10 Incompatible/important changes
11 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
13 - hroller now also plans for capacity to move non-redundant instances off
14 any node to be rebooted; the old behavior of completely ignoring any
15 non-redundant instances can be restored by adding the --ignore-non-redundant
17 - The cluster option '--no-lvm-storage' was removed in favor of the new option
18 '--enabled-disk-templates'.
19 - On instance creation, disk templates no longer need to be specified
20 with '-t'. The default disk template will be taken from the list of
21 enabled disk templates.
22 - The monitoring daemon is now running as root, in order to be able to collect
23 information only available to root (such as the state of Xen instances).
24 - The ConfD client is now IPv6 compatible.
25 - File and shared file storage is no longer dis/enabled at configure time,
26 but using the option '--enabled-disk-templates' at cluster initialization and
28 - The default directories for file and shared file storage are not anymore
29 specified at configure time, but taken from the cluster's configuration.
30 They can be set at cluster initialization and modification with
31 '--file-storage-dir' and '--shared-file-storage-dir'.
32 - Cluster verification now includes stricter checks regarding the
33 default file and shared file storage directories. It now checks that
34 the directories are explicitely allowed in the 'file-storage-paths' file and
35 that the directories exist on all nodes.
40 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
41 the correct command syntax. It is possible to use different DRBD versions
42 on different nodes as long as they are compatible to each other. This
43 enables rolling upgrades of DRBD with no downtime. As permanent operation
44 of different DRBD versions within a node group is discouraged,
45 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
46 - New "inst-status-xen" data collector for the monitoring daemon, providing
47 information about the state of the xen instances on the nodes.
48 - New "lv" data collector for the monitoring daemon, collecting data about the
49 logical volumes on the nodes, and pairing them with the name of the instances
51 - New "diskstats" data collector, collecting the data from /proc/diskstats and
52 presenting them over the monitoring daemon interface.
53 - The ConfD client is now IPv6 compatible.
57 The following new dependencies have been added.
61 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
62 for the unit tests (and only used for testing).
68 *(Released Mon, 24 Jun 2013)*
70 Incompatible/important changes
71 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
73 - Instance policy can contain multiple instance specs, as described in
74 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
75 <design-partitioned>`. As a consequence, it's not possible to partially change
76 or override instance specs. Bounding specs (min and max) can be specified as a
77 whole using the new option ``--ipolicy-bounds-specs``, while standard
78 specs use the new option ``--ipolicy-std-specs``.
79 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
80 gnt-instance is a valid YAML object.
81 - hail now honors network restrictions when allocating nodes. This led to an
82 update of the IAllocator protocol. See the IAllocator documentation for
88 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
89 to always require authentication, ``--require-authentication``. It can
90 be specified in ``$sysconfdir/default/ganeti``.
91 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
92 be used to manage the disk templates to be used by instances in the cluster.
93 Initially, it will be set to a list that includes plain, drbd, if they were
94 enabled by specifying a volume group name, and file and sharedfile, if those
95 were enabled at configure time. Additionally, it will include all disk
96 templates that are currently used by instances. The order of disk templates
97 will be based on Ganeti's history of supporting them. In the future, the
98 first entry of the list will be used as a default disk template on instance
100 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
101 configuration back to the previous stable version.
102 - Disk templates in group ipolicy can be restored to the default value.
103 - Initial support for diskless instances and virtual clusters in QA.
104 - More QA and unit tests for instance policies.
105 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
106 describing why the opcode itself was executed.
107 - The monitoring daemon is now available. It allows users to query the cluster
108 for obtaining information about the status of the system. The daemon is only
109 responsible for providing the information over the network: the actual data
110 gathering is performed by data collectors (currently, only the DRBD status
111 collector is available).
112 - In order to help developers work on Ganeti, a new script
113 (``devel/build_chroot``) is provided, for building a chroot that contains all
114 the required development libraries and tools for compiling Ganeti on a Debian
116 - A new tool, ``harep``, for performing self-repair and recreation of instances
117 in Ganeti has been added.
118 - Split queries are enabled for tags, network, exports, cluster info, groups,
120 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
121 It prints the command line to set the current policies, to ease
126 The following new dependencies have been added:
129 - The ``curl`` library is not optional anymore for compiling the Haskell code.
130 - ``snap-server`` library (if monitoring is enabled).
133 - The minimum Python version needed to run Ganeti is now 2.6.
134 - ``yaml`` library (only for running the QA).
140 *(Released Thu, 25 Jul 2013)*
142 - Add logrotate functionality in daemon-util
143 - Add logrotate example file
144 - Add missing fields to network queries over rapi
145 - Fix network object timestamps
146 - Add support for querying network timestamps
147 - Fix a typo in the example crontab
148 - Fix a documentation typo
154 *(Released Thu, 04 Jul 2013)*
156 Incompatible/important changes
157 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
159 - Instance policies for disk size were documented to be on a per-disk
160 basis, but hail applied them to the sum of all disks. This has been
162 - ``hbal`` will now exit with status 0 if, during job execution over
163 LUXI, early exit has been requested and all jobs are successful;
164 before, exit status 1 was used, which cannot be differentiated from
166 - Compatibility with newer versions of rbd has been fixed
167 - ``gnt-instance batch-create`` has been changed to use the bulk create
168 opcode from Ganeti. This lead to incompatible changes in the format of
169 the JSON file. It's now not a custom dict anymore but a dict
170 compatible with the ``OpInstanceCreate`` opcode.
171 - Parent directories for file storage need to be listed in
172 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
173 write the file automatically based on old configuration values, but it
174 can not distribute it across all nodes and the file contents should be
175 verified. Use ``gnt-cluster copyfile
176 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
177 upgraded. The reason for requiring this list of paths now is that
178 before it would have been possible to inject new paths via RPC,
179 allowing files to be created in arbitrary locations. The RPC protocol
180 is protected using SSL/X.509 certificates, but as a design principle
181 Ganeti does not permit arbitrary paths to be passed.
182 - The parsing of the variants file for OSes (see
183 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
184 lines and comment lines (starting with ``#``) are ignored for better
186 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
187 longer available. ``gnt-node add`` now invokes a new tool on the
188 destination node, named ``prepare-node-join``, to configure the SSH
189 daemon. Paramiko is no longer necessary to configure nodes' SSH
190 daemons via ``gnt-node add``.
191 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
192 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
193 the flag is not reset after a master failover.
194 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
195 is recommended. 2.6 will be mandatory from the 2.8 series.
201 - New network management functionality to support automatic allocation
202 of IP addresses and managing of network parameters. See
203 :manpage:`gnt-network(8)` for more details.
204 - New external storage backend, to allow managing arbitrary storage
205 systems external to the cluster. See
206 :manpage:`ganeti-extstorage-interface(7)`.
207 - New ``exclusive-storage`` node parameter added, restricted to
208 nodegroup level. When it's set to true, physical disks are assigned in
209 an exclusive fashion to instances, as documented in :doc:`Partitioned
210 Ganeti <design-partitioned>`. Currently, only instances using the
211 ``plain`` disk template are supported.
212 - The KVM hypervisor has been updated with many new hypervisor
213 parameters, including a generic one for passing arbitrary command line
214 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
215 compatible up to qemu 1.4.
216 - A new tool, called ``mon-collector``, is the stand-alone executor of
217 the data collectors for a monitoring system. As of this version, it
218 just includes the DRBD data collector, that can be executed by calling
219 ``mon-collector`` using the ``drbd`` parameter. See
220 :manpage:`mon-collector(7)`.
221 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
222 for RAPI users. It allows granting permissions to query for
223 information to a specific user without giving
224 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
225 - A new tool named ``node-cleanup`` has been added. It cleans remains of
226 a cluster from a machine by stopping all daemons, removing
227 certificates and ssconf files. Unless the ``--no-backup`` option is
228 given, copies of the certificates are made.
229 - Instance creations now support the use of opportunistic locking,
230 potentially speeding up the (parallel) creation of multiple instances.
231 This feature is currently only available via the :doc:`RAPI
232 <rapi>` interface and when an instance allocator is used. If the
233 ``opportunistic_locking`` parameter is set the opcode will try to
234 acquire as many locks as possible, but will not wait for any locks
235 held by other opcodes. If not enough resources can be found to
236 allocate the instance, the temporary error code
237 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
238 retried thereafter, with or without opportunistic locking.
239 - New experimental linux-ha resource scripts.
240 - Restricted-commands support: ganeti can now be asked (via command line
241 or rapi) to perform commands on a node. These are passed via ganeti
242 RPC rather than ssh. This functionality is restricted to commands
243 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
244 reasons. The file is not copied automatically.
250 - Diskless instances are now externally mirrored (Issue 237). This for
251 now has only been tested in conjunction with explicit target nodes for
253 - Queries not needing locks or RPC access to the node can now be
254 performed by the confd daemon, making them independent from jobs, and
255 thus faster to execute. This is selectable at configure time.
256 - The functionality for allocating multiple instances at once has been
257 overhauled and is now also available through :doc:`RAPI <rapi>`.
259 There are no significant changes from version 2.7.0~rc3.
265 *(Released Tue, 25 Jun 2013)*
267 - Fix permissions on the confd query socket (Issue 477)
268 - Fix permissions on the job archive dir (Issue 498)
269 - Fix handling of an internal exception in replace-disks (Issue 472)
270 - Fix gnt-node info handling of shortened names (Issue 497)
271 - Fix gnt-instance grow-disk when wiping is enabled
272 - Documentation improvements, and support for newer pandoc
273 - Fix hspace honoring ipolicy for disks (Issue 484)
274 - Improve handling of the ``kvm_extra`` HV parameter
280 *(Released Fri, 24 May 2013)*
282 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
284 - Disks added through ``gnt-instance modify`` or created through
285 ``gnt-instance recreate-disks`` are wiped, if the
286 ``prealloc_wipe_disks`` flag is set.
287 - If wiping newly created disks fails, the disks are removed. Also,
288 partial failures in creating disks through ``gnt-instance modify``
289 triggers a cleanup of the partially-created disks.
290 - Removing the master IP address doesn't fail if the address has been
292 - Fix ownership of the OS log dir
293 - Workaround missing SO_PEERCRED constant (Issue 191)
299 *(Released Fri, 3 May 2013)*
301 This was the first release candidate of the 2.7 series. Since beta3:
303 - Fix kvm compatibility with qemu 1.4 (Issue 389)
304 - Documentation updates (admin guide, upgrade notes, install
305 instructions) (Issue 372)
306 - Fix gnt-group list nodes and instances count (Issue 436)
307 - Fix compilation without non-mandatory libraries (Issue 441)
308 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
309 - Make confd logging more verbose at INFO level (Issue 435)
310 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
311 - Fix failure path for instance storage type conversion (Issue 229)
312 - Update htools text backend documentation
313 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
314 - Disable inter-cluster instance move for file-based instances, because
315 it is dependant on instance export, which is not supported for
316 file-based instances. (Issue 414)
317 - Fix gnt-job crashes on non-ascii characters (Issue 427)
318 - Fix volume group checks on non-vm-capable nodes (Issue 432)
324 *(Released Mon, 22 Apr 2013)*
326 This was the third beta release of the 2.7 series. Since beta2:
328 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
329 - Fix data loss on wrong usage of ``gnt-instance move``
330 - Properly export errors in confd-based job queries
331 - Add ``users-setup`` tool
332 - Fix iallocator protocol to report 0 as a disk size for diskless
333 instances. This avoids hail breaking when a diskless instance is
335 - Fix job queue directory permission problem that made confd job queries
336 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
337 for access to archived jobs (Issue 406).
