10 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
11 the correct command syntax. It is possible to use different DRBD versions
12 on different nodes as long as they are compatible to each other. This
13 enables rolling upgrades of DRBD with no downtime. As permanent operation
14 of different DRBD versions within a node group is discouraged,
15 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
16 - hroller now also plans for capacity to move non-redundant instances off
17 any node to be rebooted; the old behavior of completely ignoring any
18 non-redundant instances can be restored by adding the --ignore-non-redundant
20 - The cluster option '--no-lvm-storage' was removed in favor of the new option
21 '--enabled-disk-templates'.
27 *(Released Mon, 24 Jun 2013)*
29 Incompatible/important changes
30 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
32 - Instance policy can contain multiple instance specs, as described in
33 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
34 <design-partitioned>`. As a consequence, it's not possible to partially change
35 or override instance specs. Bounding specs (min and max) can be specified as a
36 whole using the new option ``--ipolicy-bounds-specs``, while standard
37 specs use the new option ``--ipolicy-std-specs``.
38 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
39 gnt-instance is a valid YAML object.
40 - hail now honors network restrictions when allocating nodes. This led to an
41 update of the IAllocator protocol. See the IAllocator documentation for
47 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
48 to always require authentication, ``--require-authentication``. It can
49 be specified in ``$sysconfdir/default/ganeti``.
50 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
51 be used to manage the disk templates to be used by instances in the cluster.
52 Initially, it will be set to a list that includes plain, drbd, if they were
53 enabled by specifying a volume group name, and file and sharedfile, if those
54 were enabled at configure time. Additionally, it will include all disk
55 templates that are currently used by instances. The order of disk templates
56 will be based on Ganeti's history of supporting them. In the future, the
57 first entry of the list will be used as a default disk template on instance
59 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
60 configuration back to the previous stable version.
61 - Disk templates in group ipolicy can be restored to the default value.
62 - Initial support for diskless instances and virtual clusters in QA.
63 - More QA and unit tests for instance policies.
64 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
65 describing why the opcode itself was executed.
66 - The monitoring daemon is now available. It allows users to query the cluster
67 for obtaining information about the status of the system. The daemon is only
68 responsible for providing the information over the network: the actual data
69 gathering is performed by data collectors (currently, only the DRBD status
70 collector is available).
71 - In order to help developers work on Ganeti, a new script
72 (``devel/build_chroot``) is provided, for building a chroot that contains all
73 the required development libraries and tools for compiling Ganeti on a Debian
75 - A new tool, ``harep``, for performing self-repair and recreation of instances
76 in Ganeti has been added.
77 - Split queries are enabled for tags, network, exports, cluster info, groups,
79 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
80 It prints the command line to set the current policies, to ease
85 The following new dependencies have been added:
88 - The ``curl`` library is not optional anymore for compiling the Haskell code.
89 - ``snap-server`` library (if monitoring is enabled).
92 - The minimum Python version needed to run Ganeti is now 2.6.
93 - ``yaml`` library (only for running the QA).
99 *(Released Tue, 25 Jun 2013)*
101 Incompatible/important changes
102 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
104 - Instance policies for disk size were documented to be on a per-disk
105 basis, but hail applied them to the sum of all disks. This has been
107 - ``hbal`` will now exit with status 0 if, during job execution over
108 LUXI, early exit has been requested and all jobs are successful;
109 before, exit status 1 was used, which cannot be differentiated from
111 - Compatibility with newer versions of rbd has been fixed
112 - ``gnt-instance batch-create`` has been changed to use the bulk create
113 opcode from Ganeti. This lead to incompatible changes in the format of
114 the JSON file. It's now not a custom dict anymore but a dict
115 compatible with the ``OpInstanceCreate`` opcode.
