improvements
2747 compared to DRBD 0.7)
2748 - Added command line tags support (see man pages for ``gnt-instance``,
2749 ``gnt-node``, ``gnt-cluster``)
2750 - Added instance rename support
2751 - Added multi-instance startup/shutdown
2752 - Added cluster rename support
2753 - Added ``gnt-node evacuate`` to simplify some node operations
2754 - Added instance reboot operation that can speedup reboot as compared to
2756 - Soften the requirement that hostnames are in FQDN format
2757 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2759 - Removed dependency on debian's patched fping that uses the
2760 non-standard ``-S`` option
2761 - Now the OS definitions are searched for in multiple, configurable
2762 paths (easier for distros to package)
2763 - Some changes to the hooks infrastructure (especially the new
2764 post-configuration update hook)
2765 - Other small bugfixes
2767 .. vim: set textwidth=72 syntax=rst :