338 - Limit the sizes of networks supported by ``gnt-network`` to something
339 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
340 - Fix bugs in instance disk template conversion
341 - Fix GHC 7 compatibility
342 - Fix ``burnin`` install path (Issue 426).
343 - Allow very small disk grows (Issue 347).
344 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
345 that noded doesn't import masterd code (Issue 419).
346 - Make sure the default metavg at cluster init is the same as the vg, if
347 unspecified (Issue 358).
348 - Fix cleanup of partially created disks (part of Issue 416)
354 *(Released Tue, 2 Apr 2013)*
356 This was the second beta release of the 2.7 series. Since beta1:
358 - Networks no longer have a "type" slot, since this information was
359 unused in Ganeti: instead of it tags should be used.
360 - The rapi client now has a ``target_node`` option to MigrateInstance.
361 - Fix early exit return code for hbal (Issue 386).
362 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
363 - Fix ``rbd showmapped`` output parsing (Issue 312).
364 - Networks are now referenced indexed by UUID, rather than name. This
365 will require running cfgupgrade, from 2.7.0beta1, if networks are in
367 - The OS environment now includes network information.
368 - Deleting of a network is now disallowed if any instance nic is using
369 it, to prevent dangling references.
370 - External storage is now documented in man pages.
371 - The exclusive_storage flag can now only be set at nodegroup level.
372 - Hbal can now submit an explicit priority with its jobs.
373 - Many network related locking fixes.
374 - Bump up the required pylint version to 0.25.1.
375 - Fix the ``no_remember`` option in RAPI client.
376 - Many ipolicy related tests, qa, and fixes.
377 - Many documentation improvements and fixes.
378 - Fix building with ``--disable-file-storage``.
379 - Fix ``-q`` option in htools, which was broken if passed more than
381 - Some haskell/python interaction improvements and fixes.
382 - Fix iallocator in case of missing LVM storage.
383 - Fix confd config load in case of ``--no-lvm-storage``.
384 - The confd/query functionality is now mentioned in the security
391 *(Released Wed, 6 Feb 2013)*
393 This was the first beta release of the 2.7 series. All important changes
394 are listed in the latest 2.7 entry.
400 *(Released Fri, 21 Dec 2012)*
402 Important behaviour change: hbal won't rebalance anymore instances which
403 have the ``auto_balance`` attribute set to false. This was the intention
404 all along, but until now it only skipped those from the N+1 memory
405 reservation (DRBD-specific).
407 A significant number of bug fixes in this release:
409 - Fixed disk adoption interaction with ipolicy checks.
410 - Fixed networking issues when instances are started, stopped or
411 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
412 - Fixed the warning in cluster verify for shared storage instances not
414 - Fixed removal of storage directory on shared file storage (issue 262).
415 - Fixed validation of LVM volume group name in OpClusterSetParams
416 (``gnt-cluster modify``) (issue 285).
417 - Fixed runtime memory increases (``gnt-instance modify -m``).
418 - Fixed live migration under Xen's ``xl`` mode.
419 - Fixed ``gnt-instance console`` with ``xl``.
420 - Fixed building with newer Haskell compiler/libraries.
421 - Fixed PID file writing in Haskell daemons (confd); this prevents
422 restart issues if confd was launched manually (outside of
423 ``daemon-util``) while another copy of it was running
424 - Fixed a type error when doing live migrations with KVM (issue 297) and
425 the error messages for failing migrations have been improved.
426 - Fixed opcode validation for the out-of-band commands (``gnt-node
428 - Fixed a type error when unsetting OS hypervisor parameters (issue
429 311); now it's possible to unset all OS-specific hypervisor
431 - Fixed the ``dry-run`` mode for many operations: verification of
432 results was over-zealous but didn't take into account the ``dry-run``
433 operation, resulting in "wrong" failures.
434 - Fixed bash completion in ``gnt-job list`` when the job queue has
435 hundreds of entries; especially with older ``bash`` versions, this
436 results in significant CPU usage.
438 And lastly, a few other improvements have been made:
440 - Added option to force master-failover without voting (issue 282).
441 - Clarified error message on lock conflict (issue 287).
442 - Logging of newly submitted jobs has been improved (issue 290).
443 - Hostname checks have been made uniform between instance rename and
445 - The ``--submit`` option is now supported by ``gnt-debug delay``.
446 - Shutting down the master daemon by sending SIGTERM now stops it from
447 processing jobs waiting for locks; instead, those jobs will be started
448 once again after the master daemon is started the next time (issue
450 - Support for Xen's ``xl`` program has been improved (besides the fixes
452 - Reduced logging noise in the Haskell confd daemon (only show one log
453 entry for each config reload, instead of two).
454 - Several man page updates and typo fixes.
460 *(Released Fri, 12 Oct 2012)*
462 A small bugfix release. Among the bugs fixed:
464 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
465 made the command unusable.
466 - Commands that issue many jobs don't fail anymore just because some jobs
467 take so long that other jobs are archived.
468 - Failures during ``gnt-instance reinstall`` are reflected by the exit
470 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
472 - When ``always_failover`` is set, ``--allow-failover`` is not required
473 in migrate commands anymore.
474 - ``bash_completion`` works even if extglob is disabled.
475 - Fixed bug with locks that made failover for RDB-based instances fail.
476 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
477 a random node instead of one based on the allocator metric.
478 - Support for newer versions of pylint and pep8.
479 - Hail doesn't fail anymore when trying to add an instance of type
480 ``file``, ``sharedfile`` or ``rbd``.
481 - Added new Makefile target to rebuild the whole distribution, so that
482 all files are included.
488 *(Released Fri, 27 Jul 2012)*
491 .. attention:: The ``LUXI`` protocol has been made more consistent
492 regarding its handling of command arguments. This, however, leads to
493 incompatibility issues with previous versions. Please ensure that you
494 restart Ganeti daemons soon after the upgrade, otherwise most
495 ``LUXI`` calls (job submission, setting/resetting the drain flag,
496 pausing/resuming the watcher, cancelling and archiving jobs, querying
497 the cluster configuration) will fail.
506 The current ``admin_up`` field, which used to denote whether an instance
507 should be running or not, has been removed. Instead, ``admin_state`` is
508 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
510 The rational behind this is that an instance being “down” can have
513 - it could be down during a reboot
514 - it could be temporarily be down for a reinstall
515 - or it could be down because it is deprecated and kept just for its
518 The previous Boolean state was making it difficult to do capacity
519 calculations: should Ganeti reserve memory for a down instance? Now, the
520 tri-state field makes it clear:
522 - in ``up`` and ``down`` state, all resources are reserved for the
523 instance, and it can be at any time brought up if it is down
524 - in ``offline`` state, only disk space is reserved for it, but not
527 The field can have an extra use: since the transition between ``up`` and
528 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
529 transition between ``offline`` and ``down`` is done via ``gnt-instance
530 modify``, it is possible to given different rights to users. For
531 example, owners of an instance could be allowed to start/stop it, but
532 not transition it out of the offline state.
534 Instance policies and specs
535 +++++++++++++++++++++++++++
537 In previous Ganeti versions, an instance creation request was not
538 limited on the minimum size and on the maximum size just by the cluster
539 resources. As such, any policy could be implemented only in third-party
540 clients (RAPI clients, or shell wrappers over ``gnt-*``
541 tools). Furthermore, calculating cluster capacity via ``hspace`` again
542 required external input with regards to instance sizes.
544 In order to improve these workflows and to allow for example better
545 per-node group differentiation, we introduced instance specs, which
548 - minimum instance disk size, disk count, memory size, cpu count
549 - maximum values for the above metrics
550 - and “standard” values (used in ``hspace`` to calculate the standard
553 The minimum/maximum values can be also customised at node-group level,
554 for example allowing more powerful hardware to support bigger instance
557 Beside the instance specs, there are a few other settings belonging to
558 the instance policy framework. It is possible now to customise, per
559 cluster and node-group:
561 - the list of allowed disk templates
562 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
563 - the maximum ratio of instance to spindles (see below for more
564 information) for local storage
566 All these together should allow all tools that talk to Ganeti to know
567 what are the ranges of allowed values for instances and the
568 over-subscription that is allowed.
570 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
571 instance configuration, and the physical CPU configuration from the
572 node. For the spindle ratios however, we didn't track before these
573 values, so new parameters have been added:
575 - a new node parameter ``spindle_count``, defaults to 1, customisable at
576 node group or node level
577 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
579 Note that spindles in this context doesn't need to mean actual
580 mechanical hard-drives; it's just a relative number for both the node
581 I/O capacity and instance I/O consumption.
583 Instance migration behaviour
584 ++++++++++++++++++++++++++++
586 While live-migration is in general desirable over failover, it is
587 possible that for some workloads it is actually worse, due to the
588 variable time of the “suspend” phase during live migration.
590 To allow the tools to work consistently over such instances (without
591 having to hard-code instance names), a new backend parameter
592 ``always_failover`` has been added to control the migration/failover
593 behaviour. When set to True, all migration requests for an instance will
594 instead fall-back to failover.
596 Instance memory ballooning
597 ++++++++++++++++++++++++++
599 Initial support for memory ballooning has been added. The memory for an
600 instance is no longer fixed (backend parameter ``memory``), but instead
601 can vary between minimum and maximum values (backend parameters
602 ``minmem`` and ``maxmem``). Currently we only change an instance's
605 - live migrating or failing over and instance and the target node
606 doesn't have enough memory
607 - user requests changing the memory via ``gnt-instance modify
613 In order to control the use of specific CPUs by instance, support for
614 controlling CPU pinning has been added for the Xen, HVM and LXC
615 hypervisors. This is controlled by a new hypervisor parameter
616 ``cpu_mask``; details about possible values for this are in the
617 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
618 VCPU-to-CPU mapping) form will work well only when all nodes in your
619 cluster have the same amount of CPUs.
624 Another area in which Ganeti was not customisable were the parameters
625 used for storage configuration, e.g. how many stripes to use for LVM,
626 DRBD resync configuration, etc.
628 To improve this area, we've added disks parameters, which are
629 customisable at cluster and node group level, and which allow to
630 specify various parameters for disks (DRBD has the most parameters
631 currently), for example:
633 - DRBD resync algorithm and parameters (e.g. speed)
634 - the default VG for meta-data volumes for DRBD
635 - number of stripes for LVM (plain disk template)
638 These parameters can be modified via ``gnt-cluster modify -D …`` and
639 ``gnt-group modify -D …``, and are used at either instance creation (in
640 case of LVM stripes, for example) or at disk “activation” time
643 Rados block device support
644 ++++++++++++++++++++++++++
646 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
647 denoted by the ``rbd`` disk template type. This is considered
648 experimental, feedback is welcome. For details on configuring it, see
649 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
654 The existing master IP functionality works well only in simple setups (a
655 single network shared by all nodes); however, if nodes belong to
656 different networks, then the ``/32`` setup and lack of routing
657 information is not enough.