116 - Parent directories for file storage need to be listed in
117 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
118 write the file automatically based on old configuration values, but it
119 can not distribute it across all nodes and the file contents should be
120 verified. Use ``gnt-cluster copyfile
121 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
122 upgraded. The reason for requiring this list of paths now is that
123 before it would have been possible to inject new paths via RPC,
124 allowing files to be created in arbitrary locations. The RPC protocol
125 is protected using SSL/X.509 certificates, but as a design principle
126 Ganeti does not permit arbitrary paths to be passed.
127 - The parsing of the variants file for OSes (see
128 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
129 lines and comment lines (starting with ``#``) are ignored for better
131 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
132 longer available. ``gnt-node add`` now invokes a new tool on the
133 destination node, named ``prepare-node-join``, to configure the SSH
134 daemon. Paramiko is no longer necessary to configure nodes' SSH
135 daemons via ``gnt-node add``.
136 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
137 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
138 the flag is not reset after a master failover.
139 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
140 is recommended. 2.6 will be mandatory from the 2.8 series.
146 - New network management functionality to support automatic allocation
147 of IP addresses and managing of network parameters. See
148 :manpage:`gnt-network(8)` for more details.
149 - New external storage backend, to allow managing arbitrary storage
150 systems external to the cluster. See
151 :manpage:`ganeti-extstorage-interface(7)`.
152 - New ``exclusive-storage`` node parameter added, restricted to
153 nodegroup level. When it's set to true, physical disks are assigned in
154 an exclusive fashion to instances, as documented in :doc:`Partitioned
155 Ganeti <design-partitioned>`. Currently, only instances using the
156 ``plain`` disk template are supported.
157 - The KVM hypervisor has been updated with many new hypervisor
158 parameters, including a generic one for passing arbitrary command line
159 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
160 compatible up to qemu 1.4.
161 - A new tool, called ``mon-collector``, is the stand-alone executor of
162 the data collectors for a monitoring system. As of this version, it
163 just includes the DRBD data collector, that can be executed by calling
164 ``mon-collector`` using the ``drbd`` parameter. See
165 :manpage:`mon-collector(7)`.
166 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
167 for RAPI users. It allows granting permissions to query for
168 information to a specific user without giving
169 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
170 - A new tool named ``node-cleanup`` has been added. It cleans remains of
171 a cluster from a machine by stopping all daemons, removing
172 certificates and ssconf files. Unless the ``--no-backup`` option is
173 given, copies of the certificates are made.
174 - Instance creations now support the use of opportunistic locking,
175 potentially speeding up the (parallel) creation of multiple instances.
176 This feature is currently only available via the :doc:`RAPI
177 <rapi>` interface and when an instance allocator is used. If the
178 ``opportunistic_locking`` parameter is set the opcode will try to
179 acquire as many locks as possible, but will not wait for any locks
180 held by other opcodes. If not enough resources can be found to
181 allocate the instance, the temporary error code
182 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
183 retried thereafter, with or without opportunistic locking.
184 - New experimental linux-ha resource scripts.
185 - Restricted-commands support: ganeti can now be asked (via command line
186 or rapi) to perform commands on a node. These are passed via ganeti
187 RPC rather than ssh. This functionality is restricted to commands
188 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
189 reasons. The file is not copied automatically.
195 - Diskless instances are now externally mirrored (Issue 237). This for
196 now has only been tested in conjunction with explicit target nodes for
198 - Queries not needing locks or RPC access to the node can now be
199 performed by the confd daemon, making them independent from jobs, and
200 thus faster to execute. This is selectable at configure time.
201 - The functionality for allocating multiple instances at once has been
202 overhauled and is now also available through :doc:`RAPI <rapi>`.
206 - Fix permissions on the confd query socket (Issue 477)
207 - Fix permissions on the job archive dir (Issue 498)
208 - Fix handling of an internal exception in replace-disks (Issue 472)
209 - Fix gnt-node info handling of shortened names (Issue 497)
210 - Fix gnt-instance grow-disk when wiping is enabled
211 - Documentation improvements, and support for newer pandoc
212 - Fix hspace honoring ipolicy for disks (Issue 484)
213 - Improve handling of the ``kvm_extra`` HV parameter
219 *(Released Fri, 24 May 2013)*
221 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
223 - Disks added through ``gnt-instance modify`` or created through
224 ``gnt-instance recreate-disks`` are wiped, if the
225 ``prealloc_wipe_disks`` flag is set.