659 To allow the master IP to function well in more complex cases, the
660 system was reworked as follows:
662 - a master IP netmask setting has been added
663 - the master IP activation/turn-down code was moved from the node daemon
665 - whether to run the Ganeti-supplied master IP script or a user-supplied
666 on is a ``gnt-cluster init`` setting
668 Details about the location of the standard and custom setup scripts are
669 in the man page :manpage:`gnt-cluster(8)`; for information about the
670 setup script protocol, look at the Ganeti-supplied script.
675 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
678 It is now possible to use TLS-protected connections, and when renewing
679 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
680 it is now possible to specify spice or spice CA certificates. Also, it
681 is possible to configure a password for SPICE sessions via the
682 hypervisor parameter ``spice_password_file``.
684 There are also new parameters to control the compression and streaming
685 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
686 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
687 for the spice parameters.
689 Lastly, it is now possible to see the SPICE connection information via
690 ``gnt-instance console``.
695 A new tool (``tools/ovfconverter``) has been added that supports
696 conversion between Ganeti and the `Open Virtualization Format
697 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
700 This relies on the ``qemu-img`` tool to convert the disk formats, so the
701 actual compatibility with other virtualization solutions depends on it.
706 The configuration query daemon (``ganeti-confd``) is now optional, and
707 has been rewritten in Haskell; whether to use the daemon at all, use the
708 Python (default) or the Haskell version is selectable at configure time
709 via the ``--enable-confd`` parameter, which can take one of the
710 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
711 daemon will result in a smaller footprint; for larger systems, we
712 welcome feedback on the Haskell version which might become the default
715 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
716 daemon running. The Python version doesn't implement the new call.
719 User interface changes
720 ~~~~~~~~~~~~~~~~~~~~~~
722 We have replaced the ``--disks`` option of ``gnt-instance
723 replace-disks`` with a more flexible ``--disk`` option, which allows
724 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
725 disk size and mode can be changed upon recreation (via ``gnt-instance
726 recreate-disks``, which accepts the same ``--disk`` option).
728 As many people are used to a ``show`` command, we have added that as an
729 alias to ``info`` on all ``gnt-*`` commands.
731 The ``gnt-instance grow-disk`` command has a new mode in which it can
732 accept the target size of the disk, instead of the delta; this can be
733 more safe since two runs in absolute mode will be idempotent, and
734 sometimes it's also easier to specify the desired size directly.
736 Also the handling of instances with regard to offline secondaries has
737 been improved. Instance operations should not fail because one of it's
738 secondary nodes is offline, even though it's safe to proceed.
740 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
741 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
742 minors to instance name.
747 RAPI coverage has improved, with (for example) new resources for
748 recreate-disks, node power-cycle, etc.
753 There is partial support for ``xl`` in the Xen hypervisor; feedback is
756 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
757 whether to still support Python 2.4 or move to Python 2.6 as minimum
760 Support for Fedora has been slightly improved; the provided example
761 init.d script should work better on it and the INSTALL file should
762 document the needed dependencies.
767 The deprecated ``QueryLocks`` LUXI request has been removed. Use
768 ``Query(what=QR_LOCK, ...)`` instead.
770 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
771 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
772 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
773 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
774 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
776 RAPI client: ``CertificateError`` now derives from
777 ``GanetiApiError``. This should make it more easy to handle Ganeti
780 Deprecation warnings due to PyCrypto/paramiko import in
781 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
782 make sure to run an up-to-date paramiko version, if you use this tool.
784 The QA scripts now depend on Python 2.5 or above (the main code base
785 still works with Python 2.4).
787 The configuration file (``config.data``) is now written without
788 indentation for performance reasons; if you want to edit it, it can be
789 re-formatted via ``tools/fmtjson``.
791 A number of bugs has been fixed in the cluster merge tool.
793 ``x509`` certification verification (used in import-export) has been
794 changed to allow the same clock skew as permitted by the cluster
795 verification. This will remove some rare but hard to diagnose errors in
802 *(Released Thu, 19 Jul 2012)*
804 Very few changes from rc4 to the final release, only bugfixes:
806 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
807 instance, fix CDROM booting for KVM instances)
808 - fixed node group modification of node parameters
809 - fixed issue in LUClusterVerifyGroup with multi-group clusters
810 - fixed generation of bash completion to ensure a stable ordering
817 *(Released Fri, 13 Jul 2012)*
819 Third release candidate for 2.6. The following changes were done from
822 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
823 - Fixed an inconsistency in the LUXI protocol with the provided
824 arguments (NOT backwards compatible)
825 - Fixed a bug with node groups ipolicy where ``min`` was greater than
826 the cluster ``std`` value
827 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
828 easier instance debugging on nodes (requires ``hconfd`` to work)
834 *(Released Tue, 03 Jul 2012)*
836 Second release candidate for 2.6. The following changes were done from
839 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
841 - Fixed a RAPI regression on missing beparams/memory
842 - Fixed redistribution of files on offline nodes
843 - Added possibility to run activate-disks even though secondaries are
844 offline. With this change it relaxes also the strictness on some other
845 commands which use activate disks internally:
846 * ``gnt-instance start|reboot|rename|backup|export``
847 - Made it possible to remove safely an instance if its secondaries are
849 - Made it possible to reinstall even though secondaries are offline
855 *(Released Mon, 25 Jun 2012)*
857 First release candidate for 2.6. The following changes were done from
860 - Fixed bugs with disk parameters and ``rbd`` templates as well as
862 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
863 behaviour. It supports now the modify operation
864 - ``hcheck`` implemented to analyze cluster health and possibility of
865 improving health by rebalance
866 - ``hbal`` has been improved in dealing with split instances
872 *(Released Mon, 11 Jun 2012)*
874 Second beta release of 2.6. The following changes were done from beta2
877 - Fixed ``daemon-util`` with non-root user models
878 - Fixed creation of plain instances with ``--no-wait-for-sync``
879 - Fix wrong iv_names when running ``cfgupgrade``
880 - Export more information in RAPI group queries
881 - Fixed bug when changing instance network interfaces
882 - Extended burnin to do NIC changes
883 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
884 - Changed default for DRBD barriers
885 - Fixed DRBD error reporting for syncer rate
886 - Verify the options on disk parameters
888 And of course various fixes to documentation and improved unittests and
895 *(Released Wed, 23 May 2012)*
897 First beta release of 2.6. The following changes were done from beta1 to
900 - integrated patch for distributions without ``start-stop-daemon``
901 - adapted example init.d script to work on Fedora
902 - fixed log handling in Haskell daemons
903 - adapted checks in the watcher for pycurl linked against libnss
904 - add partial support for ``xl`` instead of ``xm`` for Xen
905 - fixed a type issue in cluster verification
906 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
909 Plus integrated fixes from the 2.5 branch:
911 - fixed ``kvm-ifup`` to use ``/bin/bash``
912 - fixed parallel build failures
913 - KVM live migration when using a custom keymap
919 *(Released Tue, 24 Jul 2012)*
921 A small bugfix release, with no new features:
923 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
924 different default shell (e.g. Debian, Ubuntu)
925 - fixed KVM startup and live migration with a custom keymap (fixes Issue
926 243 and Debian bug #650664)
927 - fixed compatibility with KVM versions that don't support multiple boot
928 devices (fixes Issue 230 and Debian bug #624256)
930 Additionally, a few fixes were done to the build system (fixed parallel
931 build failures) and to the unittests (fixed race condition in test for
932 FileID functions, and the default enable/disable mode for QA test is now
939 *(Released Fri, 11 May 2012)*
941 A small bugfix release.
943 The main issues solved are on the topic of compatibility with newer LVM
946 - fixed parsing of ``lv_attr`` field
947 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
948 the ``--force`` flag is needed
950 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
951 to accept kernel 3.x too (was hardcoded to 2.6.*).
953 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
954 of hook results and that also made display of other errors suboptimal
955 was fixed; the code behaves now like 2.4 and earlier.
957 Another change in 2.5, the cleanup of the OS scripts environment, is too
958 aggressive: it removed even the ``PATH`` variable, which requires the OS
959 scripts to *always* need to export it. Since this is a bit too strict,
960 we now export a minimal PATH, the same that we export for hooks.
962 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
963 integrated into this release.
965 Finally, a few other miscellaneous changes were done (no new features,
966 just small improvements):
968 - Fix ``gnt-group --help`` display
969 - Fix hardcoded Xen kernel path
970 - Fix grow-disk handling of invalid units
971 - Update synopsis for ``gnt-cluster repair-disk-sizes``
972 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
978 *(Released Thu, 12 Apr 2012)*
980 Incompatible/important changes and bugfixes
981 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
983 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
984 resource's ``ip_check`` parameter changed from ``True`` to ``False``
985 to match the underlying LUXI interface.
986 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
987 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
988 not maintain backwards-compatibility as the underlying operation
989 changed in an incompatible way. The RAPI client can talk to old
990 servers, but it needs to be told so as the return value changed.
991 - When creating file-based instances via RAPI, the ``file_driver``
992 parameter no longer defaults to ``loop`` and must be specified.
993 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
995 - Support for the undocumented and deprecated RAPI instance creation
996 request format version 0 has been dropped. Use version 1, supported
997 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
998 - Pyparsing 1.4.6 or above is required, see :doc:`installation
999 documentation <install>`.
1000 - The "cluster-verify" hooks are now executed per group by the
1001 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1002 if you just run ``gnt-cluster verify``, which generates one opcode per
1004 - The environment as passed to the OS scripts is cleared, and thus no
1005 environment variables defined in the node daemon's environment will be
1006 inherited by the scripts.
1007 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1009 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1010 support operations involving multiple node groups.
1011 - Offline nodes are ignored when failing over an instance.
1012 - Support for KVM version 1.0, which changed the version reporting format
1014 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1016 - ``Makefile`` is now compatible with Automake 1.11.2
1017 - Includes all bugfixes made in the 2.4 series
1022 - The ganeti-htools project has been merged into the ganeti-core source
1023 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1024 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1025 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1026 GPT-style partition tables (requires `parted
1027 <http://www.gnu.org/s/parted/>`_).
1028 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1029 - Allowed adding tags on instance creation.
1030 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1031 - Allow instances to be started in a paused state, enabling the user to
1032 see the complete console output on boot using the console.
1033 - Added new hypervisor flag to control default reboot behaviour
1034 (``reboot_behavior``).
1035 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1036 - Improved out-of-band management support:
1038 - Added ``gnt-node health`` command reporting the health status of
1040 - Added ``gnt-node power`` command to manage power status of nodes.
1041 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1043 - Instance migration can fall back to failover if instance is not
1045 - Filters can be used when listing nodes, instances, groups and locks;
1046 see :manpage:`ganeti(7)` manpage.
1047 - Added post-execution status as variables to :doc:`hooks <hooks>`
1049 - Instance tags are exported/imported together with the instance.
1050 - When given an explicit job ID, ``gnt-job info`` will work for archived
1052 - Jobs can define dependencies on other jobs (not yet supported via
1053 RAPI or command line, but used by internal commands and usable via
1056 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1059 - Instance failover is now available as a RAPI resource
1060 (``/2/instances/[instance_name]/failover``).
1061 - ``gnt-instance info`` defaults to static information if primary node
1063 - Opcodes have a new ``comment`` attribute.
1064 - Added basic SPICE support to KVM hypervisor.
1065 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1067 Node group improvements
1068 ~~~~~~~~~~~~~~~~~~~~~~~
1070 - ``gnt-cluster verify`` has been modified to check groups separately,
1071 thereby improving performance.
1072 - Node group support has been added to ``gnt-cluster verify-disks``,
1073 which now operates per node group.
1074 - Watcher has been changed to work better with node groups.
1076 - One process and state file per node group.
1077 - Slow watcher in one group doesn't block other group's watcher.
1079 - Added new command, ``gnt-group evacuate``, to move all instances in a
1080 node group to other groups.
1081 - Added ``gnt-instance change-group`` to move an instance to another
1083 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1084 per-group operations.
1085 - Node groups can be tagged.
1086 - Some operations switch from an exclusive to a shared lock as soon as
1088 - Instance's primary and secondary nodes' groups are now available as
1089 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1090 and ``snodes.group.uuid``).
1095 - Numerous updates to documentation and manpages.
1097 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1099 - Some opcode/job results are now also documented, see :doc:`RAPI
1102 - A lockset's internal lock is now also visible in lock monitor.
1103 - Log messages from job queue workers now contain information about the
1104 opcode they're processing.
1105 - ``gnt-instance console`` no longer requires the instance lock.
1106 - A short delay when waiting for job changes reduces the number of LUXI
1107 requests significantly.
1108 - DRBD metadata volumes are overwritten with zeros during disk creation.
1109 - Out-of-band commands no longer acquire the cluster lock in exclusive
1111 - ``devel/upload`` now uses correct permissions for directories.
1117 *(Released Fri, 23 Mar 2012)*
1119 This was the sixth release candidate of the 2.5 series.
1125 *(Released Mon, 9 Jan 2012)*
1127 This was the fifth release candidate of the 2.5 series.
1133 *(Released Thu, 27 Oct 2011)*
1135 This was the fourth release candidate of the 2.5 series.
1141 *(Released Wed, 26 Oct 2011)*
1143 This was the third release candidate of the 2.5 series.
1149 *(Released Tue, 18 Oct 2011)*
1151 This was the second release candidate of the 2.5 series.
1157 *(Released Tue, 4 Oct 2011)*
1159 This was the first release candidate of the 2.5 series.
1165 *(Released Wed, 31 Aug 2011)*
1167 This was the third beta release of the 2.5 series.
1173 *(Released Mon, 22 Aug 2011)*
1175 This was the second beta release of the 2.5 series.
1181 *(Released Fri, 12 Aug 2011)*
1183 This was the first beta release of the 2.5 series.
1189 *(Released Thu, 27 Oct 2011)*
1191 - Fixed bug when parsing command line parameter values ending in
1193 - Fixed assertion error after unclean master shutdown
1194 - Disable HTTP client pool for RPC, significantly reducing memory usage
1196 - Fixed queue archive creation with wrong permissions
1202 *(Released Tue, 23 Aug 2011)*
1206 - Fixed documentation for importing with ``--src-dir`` option
1207 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1208 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1214 *(Released Fri, 5 Aug 2011)*
1216 Many bug-fixes and a few small features:
1218 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1219 issues when you tried to add an instance with two MAC addresses in one
1221 - KVM: fixed per-instance stored UID value
1222 - KVM: configure bridged NICs at migration start
1223 - KVM: Fix a bug where instance will not start with never KVM versions
1225 - Added OS search path to ``gnt-cluster info``
1226 - Fixed an issue with ``file_storage_dir`` where you were forced to
1227 provide an absolute path, but the documentation states it is a
1228 relative path, the documentation was right
1229 - Added a new parameter to instance stop/start called ``--no-remember``
1230 that will make the state change to not be remembered
1231 - Implemented ``no_remember`` at RAPI level
1232 - Improved the documentation
1233 - Node evacuation: don't call IAllocator if node is already empty
1234 - Fixed bug in DRBD8 replace disks on current nodes
1235 - Fixed bug in recreate-disks for DRBD instances
1236 - Moved assertion checking locks in ``gnt-instance replace-disks``
1237 causing it to abort with not owning the right locks for some situation
1238 - Job queue: Fixed potential race condition when cancelling queued jobs
1239 - Fixed off-by-one bug in job serial generation
1240 - ``gnt-node volumes``: Fix instance names
1241 - Fixed aliases in bash completion
1242 - Fixed a bug in reopening log files after being sent a SIGHUP
1243 - Added a flag to burnin to allow specifying VCPU count
1244 - Bugfixes to non-root Ganeti configuration
1250 *(Released Thu, 12 May 2011)*
1252 Many bug-fixes and a few new small features:
1254 - Fixed a bug related to log opening failures
1255 - Fixed a bug in instance listing with orphan instances
1256 - Fixed a bug which prevented resetting the cluster-level node parameter
1257 ``oob_program`` to the default
1258 - Many fixes related to the ``cluster-merge`` tool
1259 - Fixed a race condition in the lock monitor, which caused failures
1260 during (at least) creation of many instances in parallel
1261 - Improved output for gnt-job info
1262 - Removed the quiet flag on some ssh calls which prevented debugging
1264 - Improved the N+1 failure messages in cluster verify by actually
1265 showing the memory values (needed and available)
1266 - Increased lock attempt timeouts so that when executing long operations
1267 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1268 too early and thus prevent the use of the 'fair' mechanism
1269 - Changed instance query data (``gnt-instance info``) to not acquire
1270 locks unless needed, thus allowing its use on locked instance if only
1271 static information is asked for
1272 - Improved behaviour with filesystems that do not support rename on an
1274 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1275 kept locks on all nodes during the wipe, which is unneeded
1276 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1277 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1278 10GiB) which caused the wipe to fail right at the end in some cases
1279 - Fixed master IP activation when doing master failover with no-voting
1280 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1281 the master node itself
1282 - Fixed potential data-loss in under disk full conditions, where Ganeti
1283 wouldn't check correctly the return code and would consider
1284 partially-written files 'correct'
1285 - Fixed bug related to multiple VGs and DRBD disk replacing
1286 - Added new disk parameter ``metavg`` that allows placement of the meta
1287 device for DRBD in a different volume group
1288 - Fixed error handling in the node daemon when the system libc doesn't
1289 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1290 - Fixed lock release during replace-disks, which kept cluster-wide locks
1291 when doing disk replaces with an iallocator script
1292 - Added check for missing bridges in cluster verify
1293 - Handle EPIPE errors while writing to the terminal better, so that
1294 piping the output to e.g. ``less`` doesn't cause a backtrace
1295 - Fixed rare case where a ^C during Luxi calls could have been
1296 interpreted as server errors, instead of simply terminating
1297 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1299 - Added a few more parameters to the KVM hypervisor, allowing a second
1300 CDROM, custom disk type for CDROMs and a floppy image
1301 - Removed redundant message in instance rename when the name is given
1303 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1304 disks on new nodes, allowing recreation when the original instance
1305 nodes are completely gone
1306 - Added option when converting disk templates to DRBD to skip waiting
1307 for the resync, in order to make the instance available sooner
1308 - Added two new variables to the OS scripts environment (containing the
1310 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1311 to allow use of ``pvgrub`` as bootloader
1312 - Changed the instance memory modifications to only check out-of-memory
1313 conditions on memory increases, and turned the secondary node warnings
1314 into errors (they can still be overridden via ``--force``)
1315 - Fixed the handling of a corner case when the Python installation gets
1316 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1317 to execute a command that doesn't exist
1318 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1319 node of the instance returned failures during instance shutdown; this
1320 adds the option ``--ignore-consistency`` to gnt-instance move
1322 And as usual, various improvements to the error messages, documentation
1329 *(Released Wed, 09 Mar 2011)*
1331 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1332 the RAPI users file if run twice (even with ``--dry-run``).
1334 The release fixes that bug (nothing else changed).
1340 *(Released Mon, 07 Mar 2011)*
1342 Final 2.4.0 release. Just a few small fixes:
1344 - Fixed RAPI node evacuate
1345 - Fixed the kvm-ifup script
1346 - Fixed internal error handling for special job cases
1347 - Updated man page to specify the escaping feature for options
1353 *(Released Mon, 28 Feb 2011)*
1355 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1356 that this feature wiped the disks of the wrong instance, leading to loss
1361 - Fixed title of query field containing instance name
1362 - Expanded the glossary in the documentation
1363 - Fixed one unittest (internal issue)
1369 *(Released Mon, 21 Feb 2011)*
1371 A number of bug fixes plus just a couple functionality changes.
1373 On the user-visible side, the ``gnt-* list`` command output has changed
1374 with respect to "special" field states. The current rc1 style of display
1375 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1376 the default output mode special fields states are displayed as follows:
1378 - Offline resource: ``*``
1379 - Unavailable/not applicable: ``-``
1380 - Data missing (RPC failure): ``?``
1381 - Unknown field: ``??``
1383 Another user-visible change is the addition of ``--force-join`` to
1388 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1389 - Fixed regression in RAPI/instance reinstall where all parameters were
1390 required (instead of optional)
1391 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1392 - Fixed iallocator usage (offline nodes were not considered offline)
1393 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1394 - Fixed hypervisor and OS parameter validation with respect to
1395 non-vm_capable nodes
1396 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1398 - Fixed ``tools/listrunner`` with respect to agent-based usage
1404 *(Released Fri, 4 Feb 2011)*
1406 Many changes and fixes since the beta1 release. While there were some
1407 internal changes, the code has been mostly stabilised for the RC
1410 Note: the dumb allocator was removed in this release, as it was not kept
1411 up-to-date with the IAllocator protocol changes. It is recommended to
1412 use the ``hail`` command from the ganeti-htools package.
1414 Note: the 2.4 and up versions of Ganeti are not compatible with the
1415 0.2.x branch of ganeti-htools. You need to upgrade to
1416 ganeti-htools-0.3.0 (or later).