226 - If wiping newly created disks fails, the disks are removed. Also,
227 partial failures in creating disks through ``gnt-instance modify``
228 triggers a cleanup of the partially-created disks.
229 - Removing the master IP address doesn't fail if the address has been
231 - Fix ownership of the OS log dir
232 - Workaround missing SO_PEERCRED constant (Issue 191)
238 *(Released Fri, 3 May 2013)*
240 This was the first release candidate of the 2.7 series. Since beta3:
242 - Fix kvm compatibility with qemu 1.4 (Issue 389)
243 - Documentation updates (admin guide, upgrade notes, install
244 instructions) (Issue 372)
245 - Fix gnt-group list nodes and instances count (Issue 436)
246 - Fix compilation without non-mandatory libraries (Issue 441)
247 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
248 - Make confd logging more verbose at INFO level (Issue 435)
249 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
250 - Fix failure path for instance storage type conversion (Issue 229)
251 - Update htools text backend documentation
252 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
253 - Disable inter-cluster instance move for file-based instances, because
254 it is dependant on instance export, which is not supported for
255 file-based instances. (Issue 414)
256 - Fix gnt-job crashes on non-ascii characters (Issue 427)
257 - Fix volume group checks on non-vm-capable nodes (Issue 432)
263 *(Released Mon, 22 Apr 2013)*
265 This was the third beta release of the 2.7 series. Since beta2:
267 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
268 - Fix data loss on wrong usage of ``gnt-instance move``
269 - Properly export errors in confd-based job queries
270 - Add ``users-setup`` tool
271 - Fix iallocator protocol to report 0 as a disk size for diskless
272 instances. This avoids hail breaking when a diskless instance is
274 - Fix job queue directory permission problem that made confd job queries
275 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
276 for access to archived jobs (Issue 406).
277 - Limit the sizes of networks supported by ``gnt-network`` to something
278 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
279 - Fix bugs in instance disk template conversion
280 - Fix GHC 7 compatibility
281 - Fix ``burnin`` install path (Issue 426).
282 - Allow very small disk grows (Issue 347).
283 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
284 that noded doesn't import masterd code (Issue 419).
285 - Make sure the default metavg at cluster init is the same as the vg, if
286 unspecified (Issue 358).
287 - Fix cleanup of partially created disks (part of Issue 416)
293 *(Released Tue, 2 Apr 2013)*
295 This was the second beta release of the 2.7 series. Since beta1:
297 - Networks no longer have a "type" slot, since this information was
298 unused in Ganeti: instead of it tags should be used.
299 - The rapi client now has a ``target_node`` option to MigrateInstance.
300 - Fix early exit return code for hbal (Issue 386).
301 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
302 - Fix ``rbd showmapped`` output parsing (Issue 312).
303 - Networks are now referenced indexed by UUID, rather than name. This
304 will require running cfgupgrade, from 2.7.0beta1, if networks are in
306 - The OS environment now includes network information.
307 - Deleting of a network is now disallowed if any instance nic is using
308 it, to prevent dangling references.
309 - External storage is now documented in man pages.
310 - The exclusive_storage flag can now only be set at nodegroup level.
311 - Hbal can now submit an explicit priority with its jobs.
312 - Many network related locking fixes.
313 - Bump up the required pylint version to 0.25.1.
314 - Fix the ``no_remember`` option in RAPI client.
315 - Many ipolicy related tests, qa, and fixes.
316 - Many documentation improvements and fixes.
317 - Fix building with ``--disable-file-storage``.
318 - Fix ``-q`` option in htools, which was broken if passed more than
320 - Some haskell/python interaction improvements and fixes.