1418 Regressions fixed from 2.3
1419 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1421 - Fixed the ``gnt-cluster verify-disks`` command
1422 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1424 - Fixed disk adoption breakage
1425 - Fixed wrong headers in instance listing for field aliases
1430 - Fixed corner case in KVM handling of NICs
1431 - Fixed many cases of wrong handling of non-vm_capable nodes
1432 - Fixed a bug where a missing instance symlink was not possible to
1433 recreate with any ``gnt-*`` command (now ``gnt-instance
1434 activate-disks`` does it)
1435 - Fixed the volume group name as reported by ``gnt-cluster
1437 - Increased timeouts for the import-export code, hopefully leading to
1438 fewer aborts due network or instance timeouts
1439 - Fixed bug in ``gnt-node list-storage``
1440 - Fixed bug where not all daemons were started on cluster
1441 initialisation, but only at the first watcher run
1442 - Fixed many bugs in the OOB implementation
1443 - Fixed watcher behaviour in presence of instances with offline
1445 - Fixed instance list output for instances running on the wrong node
1446 - a few fixes to the cluster-merge tool, but it still cannot merge
1447 multi-node groups (currently it is not recommended to use this tool)
1453 - Improved network configuration for the KVM hypervisor
1454 - Added e1000 as a supported NIC for Xen-HVM
1455 - Improved the lvmstrap tool to also be able to use partitions, as
1456 opposed to full disks
1457 - Improved speed of disk wiping (the cluster parameter
1458 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1459 of instance creations
1460 - Added documentation for the OS parameters
1461 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1462 hypervisor is not responding
1463 - Added display of blacklisted and hidden OS information in
1464 ``gnt-cluster info``
1465 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1466 NIC and node parameters, which might create problems with currently
1467 invalid (but undetected) configuration files, but prevents validation
1468 failures when unrelated parameters are modified
1469 - Changed cluster initialisation to wait for the master daemon to become
1471 - Expanded the RAPI interface:
1473 - Added config redistribution resource
1474 - Added activation/deactivation of instance disks
1475 - Added export of console information
1477 - Implemented log file reopening on SIGHUP, which allows using
1478 logrotate(8) for the Ganeti log files
1479 - Added a basic OOB helper script as an example
1485 *(Released Fri, 14 Jan 2011)*
1490 - Fixed timezone issues when formatting timestamps
1491 - Added support for node groups, available via ``gnt-group`` and other
1493 - Added out-of-band framework and management, see :doc:`design
1494 document <design-oob>`
1495 - Removed support for roman numbers from ``gnt-node list`` and
1496 ``gnt-instance list``.
1497 - Allowed modification of master network interface via ``gnt-cluster
1498 modify --master-netdev``
1499 - Accept offline secondaries while shutting down instance disks
1500 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1501 - Added support for multiple LVM volume groups
1502 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1504 - Added commands to list available fields:
1506 - ``gnt-node list-fields``
1507 - ``gnt-group list-fields``
1508 - ``gnt-instance list-fields``
1510 - Updated documentation and man pages
1515 - Moved ``rapi_users`` file into separate directory, now named
1516 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1518 - Added new tool for running commands on many machines,
1519 ``tools/ganeti-listrunner``
1520 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1521 improving the ``gnt-instance console`` output
1522 - Allowed customisation of disk index separator at ``configure`` time
1523 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1524 - Added support for non-partitioned md disks in ``lvmstrap``
1525 - Added script to gracefully power off KVM instances
1526 - Split ``utils`` module into smaller parts
1527 - Changed query operations to return more detailed information, e.g.
1528 whether an information is unavailable due to an offline node. To use
1529 this new functionality, the LUXI call ``Query`` must be used. Field
1530 information is now stored by the master daemon and can be retrieved
1531 using ``QueryFields``. Instances, nodes and groups can also be queried
1532 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1533 exposed via RAPI). The following commands make use of this
1534 infrastructure change:
1536 - ``gnt-group list``
1537 - ``gnt-group list-fields``
1539 - ``gnt-node list-fields``
1540 - ``gnt-instance list``
1541 - ``gnt-instance list-fields``
1542 - ``gnt-debug locks``
1547 - New RAPI resources (see :doc:`rapi`):
1551 - ``/2/groups/[group_name]``
1552 - ``/2/groups/[group_name]/assign-nodes``
1553 - ``/2/groups/[group_name]/modify``
1554 - ``/2/groups/[group_name]/rename``
1555 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1559 - Implemented ``no_install`` for instance creation
1560 - Implemented OS parameters for instance reinstallation, allowing
1561 use of special settings on reinstallation (e.g. for preserving data)
1566 - Added IPv6 support in import/export
1567 - Pause DRBD synchronization while wiping disks on instance creation
1568 - Updated unittests and QA scripts
1569 - Improved network parameters passed to KVM
1570 - Converted man pages from docbook to reStructuredText
1576 *(Released Mon, 20 Dec 2010)*
1578 Released version 2.3.1~rc1 without any changes.
1584 *(Released Wed, 1 Dec 2010)*
1586 - impexpd: Disable OpenSSL compression in socat if possible (backport
1587 from master, commit e90739d625b, see :doc:`installation guide
1588 <install-quick>` for details)
1589 - Changed unittest coverage report to exclude test scripts
1590 - Added script to check version format
1596 *(Released Wed, 1 Dec 2010)*
1598 Released version 2.3.0~rc1 without any changes.
1604 *(Released Fri, 19 Nov 2010)*
1606 A number of bugfixes and documentation updates:
1608 - Update ganeti-os-interface documentation
1609 - Fixed a bug related to duplicate MACs or similar items which should be
1611 - Fix breakage in OS state modify
1612 - Reinstall instance: disallow offline secondaries (fixes bug related to
1613 OS changing but reinstall failing)
1614 - plus all the other fixes between 2.2.1 and 2.2.2
1620 *(Released Tue, 2 Nov 2010)*
1622 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1623 - Fixed master failover race with watcher
1624 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1626 - Accept previously stopped instance for export with instance removal
1627 - Simplify and extend the environment variables for instance OS scripts
1628 - Added new node flags, ``master_capable`` and ``vm_capable``
1629 - Added optional instance disk wiping prior during allocation. This is a
1630 cluster-wide option and can be set/modified using
1631 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1632 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1633 :doc:`install-quick`
1634 - Added a new watcher option (``--ignore-pause``)
1635 - Added option to ignore offline node on instance start/stop
1636 (``--ignore-offline``)
1637 - Allow overriding OS parameters with ``gnt-instance reinstall``
1638 - Added ability to change node's secondary IP address using ``gnt-node
1640 - Implemented privilege separation for all daemons except
1641 ``ganeti-noded``, see ``configure`` options
1642 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1644 - Implemented job priorities (see ``ganeti(7)`` manpage)
1645 - Ignore failures while shutting down instances during failover from
1647 - Exit daemon's bootstrap process only once daemon is ready
1648 - Export more information via ``LUInstanceQuery``/remote API
1649 - Improved documentation, QA and unittests
1650 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1651 restart if the file was created or changed
1652 - Added LUXI protocol version sent with each request and response,
1653 allowing detection of server/client mismatches
1654 - Moved the Python scripts among gnt-* and ganeti-* into modules
1655 - Moved all code related to setting up SSH to an external script,
1657 - Infrastructure changes for node group support in future versions
1663 *(Released Fri, 19 Nov 2010)*
1665 A few small bugs fixed, and some improvements to the build system:
1667 - Fix documentation regarding conversion to drbd
1668 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1670 - Fix error handling in node modify with multiple changes
1671 - Allow remote imports without checked names
1677 *(Released Tue, 19 Oct 2010)*
1679 - Disable SSL session ID cache in RPC client
1685 *(Released Thu, 14 Oct 2010)*
1687 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1688 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1690 - Fix problems with interaction between Curl and Python's HTTP server,
1691 resulting in increased speed in many RPC calls
1692 - Improve our release script to prevent breakage with older aclocal and
1699 *(Released Thu, 7 Oct 2010)*
1701 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1702 value retrieved from master
1703 - Added support for blacklisted or hidden OS definitions
1704 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1705 - Added support for -mem-path in KVM hypervisor abstraction layer
1706 - Allow overriding instance parameters in tool for inter-cluster
1707 instance moves (``tools/move-instance``)
1708 - Improved opcode summaries (e.g. in ``gnt-job list``)
1709 - Improve consistency of OS listing by sorting it
1710 - Documentation updates
1716 *(Released Fri, 8 Oct 2010)*
1718 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1724 *(Released Mon, 4 Oct 2010)*
1726 - Fixed regression in ``gnt-instance rename``
1732 *(Released Wed, 22 Sep 2010)*
1734 - Fixed OS_VARIANT variable for OS scripts
1735 - Fixed cluster tag operations via RAPI
1736 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1737 - Disabled RAPI CA checks in watcher
1743 *(Released Mon, 23 Aug 2010)*
1745 - Support DRBD versions of the format "a.b.c.d"
1747 - Re-introduce support for usage from multiple threads in RAPI client
1748 - Instance renames and modify via RAPI
1749 - Work around race condition between processing and archival in job
1751 - Mark opcodes following failed one as failed, too
1752 - Job field ``lock_status`` was removed due to difficulties making it
1753 work with the changed job queue in Ganeti 2.2; a better way to monitor
1754 locks is expected for a later 2.2.x release
1755 - Fixed dry-run behaviour with many commands
1756 - Support ``ssh-agent`` again when adding nodes
1757 - Many additional bugfixes
1763 *(Released Fri, 30 Jul 2010)*
1765 Important change: the internal RPC mechanism between Ganeti nodes has
1766 changed from using a home-grown http library (based on the Python base
1767 libraries) to use the PycURL library. This requires that PycURL is
1768 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1769 against GnuTLS by default. cURL's support for GnuTLS had known issues
1770 before cURL 7.21.0 and we recommend using the latest cURL release or
1771 linking against OpenSSL. Most other distributions already link PycURL
1772 and cURL against OpenSSL. The command::
1774 python -c 'import pycurl; print pycurl.version'
1776 can be used to determine the libraries PycURL and cURL are linked
1779 Other significant changes:
1781 - Rewrote much of the internals of the job queue, in order to achieve
1782 better parallelism; this decouples job query operations from the job
1783 processing, and it should allow much nicer behaviour of the master
1784 daemon under load, and it also has uncovered some long-standing bugs
1785 related to the job serialisation (now fixed)
1786 - Added a default iallocator setting to the cluster parameters,
1787 eliminating the need to always pass nodes or an iallocator for
1788 operations that require selection of new node(s)
1789 - Added experimental support for the LXC virtualization method
1790 - Added support for OS parameters, which allows the installation of
1791 instances to pass parameter to OS scripts in order to customise the
1793 - Added a hypervisor parameter controlling the migration type (live or
1794 non-live), since hypervisors have various levels of reliability; this
1795 has renamed the 'live' parameter to 'mode'
1796 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1797 logical volumes, meaning that cluster verify will ignore them and not
1798 flag their presence as errors
1799 - The watcher will now reset the error count for failed instances after
1800 8 hours, thus allowing self-healing if the problem that caused the
1801 instances to be down/fail to start has cleared in the meantime
1802 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1803 check for, and warn, if the drbd module parameter ``usermode_helper``
1804 is not consistent with the cluster-wide setting; this is needed to
1805 make diagnose easier of failed drbd creations
1806 - Started adding base IPv6 support, but this is not yet
1807 enabled/available for use
1808 - Rename operations (cluster, instance) will now return the new name,
1809 which is especially useful if a short name was passed in
1810 - Added support for instance migration in RAPI
1811 - Added a tool to pre-configure nodes for the SSH setup, before joining
1812 them to the cluster; this will allow in the future a simplified model
1813 for node joining (but not yet fully enabled in 2.2); this needs the
1814 paramiko python library
1815 - Fixed handling of name-resolving errors
1816 - Fixed consistency of job results on the error path
1817 - Fixed master-failover race condition when executed multiple times in
1819 - Fixed many bugs related to the job queue (mostly introduced during the
1820 2.2 development cycle, so not all are impacting 2.1)
1821 - Fixed instance migration with missing disk symlinks
1822 - Fixed handling of unknown jobs in ``gnt-job archive``
1823 - And many other small fixes/improvements
1827 - Enhanced both the unittest and the QA coverage
1828 - Switched the opcode validation to a generic model, and extended the
1829 validation to all opcode parameters
1830 - Changed more parts of the code that write shell scripts to use the
1832 - Switched the master daemon to use the asyncore library for the Luxi
1839 *(Released Thu, 17 Jun 2010)*
1841 - Added tool (``move-instance``) and infrastructure to move instances
1842 between separate clusters (see :doc:`separate documentation
1843 <move-instance>` and :doc:`design document <design-2.2>`)
1844 - Added per-request RPC timeout
1845 - RAPI now requires a Content-Type header for requests with a body (e.g.
1846 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1847 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1848 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1850 - Implemented initial support for running Ganeti daemons as separate
1851 users, see configure-time flags ``--with-user-prefix`` and
1852 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1854 - Instances can be removed after export (``gnt-backup export
1855 --remove-instance``)
1856 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1857 key (instead of 1024 bit)
1858 - Added new cluster configuration file for cluster domain secret
1859 - Import/export now use SSL instead of SSH
1860 - Added support for showing estimated time when exporting an instance,
1861 see the ``ganeti-os-interface(7)`` manpage and look for
1868 *(Released Tue, 16 Nov 2010)*
1870 Some more bugfixes. Unless critical bugs occur, this will be the last
1873 - Fix case of MAC special-values
1874 - Fix mac checker regex
1875 - backend: Fix typo causing "out of range" error
1876 - Add missing --units in gnt-instance list man page
1882 *(Released Tue, 24 Aug 2010)*
1885 - Don't ignore secondary node silently on non-mirrored disk templates
1887 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1888 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1889 - Properly document the usb_mouse parameter
1890 - Fix path in ganeti-rapi(8) (issue 116)
1891 - Adjust error message when the ganeti user's .ssh directory is
1893 - Add same-node-check when changing the disk template to drbd
1899 *(Released Fri, 16 Jul 2010)*
1902 - Add an option to only select some reboot types during qa/burnin.
1903 (on some hypervisors consequent reboots are not supported)
1904 - Fix infrequent race condition in master failover. Sometimes the old
1905 master ip address would be still detected as up for a short time
1906 after it was removed, causing failover to fail.
1907 - Decrease mlockall warnings when the ctypes module is missing. On
1908 Python 2.4 we support running even if no ctypes module is installed,
1909 but we were too verbose about this issue.
1910 - Fix building on old distributions, on which man doesn't have a
1912 - Fix RAPI not to ignore the MAC address on instance creation
1913 - Implement the old instance creation format in the RAPI client.
1919 *(Released Thu, 01 Jul 2010)*
1921 A small bugfix release:
1922 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1923 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1924 a non-existing option
1925 - Fix instance create: the --force-variant option was ignored
1926 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1927 - Fix modify node storage with non-FQDN arguments
1928 - Fix RAPI client to authenticate under Python 2.6 when used
1929 for more than 5 requests needing authentication
1930 - Fix gnt-instance modify -t (storage) giving a wrong error message
1931 when converting a non-shutdown drbd instance to plain
1937 *(Released Fri, 18 Jun 2010)*
1939 A small bugfix release:
1941 - Fix live migration of KVM instances started with older Ganeti
1942 versions which had fewer hypervisor parameters
1943 - Fix gnt-instance grow-disk on down instances
1944 - Fix an error-reporting bug during instance migration
1945 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1946 silently ignoring broken ones
1947 - Fix an RPC error reporting bug affecting, for example, RAPI client
1949 - Fix bug triggered by different API version os-es on different nodes
1950 - Fix a bug in instance startup with custom hvparams: OS level
1951 parameters would fail to be applied.
1952 - Fix the RAPI client under Python 2.6 (but more work is needed to
1953 make it work completely well with OpenSSL)
1954 - Fix handling of errors when resolving names from DNS
1960 *(Released Thu, 3 Jun 2010)*
1962 A medium sized development cycle. Some new features, and some
1963 fixes/small improvements/cleanups.
1965 Significant features
1966 ~~~~~~~~~~~~~~~~~~~~
1968 The node deamon now tries to mlock itself into memory, unless the
1969 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1970 its logs, and falls back to console logging. This allows emergency
1971 features such as ``gnt-node powercycle`` to work even in the event of a
1972 broken node disk (tested offlining the disk hosting the node's
1973 filesystem and dropping its memory caches; don't try this at home)
1975 KVM: add vhost-net acceleration support. It can be tested with a new
1976 enough version of the kernel and of qemu-kvm.
1978 KVM: Add instance chrooting feature. If you use privilege dropping for
1979 your VMs you can also now force them to chroot to an empty directory,
1980 before starting the emulated guest.
1982 KVM: Add maximum migration bandwith and maximum downtime tweaking
1983 support (requires a new-enough version of qemu-kvm).
1985 Cluster verify will now warn if the master node doesn't have the master
1986 ip configured on it.
1988 Add a new (incompatible) instance creation request format to RAPI which
1989 supports all parameters (previously only a subset was supported, and it
1990 wasn't possible to extend the old format to accomodate all the new
1991 features. The old format is still supported, and a client can check for
1992 this feature, before using it, by checking for its presence in the
1993 ``features`` RAPI resource.
1995 Now with ancient latin support. Try it passing the ``--roman`` option to
1996 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1997 (requires the python-roman module to be installed, in order to work).
2002 As usual many internal code refactorings, documentation updates, and
2005 - Lots of improvements and cleanups to the experimental Remote API
2006 (RAPI) client library.
2007 - A new unit test suite for the core daemon libraries.
2008 - A fix to creating missing directories makes sure the umask is not
2009 applied anymore. This enforces the same directory permissions
2011 - Better handling terminating daemons with ctrl+c (used when running
2012 them in debugging mode).
2013 - Fix a race condition in live migrating a KVM instance, when stat()
2014 on the old proc status file returned EINVAL, which is an unexpected
2016 - Fixed manpage checking with newer man and utf-8 charachters. But now
2017 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2023 *(Released Fri, 7 May 2010)*
2025 Fix a bug which prevented untagged KVM instances from starting.
2031 *(Released Fri, 7 May 2010)*
2033 Another release with a long development cycle, during which many
2034 different features were added.
2036 Significant features
2037 ~~~~~~~~~~~~~~~~~~~~
2039 The KVM hypervisor now can run the individual instances as non-root, to
2040 reduce the impact of a VM being hijacked due to bugs in the
2041 hypervisor. It is possible to run all instances as a single (non-root)
2042 user, to manually specify a user for each instance, or to dynamically
2043 allocate a user out of a cluster-wide pool to each instance, with the
2044 guarantee that no two instances will run under the same user ID on any
2047 An experimental RAPI client library, that can be used standalone
2048 (without the other Ganeti libraries), is provided in the source tree as
2049 ``lib/rapi/client.py``. Note this client might change its interface in
2050 the future, as we iterate on its capabilities.
2052 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2053 replace the cluster's certificates and crypto keys. This might help in
2054 case they have been compromised, or have simply expired.
2056 A new disk option for instance creation has been added that allows one
2057 to "adopt" currently existing logical volumes, with data
2058 preservation. This should allow easier migration to Ganeti from
2059 unmanaged (or managed via other software) instances.
2061 Another disk improvement is the possibility to convert between redundant
2062 (DRBD) and plain (LVM) disk configuration for an instance. This should
2063 allow better scalability (starting with one node and growing the
2064 cluster, or shrinking a two-node cluster to one node).
2066 A new feature that could help with automated node failovers has been
2067 implemented: if a node sees itself as offline (by querying the master
2068 candidates), it will try to shutdown (hard) all instances and any active
2069 DRBD devices. This reduces the risk of duplicate instances if an
2070 external script automatically failovers the instances on such nodes. To
2071 enable this, the cluster parameter ``maintain_node_health`` should be
2072 enabled; in the future this option (per the name) will enable other
2073 automatic maintenance features.
2075 Instance export/import now will reuse the original instance
2076 specifications for all parameters; that means exporting an instance,
2077 deleting it and the importing it back should give an almost identical
2078 instance. Note that the default import behaviour has changed from
2079 before, where it created only one NIC; now it recreates the original
2082 Cluster verify has added a few new checks: SSL certificates validity,
2083 /etc/hosts consistency across the cluster, etc.
2088 As usual, many internal changes were done, documentation fixes,
2091 - Fixed cluster initialization with disabled cluster storage (regression
2092 introduced in 2.1.1)
2093 - File-based storage supports growing the disks
2094 - Fixed behaviour of node role changes
2095 - Fixed cluster verify for some corner cases, plus a general rewrite of
2096 cluster verify to allow future extension with more checks
2097 - Fixed log spamming by watcher and node daemon (regression introduced
2099 - Fixed possible validation issues when changing the list of enabled
2101 - Fixed cleanup of /etc/hosts during node removal
2102 - Fixed RAPI response for invalid methods
2103 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2104 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2105 from ide disks, etc.)
2106 - Allow OS changes without re-installation (to record a changed OS
2107 outside of Ganeti, or to allow OS renames)
2108 - Allow instance creation without OS installation (useful for example if
2109 the OS will be installed manually, or restored from a backup not in
2111 - Implemented option to make cluster ``copyfile`` use the replication
2113 - Added list of enabled hypervisors to ssconf (possibly useful for
2115 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2117 - A partial form of node re-IP is possible via node readd, which now
2118 allows changed node primary IP
2119 - Command line utilities now show an informational message if the job is
2121 - The logs of the master daemon now show the PID/UID/GID of the
2128 *(Released Fri, 12 Mar 2010)*
2130 During the 2.1.0 long release candidate cycle, a lot of improvements and
2131 changes have accumulated with were released later as 2.1.1.
2136 The node evacuate command (``gnt-node evacuate``) was significantly
2137 rewritten, and as such the IAllocator protocol was changed - a new
2138 request type has been added. This unfortunate change during a stable
2139 series is designed to improve performance of node evacuations; on
2140 clusters with more than about five nodes and which are well-balanced,
2141 evacuation should proceed in parallel for all instances of the node
2142 being evacuated. As such, any existing IAllocator scripts need to be
2143 updated, otherwise the above command will fail due to the unknown
2144 request. The provided "dumb" allocator has not been updated; but the
2145 ganeti-htools package supports the new protocol since version 0.2.4.
2147 Another important change is increased validation of node and instance
2148 names. This might create problems in special cases, if invalid host
2149 names are being used.
2151 Also, a new layer of hypervisor parameters has been added, that sits at
2152 OS level between the cluster defaults and the instance ones. This allows
2153 customisation of virtualization parameters depending on the installed
2154 OS. For example instances with OS 'X' may have a different KVM kernel
2155 (or any other parameter) than the cluster defaults. This is intended to
2156 help managing a multiple OSes on the same cluster, without manual
2157 modification of each instance's parameters.
2159 A tool for merging clusters, ``cluster-merge``, has been added in the
2160 tools sub-directory.