321 - Fix iallocator in case of missing LVM storage.
322 - Fix confd config load in case of ``--no-lvm-storage``.
323 - The confd/query functionality is now mentioned in the security
330 *(Released Wed, 6 Feb 2013)*
332 This was the first beta release of the 2.7 series. All important changes
333 are listed in the latest 2.7 entry.
339 *(Released Fri, 21 Dec 2012)*
341 Important behaviour change: hbal won't rebalance anymore instances which
342 have the ``auto_balance`` attribute set to false. This was the intention
343 all along, but until now it only skipped those from the N+1 memory
344 reservation (DRBD-specific).
346 A significant number of bug fixes in this release:
348 - Fixed disk adoption interaction with ipolicy checks.
349 - Fixed networking issues when instances are started, stopped or
350 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
351 - Fixed the warning in cluster verify for shared storage instances not
353 - Fixed removal of storage directory on shared file storage (issue 262).
354 - Fixed validation of LVM volume group name in OpClusterSetParams
355 (``gnt-cluster modify``) (issue 285).
356 - Fixed runtime memory increases (``gnt-instance modify -m``).
357 - Fixed live migration under Xen's ``xl`` mode.
358 - Fixed ``gnt-instance console`` with ``xl``.
359 - Fixed building with newer Haskell compiler/libraries.
360 - Fixed PID file writing in Haskell daemons (confd); this prevents
361 restart issues if confd was launched manually (outside of
362 ``daemon-util``) while another copy of it was running
363 - Fixed a type error when doing live migrations with KVM (issue 297) and
364 the error messages for failing migrations have been improved.
365 - Fixed opcode validation for the out-of-band commands (``gnt-node
367 - Fixed a type error when unsetting OS hypervisor parameters (issue
368 311); now it's possible to unset all OS-specific hypervisor
370 - Fixed the ``dry-run`` mode for many operations: verification of
371 results was over-zealous but didn't take into account the ``dry-run``
372 operation, resulting in "wrong" failures.
373 - Fixed bash completion in ``gnt-job list`` when the job queue has
374 hundreds of entries; especially with older ``bash`` versions, this
375 results in significant CPU usage.
377 And lastly, a few other improvements have been made:
379 - Added option to force master-failover without voting (issue 282).
380 - Clarified error message on lock conflict (issue 287).
381 - Logging of newly submitted jobs has been improved (issue 290).
382 - Hostname checks have been made uniform between instance rename and
384 - The ``--submit`` option is now supported by ``gnt-debug delay``.
385 - Shutting down the master daemon by sending SIGTERM now stops it from
386 processing jobs waiting for locks; instead, those jobs will be started
387 once again after the master daemon is started the next time (issue
389 - Support for Xen's ``xl`` program has been improved (besides the fixes
391 - Reduced logging noise in the Haskell confd daemon (only show one log
392 entry for each config reload, instead of two).
393 - Several man page updates and typo fixes.
399 *(Released Fri, 12 Oct 2012)*
401 A small bugfix release. Among the bugs fixed:
403 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
404 made the command unusable.
405 - Commands that issue many jobs don't fail anymore just because some jobs
406 take so long that other jobs are archived.
407 - Failures during ``gnt-instance reinstall`` are reflected by the exit
409 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
411 - When ``always_failover`` is set, ``--allow-failover`` is not required
412 in migrate commands anymore.
413 - ``bash_completion`` works even if extglob is disabled.
414 - Fixed bug with locks that made failover for RDB-based instances fail.
415 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
416 a random node instead of one based on the allocator metric.
417 - Support for newer versions of pylint and pep8.
418 - Hail doesn't fail anymore when trying to add an instance of type
419 ``file``, ``sharedfile`` or ``rbd``.
420 - Added new Makefile target to rebuild the whole distribution, so that
421 all files are included.
427 *(Released Fri, 27 Jul 2012)*
430 .. attention:: The ``LUXI`` protocol has been made more consistent
431 regarding its handling of command arguments. This, however, leads to
432 incompatibility issues with previous versions. Please ensure that you
433 restart Ganeti daemons soon after the upgrade, otherwise most
434 ``LUXI`` calls (job submission, setting/resetting the drain flag,
435 pausing/resuming the watcher, cancelling and archiving jobs, querying
436 the cluster configuration) will fail.
445 The current ``admin_up`` field, which used to denote whether an instance
446 should be running or not, has been removed. Instead, ``admin_state`` is
447 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
449 The rational behind this is that an instance being “down” can have
452 - it could be down during a reboot
453 - it could be temporarily be down for a reinstall
454 - or it could be down because it is deprecated and kept just for its
457 The previous Boolean state was making it difficult to do capacity
458 calculations: should Ganeti reserve memory for a down instance? Now, the
459 tri-state field makes it clear:
461 - in ``up`` and ``down`` state, all resources are reserved for the
462 instance, and it can be at any time brought up if it is down
463 - in ``offline`` state, only disk space is reserved for it, but not
466 The field can have an extra use: since the transition between ``up`` and
467 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
468 transition between ``offline`` and ``down`` is done via ``gnt-instance
469 modify``, it is possible to given different rights to users. For
470 example, owners of an instance could be allowed to start/stop it, but
471 not transition it out of the offline state.
473 Instance policies and specs
474 +++++++++++++++++++++++++++
476 In previous Ganeti versions, an instance creation request was not
477 limited on the minimum size and on the maximum size just by the cluster
478 resources. As such, any policy could be implemented only in third-party
479 clients (RAPI clients, or shell wrappers over ``gnt-*``
480 tools). Furthermore, calculating cluster capacity via ``hspace`` again
481 required external input with regards to instance sizes.
483 In order to improve these workflows and to allow for example better
484 per-node group differentiation, we introduced instance specs, which
487 - minimum instance disk size, disk count, memory size, cpu count
488 - maximum values for the above metrics
489 - and “standard” values (used in ``hspace`` to calculate the standard
492 The minimum/maximum values can be also customised at node-group level,
493 for example allowing more powerful hardware to support bigger instance
496 Beside the instance specs, there are a few other settings belonging to
497 the instance policy framework. It is possible now to customise, per
498 cluster and node-group:
500 - the list of allowed disk templates
501 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
502 - the maximum ratio of instance to spindles (see below for more
503 information) for local storage
505 All these together should allow all tools that talk to Ganeti to know
506 what are the ranges of allowed values for instances and the
507 over-subscription that is allowed.
509 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
510 instance configuration, and the physical CPU configuration from the
511 node. For the spindle ratios however, we didn't track before these
512 values, so new parameters have been added:
514 - a new node parameter ``spindle_count``, defaults to 1, customisable at
515 node group or node level
516 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
518 Note that spindles in this context doesn't need to mean actual
519 mechanical hard-drives; it's just a relative number for both the node
520 I/O capacity and instance I/O consumption.
522 Instance migration behaviour
523 ++++++++++++++++++++++++++++
525 While live-migration is in general desirable over failover, it is
526 possible that for some workloads it is actually worse, due to the
527 variable time of the “suspend” phase during live migration.
529 To allow the tools to work consistently over such instances (without
530 having to hard-code instance names), a new backend parameter
531 ``always_failover`` has been added to control the migration/failover
532 behaviour. When set to True, all migration requests for an instance will
533 instead fall-back to failover.
535 Instance memory ballooning
536 ++++++++++++++++++++++++++
538 Initial support for memory ballooning has been added. The memory for an
539 instance is no longer fixed (backend parameter ``memory``), but instead
540 can vary between minimum and maximum values (backend parameters
541 ``minmem`` and ``maxmem``). Currently we only change an instance's
544 - live migrating or failing over and instance and the target node
545 doesn't have enough memory
546 - user requests changing the memory via ``gnt-instance modify
552 In order to control the use of specific CPUs by instance, support for
553 controlling CPU pinning has been added for the Xen, HVM and LXC
554 hypervisors. This is controlled by a new hypervisor parameter
555 ``cpu_mask``; details about possible values for this are in the
556 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
557 VCPU-to-CPU mapping) form will work well only when all nodes in your
558 cluster have the same amount of CPUs.