2165 - Improved the int/float conversions that should make the code more
2166 robust in face of errors from the node daemons
2167 - Fixed the remove node code in case of internal configuration errors
2168 - Fixed the node daemon behaviour in face of inconsistent queue
2169 directory (e.g. read-only file-system where we can't open the files
2171 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2172 now it either aborts cleanly or, if given the new "auto_promote"
2173 parameter, will automatically promote other nodes as needed
2174 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2175 completely prevent Ganeti from working
2176 - Fixed bug for instance export when not all disks were successfully
2178 - Fixed behaviour of node add when the new node is slow in starting up
2180 - Fixed handling of signals in the LUXI client, which should improve
2181 behaviour of command-line scripts
2182 - Added checks for invalid node/instance names in the configuration (now
2183 flagged during cluster verify)
2184 - Fixed watcher behaviour for disk activation errors
2185 - Fixed two potentially endless loops in http library, which led to the
2186 RAPI daemon hanging and consuming 100% CPU in some cases
2187 - Fixed bug in RAPI daemon related to hashed passwords
2188 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2190 - Enhanced compatibility with non-Debian OSes, but not using absolute
2191 path in some commands and allowing customisation of the ssh
2192 configuration directory
2193 - Fixed possible future issue with new Python versions by abiding to the
2194 proper use of ``__slots__`` attribute on classes
2195 - Added checks that should prevent directory traversal attacks
2196 - Many documentation fixes based on feedback from users
2201 - Added an "early_release" more for instance replace disks and node
2202 evacuate, where we release locks earlier and thus allow higher
2203 parallelism within the cluster
2204 - Added watcher hooks, intended to allow the watcher to restart other
2205 daemons (e.g. from the ganeti-nbma project), but they can be used of
2206 course for any other purpose
2207 - Added a compile-time disable for DRBD barriers, to increase
2208 performance if the administrator trusts the power supply or the
2209 storage system to not lose writes
2210 - Added the option of using syslog for logging instead of, or in
2211 addition to, Ganeti's own log files
2212 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2213 can indeed boot from a paravirtual NIC
2214 - Added a generic debug level for many operations; while this is not
2215 used widely yet, it allows one to pass the debug value all the way to
2217 - Enhanced the hooks environment for instance moves (failovers,
2218 migrations) where the primary/secondary nodes changed during the
2219 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2220 - Enhanced data validations for many user-supplied values; one important
2221 item is the restrictions imposed on instance and node names, which
2222 might reject some (invalid) host names
2223 - Add a configure-time option to disable file-based storage, if it's not
2224 needed; this allows greater security separation between the master
2225 node and the other nodes from the point of view of the inter-node RPC
2227 - Added user notification in interactive tools if job is waiting in the
2228 job queue or trying to acquire locks
2229 - Added log messages when a job is waiting for locks
2230 - Added filtering by node tags in instance operations which admit
2231 multiple instances (start, stop, reboot, reinstall)
2232 - Added a new tool for cluster mergers, ``cluster-merge``
2233 - Parameters from command line which are of the form ``a=b,c=d`` can now
2234 use backslash escapes to pass in values which contain commas,
2235 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2237 - For KVM, the instance name is the first parameter passed to KVM, so
2238 that it's more visible in the process list
2244 *(Released Tue, 2 Mar 2010)*
2246 Ganeti 2.1 brings many improvements with it. Major changes:
2248 - Added infrastructure to ease automated disk repairs
2249 - Added new daemon to export configuration data in a cheaper way than
2250 using the remote API
2251 - Instance NICs can now be routed instead of being associated with a
2253 - Improved job locking logic to reduce impact of jobs acquiring multiple
2254 locks waiting for other long-running jobs
2256 In-depth implementation details can be found in the Ganeti 2.1 design
2262 - Added chroot hypervisor
2263 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2265 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2266 ``bootloader_path`` and ``bootloader_args``)
2267 - Added the ``use_localtime`` option for the xen-hvm and kvm
2268 hypervisors, and the default value for this has changed to false (in
2269 2.0 xen-hvm always enabled it)
2270 - Added luxi call to submit multiple jobs in one go
2271 - Added cluster initialization option to not modify ``/etc/hosts``
2273 - Added network interface parameters
2274 - Added dry run mode to some LUs
2275 - Added RAPI resources:
2277 - ``/2/instances/[instance_name]/info``
2278 - ``/2/instances/[instance_name]/replace-disks``
2279 - ``/2/nodes/[node_name]/evacuate``
2280 - ``/2/nodes/[node_name]/migrate``
2281 - ``/2/nodes/[node_name]/role``
2282 - ``/2/nodes/[node_name]/storage``
2283 - ``/2/nodes/[node_name]/storage/modify``
2284 - ``/2/nodes/[node_name]/storage/repair``
2286 - Added OpCodes to evacuate or migrate all instances on a node
2287 - Added new command to list storage elements on nodes (``gnt-node
2288 list-storage``) and modify them (``gnt-node modify-storage``)
2289 - Added new ssconf files with master candidate IP address
2290 (``ssconf_master_candidates_ips``), node primary IP address
2291 (``ssconf_node_primary_ips``) and node secondary IP address
2292 (``ssconf_node_secondary_ips``)
2293 - Added ``ganeti-confd`` and a client library to query the Ganeti
2294 configuration via UDP
2295 - Added ability to run hooks after cluster initialization and before
2297 - Added automatic mode for disk replace (``gnt-instance replace-disks
2299 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2300 after catastrophic data-loss
2301 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2303 - Added ``gnt-instance move`` command to move instances
2304 - Added ``gnt-cluster watcher`` command to control watcher
2305 - Added ``gnt-node powercycle`` command to powercycle nodes
2306 - Added new job status field ``lock_status``
2307 - Added parseable error codes to cluster verification (``gnt-cluster
2308 verify --error-codes``) and made output less verbose (use
2309 ``--verbose`` to restore previous behaviour)
2310 - Added UUIDs to the main config entities (cluster, nodes, instances)
2311 - Added support for OS variants
2312 - Added support for hashed passwords in the Ganeti remote API users file
2314 - Added option to specify maximum timeout on instance shutdown
2315 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2316 - Added new helper script to start and stop Ganeti daemons
2317 (``daemon-util``), with the intent to reduce the work necessary to
2318 adjust Ganeti for non-Debian distributions and to start/stop daemons
2320 - Added more unittests
2321 - Fixed critical bug in ganeti-masterd startup
2322 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2323 now customisable at the cluster level for both the KVM and Xen
2324 hypervisors using the new ``migration_port`` parameter
2325 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2326 reinstalling an instance
2327 - Allowed ``@`` in tag names
2328 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2329 - Many documentation updates
2330 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2331 - ``gnt-instance reinstall`` can now reinstall multiple instances
2332 - Updated many command line parameters
2333 - Introduced new OS API version 15
2334 - No longer support a default hypervisor
2335 - Treat virtual LVs as inexistent
2336 - Improved job locking logic to reduce lock contention
2337 - Match instance and node names case insensitively
2338 - Reimplemented bash completion script to be more complete
2345 *(Released Thu, 4 Feb 2010)*
2347 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2348 - Fix a string formatting bug
2349 - Improve safety of the code in some error paths
2350 - Improve data validation in the master of values returned from nodes
2356 *(Released Thu, 17 Dec 2009)*
2358 - Fix security issue due to missing validation of iallocator names; this
2359 allows local and remote execution of arbitrary executables
2360 - Fix failure of gnt-node list during instance removal
2361 - Ship the RAPI documentation in the archive
2367 *(Released Wed, 30 Sep 2009)*
2369 - Fixed many wrong messages
2370 - Fixed a few bugs related to the locking library
2371 - Fixed MAC checking at instance creation time
2372 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2373 - Fixed a few issues related to signal handling in both daemons and
2375 - Fixed the example startup script provided
2376 - Fixed insserv dependencies in the example startup script (patch from
2378 - Fixed handling of drained nodes in the iallocator framework
2379 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2381 - Fixed error related to invalid job IDs in job polling
2382 - Fixed job/opcode persistence on unclean master shutdown
2383 - Fixed handling of partial job processing after unclean master
2385 - Fixed error reporting from LUs, previously all errors were converted
2386 into execution errors
2387 - Fixed error reporting from burnin
2388 - Decreased significantly the memory usage of the job queue
2389 - Optimised slightly multi-job submission
2390 - Optimised slightly opcode loading
2391 - Backported the multi-job submit framework from the development
2392 branch; multi-instance start and stop should be faster
2393 - Added script to clean archived jobs after 21 days; this will reduce
2394 the size of the queue directory
2395 - Added some extra checks in disk size tracking
2396 - Added an example ethers hook script
2397 - Added a cluster parameter that prevents Ganeti from modifying of
2399 - Added more node information to RAPI responses
2400 - Added a ``gnt-job watch`` command that allows following the ouput of a
2402 - Added a bind-address option to ganeti-rapi
2403 - Added more checks to the configuration verify
2404 - Enhanced the burnin script such that some operations can be retried
2406 - Converted instance reinstall to multi-instance model
2412 *(Released Fri, 7 Aug 2009)*
2414 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2415 to allow using the pre-2.0.2 behaviour in activation, if any existing
2416 instances have mismatched disk sizes in the configuration
2417 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2418 any configuration mismatches for disk sizes
2419 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2420 failover to work on two-node clusters
2421 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2423 - Fixed detection of OS script errors in ``gnt-backup export``
2424 - Fixed exit code of ``gnt-backup export``
2430 *(Released Fri, 17 Jul 2009)*
2432 - Added experimental support for stripped logical volumes; this should
2433 enhance performance but comes with a higher complexity in the block
2434 device handling; stripping is only enabled when passing
2435 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2436 affected even in the non-stripped mode
2437 - Improved resiliency against transient failures at the end of DRBD
2438 resyncs, and in general of DRBD resync checks
2439 - Fixed a couple of issues with exports and snapshot errors
2440 - Fixed a couple of issues in instance listing
2441 - Added display of the disk size in ``gnt-instance info``
2442 - Fixed checking for valid OSes in instance creation
2443 - Fixed handling of the "vcpus" parameter in instance listing and in
2444 general of invalid parameters
2445 - Fixed http server library, and thus RAPI, to handle invalid
2446 username/password combinations correctly; this means that now they
2447 report unauthorized for queries too, not only for modifications,
2448 allowing earlier detect of configuration problems
2449 - Added a new "role" node list field, equivalent to the master/master
2450 candidate/drained/offline flags combinations
2451 - Fixed cluster modify and changes of candidate pool size
2452 - Fixed cluster verify error messages for wrong files on regular nodes
2453 - Fixed a couple of issues with node demotion from master candidate role
2454 - Fixed node readd issues
2455 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2456 - Added a new ``--no-voting`` option for masterfailover to fix failover
2457 on two-nodes clusters when the former master node is unreachable
2458 - Added instance reinstall over RAPI
2464 *(Released Tue, 16 Jun 2009)*
2466 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2467 allow re-adding the start in single-user option (regression from 1.2)
2468 - the watcher writes the instance status to a file, to allow monitoring
2469 to report the instance status (from the master) based on cached
2470 results of the watcher's queries; while this can get stale if the
2471 watcher is being locked due to other work on the cluster, this is
2472 still an improvement
2473 - the watcher now also restarts the node daemon and the rapi daemon if