563 Another area in which Ganeti was not customisable were the parameters
564 used for storage configuration, e.g. how many stripes to use for LVM,
565 DRBD resync configuration, etc.
567 To improve this area, we've added disks parameters, which are
568 customisable at cluster and node group level, and which allow to
569 specify various parameters for disks (DRBD has the most parameters
570 currently), for example:
572 - DRBD resync algorithm and parameters (e.g. speed)
573 - the default VG for meta-data volumes for DRBD
574 - number of stripes for LVM (plain disk template)
577 These parameters can be modified via ``gnt-cluster modify -D …`` and
578 ``gnt-group modify -D …``, and are used at either instance creation (in
579 case of LVM stripes, for example) or at disk “activation” time
582 Rados block device support
583 ++++++++++++++++++++++++++
585 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
586 denoted by the ``rbd`` disk template type. This is considered
587 experimental, feedback is welcome. For details on configuring it, see
588 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
593 The existing master IP functionality works well only in simple setups (a
594 single network shared by all nodes); however, if nodes belong to
595 different networks, then the ``/32`` setup and lack of routing
596 information is not enough.
598 To allow the master IP to function well in more complex cases, the
599 system was reworked as follows:
601 - a master IP netmask setting has been added
602 - the master IP activation/turn-down code was moved from the node daemon
604 - whether to run the Ganeti-supplied master IP script or a user-supplied
605 on is a ``gnt-cluster init`` setting
607 Details about the location of the standard and custom setup scripts are
608 in the man page :manpage:`gnt-cluster(8)`; for information about the
609 setup script protocol, look at the Ganeti-supplied script.
614 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
617 It is now possible to use TLS-protected connections, and when renewing
618 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
619 it is now possible to specify spice or spice CA certificates. Also, it
620 is possible to configure a password for SPICE sessions via the
621 hypervisor parameter ``spice_password_file``.
623 There are also new parameters to control the compression and streaming
624 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
625 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
626 for the spice parameters.
628 Lastly, it is now possible to see the SPICE connection information via
629 ``gnt-instance console``.
634 A new tool (``tools/ovfconverter``) has been added that supports
635 conversion between Ganeti and the `Open Virtualization Format
636 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
639 This relies on the ``qemu-img`` tool to convert the disk formats, so the
640 actual compatibility with other virtualization solutions depends on it.
645 The configuration query daemon (``ganeti-confd``) is now optional, and
646 has been rewritten in Haskell; whether to use the daemon at all, use the
647 Python (default) or the Haskell version is selectable at configure time
648 via the ``--enable-confd`` parameter, which can take one of the
649 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
650 daemon will result in a smaller footprint; for larger systems, we
651 welcome feedback on the Haskell version which might become the default
654 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
655 daemon running. The Python version doesn't implement the new call.
658 User interface changes
659 ~~~~~~~~~~~~~~~~~~~~~~
661 We have replaced the ``--disks`` option of ``gnt-instance
662 replace-disks`` with a more flexible ``--disk`` option, which allows
663 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
664 disk size and mode can be changed upon recreation (via ``gnt-instance
665 recreate-disks``, which accepts the same ``--disk`` option).
667 As many people are used to a ``show`` command, we have added that as an
668 alias to ``info`` on all ``gnt-*`` commands.
670 The ``gnt-instance grow-disk`` command has a new mode in which it can
671 accept the target size of the disk, instead of the delta; this can be
672 more safe since two runs in absolute mode will be idempotent, and
673 sometimes it's also easier to specify the desired size directly.