2475 - fixed the watcher to handle full and drained queue cases
2476 - hooks export more instance data in the environment, which helps if
2477 hook scripts need to take action based on the instance's properties
2478 (no longer need to query back into ganeti)
2479 - instance failovers when the instance is stopped do not check for free
2480 RAM, so that failing over a stopped instance is possible in low memory
2482 - rapi uses queries for tags instead of jobs (for less job traffic), and
2483 for cluster tags it won't talk to masterd at all but read them from
2485 - a couple of error handling fixes in RAPI
2486 - drbd handling: improved the error handling of inconsistent disks after
2487 resync to reduce the frequency of "there are some degraded disks for
2488 this instance" messages
2489 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2490 error handling path called a wrong function name)
2496 *(Released Wed, 27 May 2009)*
2498 - no changes from rc5
2504 *(Released Wed, 20 May 2009)*
2506 - fix a couple of bugs (validation, argument checks)
2507 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2508 - some small improvements to RAPI and IAllocator
2509 - make watcher automatically start the master daemon if down
2515 *(Released Mon, 27 Apr 2009)*
2517 - change the OS list to not require locks; this helps with big clusters
2518 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2519 volume group is broken
2520 - ``gnt-instance info``, without any arguments, doesn't run for all
2521 instances anymore; either pass ``--all`` or pass the desired
2522 instances; this helps against mistakes on big clusters where listing
2523 the information for all instances takes a long time
2524 - miscellaneous doc and man pages fixes
2530 *(Released Wed, 8 Apr 2009)*
2532 - Change the internal locking model of some ``gnt-node`` commands, in
2533 order to reduce contention (and blocking of master daemon) when
2534 batching many creation/reinstall jobs
2535 - Fixes to Xen soft reboot
2536 - No longer build documentation at build time, instead distribute it in
2537 the archive, in order to reduce the need for the whole docbook/rst
2544 *(Released Fri, 27 Mar 2009)*
2546 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2547 - Fix watcher startup sequence, improves the behaviour of busy clusters
2548 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2549 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2550 bind address and other places
2551 - Some documentation fixes and updates
2557 *(Released Mon, 2 Mar 2009)*
2559 - More documentation updates, now all docs should be more-or-less
2561 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2563 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2565 - Fix an issue related to $libdir/run/ganeti and cluster creation
2571 *(Released Thu, 19 Feb 2009)*
2573 - Xen PVM and KVM have switched the default value for the instance root
2574 disk to the first partition on the first drive, instead of the whole
2575 drive; this means that the OS installation scripts must be changed
2577 - Man pages have been updated
2578 - RAPI has been switched by default to HTTPS, and the exported functions
2579 should all work correctly
2580 - RAPI v1 has been removed
2581 - Many improvements to the KVM hypervisor
2582 - Block device errors are now better reported
2583 - Many other bugfixes and small improvements
2589 *(Released Mon, 26 Jan 2009)*
2591 - Version 2 is a general rewrite of the code and therefore the
2592 differences are too many to list, see the design document for 2.0 in
2593 the ``doc/`` subdirectory for more details
2594 - In this beta version there is not yet a migration path from 1.2 (there
2595 will be one in the final 2.0 release)
2596 - A few significant changes are:
2598 - all commands are executed by a daemon (``ganeti-masterd``) and the
2599 various ``gnt-*`` commands are just front-ends to it
2600 - all the commands are entered into, and executed from a job queue,
2601 see the ``gnt-job(8)`` manpage
2602 - the RAPI daemon supports read-write operations, secured by basic
2603 HTTP authentication on top of HTTPS
2604 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2605 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2606 to migrate to DRBD 8 first while still running Ganeti 1.2)
2607 - DRBD devices are using statically allocated minor numbers, which
2608 will be assigned to existing instances during the migration process
2609 - there is support for both Xen PVM and Xen HVM instances running on
2611 - KVM virtualization is supported too
2612 - file-based storage has been implemented, which means that it is
2613 possible to run the cluster without LVM and DRBD storage, for
2614 example using a shared filesystem exported from shared storage (and
2615 still have live migration)
2621 *(Released Tue, 13 Jan 2009)*
2623 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2624 - Reuse the old instance mac address by default on instance import, if
2625 the instance name is the same.
2626 - Handle situations in which the node info rpc returns incomplete
2628 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2629 - Improved version of batcher:
2631 - state file support
2632 - instance mac address support
2633 - support for HVM clusters/instances
2635 - Add an option to show the number of cpu sockets and nodes in
2637 - Support OSes that handle more than one version of the OS api (but do
2638 not change the current API in any other way)
2639 - Fix ``gnt-node migrate``
2640 - ``gnt-debug`` man page
2641 - Fixes various more typos and small issues
2642 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2648 *(Released Wed, 24 Sep 2008)*
2650 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2651 type of disk exported to fully virtualized instances.
2652 - provide access to the serial console of HVM instances
2653 - instance auto_balance flag, set by default. If turned off it will
2654 avoid warnings on cluster verify if there is not enough memory to fail
2655 over an instance. in the future it will prevent automatically failing
2656 it over when we will support that.
2657 - batcher tool for instance creation, see ``tools/README.batcher``
2658 - ``gnt-instance reinstall --select-os`` to interactively select a new
2659 operating system when reinstalling an instance.
2660 - when changing the memory amount on instance modify a check has been
2661 added that the instance will be able to start. also warnings are
2662 emitted if the instance will not be able to fail over, if auto_balance
2664 - documentation fixes
2665 - sync fields between ``gnt-instance list/modify/add/import``
2666 - fix a race condition in drbd when the sync speed was set after giving
2667 the device a remote peer.
2673 *(Released Tue, 22 Jul 2008)*
2675 - note: the allowed size and number of tags per object were reduced
2676 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2677 - fixed twisted 8.x compatibility
2678 - fixed ``gnt-instance replace-disks`` with iallocator
2679 - add TCP keepalives on twisted connections to detect restarted nodes
2680 - disk increase support, see ``gnt-instance grow-disk``
2681 - implement bulk node/instance query for RAPI
2682 - add tags in node/instance listing (optional)
2683 - experimental migration (and live migration) support, read the man page
2684 for ``gnt-instance migrate``
2685 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2686 has some small improvements in handling its state file
2692 *(Released Fri, 13 Jun 2008)*
2694 - Experimental readonly, REST-based remote API implementation;
2695 automatically started on master node, TCP port 5080, if enabled by
2696 ``--enable-rapi`` parameter to configure script.
2697 - Instance allocator support. Add and import instance accept a
2698 ``--iallocator`` parameter, and call that instance allocator to decide
2699 which node to use for the instance. The iallocator document describes
2700 what's expected from an allocator script.
2701 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2702 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2703 node is lost there is still enough memory to fail over the instances
2705 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2706 ``gnt-cluster verify``, to check for site-specific compliance. All the
2707 hooks will run, and their output, if any, will be displayed. Any
2708 failing hook will make the verification return an error value.
2709 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2710 primary and secondary interfaces
2711 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2712 that is still declared as part of the cluster and has failed.
2713 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2714 specifying output fields, that just adds the chosen fields to the
2716 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2717 export from the filesystem.
2718 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2719 have been added. Using them you can enable/disable acpi and pae
2720 support, and specify a path for a cd image to be exported to the
2721 instance. These parameters as the name suggest only work on HVM
2723 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2724 PAE support will be set to the previously hardcoded values, but the
2725 (previously hardcoded) path to the CDROM ISO image will be unset and
2726 if required, needs to be set manually with ``gnt-instance modify``
2728 - The address to which an instance's VNC console is bound is now
2729 selectable per-instance, rather than being cluster wide. Of course
2730 this only applies to instances controlled via VNC, so currently just
2731 applies to HVM clusters.
2737 *(Released Mon, 18 Feb 2008)*
2739 - more tweaks to the disk activation code (especially helpful for DRBD)
2740 - change the default ``gnt-instance list`` output format, now there is
2741 one combined status field (see the manpage for the exact values this
2743 - some more fixes for the mac export to hooks change
2744 - make Ganeti not break with DRBD 8.2.x (which changed the version
2745 format in ``/proc/drbd``) (issue 24)
2746 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2747 template, allowing migration from DRBD0.7+MD to DRBD8
2753 *(Released Wed, 30 Jan 2008)*
2755 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2757 - add command aliases infrastructure and a few aliases
2758 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2759 man pages and the ``--help`` option of ``gnt-node
2760 list``/``gnt-instance list``
2761 - fix ``gnt-backup list`` with down nodes (issue 21)
2762 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2763 - fix the dist archive and add a check for including svn/git files in
2765 - some developer-related changes: improve the burnin and the QA suite,
2766 add an upload script for testing during development
2772 *(Released Wed, 16 Jan 2008)*
2774 - experimental HVM support, read the install document, section
2775 "Initializing the cluster"
2776 - allow for the PVM hypervisor per-instance kernel and initrd paths
2777 - add a new command ``gnt-cluster verify-disks`` which uses a new
2778 algorithm to improve the reconnection of the DRBD pairs if the device
2779 on the secondary node has gone away
2780 - make logical volume code auto-activate LVs at disk activation time
2781 - slightly improve the speed of activating disks
2782 - allow specification of the MAC address at instance creation time, and
2783 changing it later via ``gnt-instance modify``
2784 - fix handling of external commands that generate lots of output on
2786 - update documentation with regard to minimum version of DRBD8 supported
2792 *(Released Tue, 4 Dec 2007)*
2794 - Log the ``xm create`` output to the node daemon log on failure (to
2795 help diagnosing the error)
2796 - In debug mode, log all external commands output if failed to the logs
2797 - Change parsing of lvm commands to ignore stderr
2803 *(Released Wed, 28 Nov 2007)*
2805 - Another round of updates to the DRBD 8 code to deal with more failures
2806 in the replace secondary node operation
2807 - Some more logging of failures in disk operations (lvm, drbd)
2808 - A few documentation updates
2815 *(Released Tue, 13 Nov 2007)*
2817 - Change configuration file format from Python's Pickle to JSON.
2818 Upgrading is possible using the cfgupgrade utility.
2819 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2820 faster replace disks and is more stable (DRBD 8 has many improvements
2821 compared to DRBD 0.7)
2822 - Added command line tags support (see man pages for ``gnt-instance``,
2823 ``gnt-node``, ``gnt-cluster``)
2824 - Added instance rename support
2825 - Added multi-instance startup/shutdown
2826 - Added cluster rename support
2827 - Added ``gnt-node evacuate`` to simplify some node operations
2828 - Added instance reboot operation that can speedup reboot as compared to
2830 - Soften the requirement that hostnames are in FQDN format
2831 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2833 - Removed dependency on debian's patched fping that uses the
2834 non-standard ``-S`` option
2835 - Now the OS definitions are searched for in multiple, configurable
2836 paths (easier for distros to package)
2837 - Some changes to the hooks infrastructure (especially the new
2838 post-configuration update hook)
2839 - Other small bugfixes
2841 .. vim: set textwidth=72 syntax=rst :