675 Also the handling of instances with regard to offline secondaries has
676 been improved. Instance operations should not fail because one of it's
677 secondary nodes is offline, even though it's safe to proceed.
679 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
680 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
681 minors to instance name.
686 RAPI coverage has improved, with (for example) new resources for
687 recreate-disks, node power-cycle, etc.
692 There is partial support for ``xl`` in the Xen hypervisor; feedback is
695 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
696 whether to still support Python 2.4 or move to Python 2.6 as minimum
699 Support for Fedora has been slightly improved; the provided example
700 init.d script should work better on it and the INSTALL file should
701 document the needed dependencies.
706 The deprecated ``QueryLocks`` LUXI request has been removed. Use
707 ``Query(what=QR_LOCK, ...)`` instead.
709 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
710 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
711 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
712 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
713 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
715 RAPI client: ``CertificateError`` now derives from
716 ``GanetiApiError``. This should make it more easy to handle Ganeti
719 Deprecation warnings due to PyCrypto/paramiko import in
720 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
721 make sure to run an up-to-date paramiko version, if you use this tool.
723 The QA scripts now depend on Python 2.5 or above (the main code base
724 still works with Python 2.4).
726 The configuration file (``config.data``) is now written without
727 indentation for performance reasons; if you want to edit it, it can be
728 re-formatted via ``tools/fmtjson``.
730 A number of bugs has been fixed in the cluster merge tool.
732 ``x509`` certification verification (used in import-export) has been
733 changed to allow the same clock skew as permitted by the cluster
734 verification. This will remove some rare but hard to diagnose errors in
741 *(Released Thu, 19 Jul 2012)*
743 Very few changes from rc4 to the final release, only bugfixes:
745 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
746 instance, fix CDROM booting for KVM instances)
747 - fixed node group modification of node parameters
748 - fixed issue in LUClusterVerifyGroup with multi-group clusters
749 - fixed generation of bash completion to ensure a stable ordering
756 *(Released Fri, 13 Jul 2012)*
758 Third release candidate for 2.6. The following changes were done from
761 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
762 - Fixed an inconsistency in the LUXI protocol with the provided
763 arguments (NOT backwards compatible)
764 - Fixed a bug with node groups ipolicy where ``min`` was greater than
765 the cluster ``std`` value
766 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
767 easier instance debugging on nodes (requires ``hconfd`` to work)
773 *(Released Tue, 03 Jul 2012)*
775 Second release candidate for 2.6. The following changes were done from
778 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
780 - Fixed a RAPI regression on missing beparams/memory
781 - Fixed redistribution of files on offline nodes
782 - Added possibility to run activate-disks even though secondaries are
783 offline. With this change it relaxes also the strictness on some other
784 commands which use activate disks internally:
785 * ``gnt-instance start|reboot|rename|backup|export``
786 - Made it possible to remove safely an instance if its secondaries are
788 - Made it possible to reinstall even though secondaries are offline
794 *(Released Mon, 25 Jun 2012)*
796 First release candidate for 2.6. The following changes were done from
799 - Fixed bugs with disk parameters and ``rbd`` templates as well as
801 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
802 behaviour. It supports now the modify operation
803 - ``hcheck`` implemented to analyze cluster health and possibility of
804 improving health by rebalance
805 - ``hbal`` has been improved in dealing with split instances
811 *(Released Mon, 11 Jun 2012)*
813 Second beta release of 2.6. The following changes were done from beta2
816 - Fixed ``daemon-util`` with non-root user models
817 - Fixed creation of plain instances with ``--no-wait-for-sync``
818 - Fix wrong iv_names when running ``cfgupgrade``
819 - Export more information in RAPI group queries
820 - Fixed bug when changing instance network interfaces
821 - Extended burnin to do NIC changes
822 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
823 - Changed default for DRBD barriers
824 - Fixed DRBD error reporting for syncer rate
825 - Verify the options on disk parameters
827 And of course various fixes to documentation and improved unittests and
834 *(Released Wed, 23 May 2012)*
836 First beta release of 2.6. The following changes were done from beta1 to
839 - integrated patch for distributions without ``start-stop-daemon``
840 - adapted example init.d script to work on Fedora
841 - fixed log handling in Haskell daemons
842 - adapted checks in the watcher for pycurl linked against libnss
843 - add partial support for ``xl`` instead of ``xm`` for Xen
844 - fixed a type issue in cluster verification
845 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
848 Plus integrated fixes from the 2.5 branch:
850 - fixed ``kvm-ifup`` to use ``/bin/bash``
851 - fixed parallel build failures
852 - KVM live migration when using a custom keymap
858 *(Released Tue, 24 Jul 2012)*
860 A small bugfix release, with no new features:
862 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
863 different default shell (e.g. Debian, Ubuntu)
864 - fixed KVM startup and live migration with a custom keymap (fixes Issue
865 243 and Debian bug #650664)
866 - fixed compatibility with KVM versions that don't support multiple boot
867 devices (fixes Issue 230 and Debian bug #624256)
869 Additionally, a few fixes were done to the build system (fixed parallel
870 build failures) and to the unittests (fixed race condition in test for
871 FileID functions, and the default enable/disable mode for QA test is now
878 *(Released Fri, 11 May 2012)*
880 A small bugfix release.
882 The main issues solved are on the topic of compatibility with newer LVM
885 - fixed parsing of ``lv_attr`` field
886 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
887 the ``--force`` flag is needed
889 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
890 to accept kernel 3.x too (was hardcoded to 2.6.*).
892 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
893 of hook results and that also made display of other errors suboptimal
894 was fixed; the code behaves now like 2.4 and earlier.
896 Another change in 2.5, the cleanup of the OS scripts environment, is too
897 aggressive: it removed even the ``PATH`` variable, which requires the OS
898 scripts to *always* need to export it. Since this is a bit too strict,
899 we now export a minimal PATH, the same that we export for hooks.
901 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
902 integrated into this release.
904 Finally, a few other miscellaneous changes were done (no new features,
905 just small improvements):
907 - Fix ``gnt-group --help`` display
908 - Fix hardcoded Xen kernel path
909 - Fix grow-disk handling of invalid units
910 - Update synopsis for ``gnt-cluster repair-disk-sizes``
911 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
917 *(Released Thu, 12 Apr 2012)*
919 Incompatible/important changes and bugfixes
920 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
922 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
923 resource's ``ip_check`` parameter changed from ``True`` to ``False``
924 to match the underlying LUXI interface.
925 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
926 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
927 not maintain backwards-compatibility as the underlying operation
928 changed in an incompatible way. The RAPI client can talk to old
929 servers, but it needs to be told so as the return value changed.
930 - When creating file-based instances via RAPI, the ``file_driver``
931 parameter no longer defaults to ``loop`` and must be specified.
932 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
934 - Support for the undocumented and deprecated RAPI instance creation
935 request format version 0 has been dropped. Use version 1, supported
936 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
937 - Pyparsing 1.4.6 or above is required, see :doc:`installation
938 documentation <install>`.
939 - The "cluster-verify" hooks are now executed per group by the
940 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
941 if you just run ``gnt-cluster verify``, which generates one opcode per
943 - The environment as passed to the OS scripts is cleared, and thus no
944 environment variables defined in the node daemon's environment will be
945 inherited by the scripts.
946 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
948 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
949 support operations involving multiple node groups.
950 - Offline nodes are ignored when failing over an instance.
951 - Support for KVM version 1.0, which changed the version reporting format
953 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
955 - ``Makefile`` is now compatible with Automake 1.11.2
956 - Includes all bugfixes made in the 2.4 series
961 - The ganeti-htools project has been merged into the ganeti-core source
962 tree and will be built as part of Ganeti (see :doc:`install-quick`).
963 - Implemented support for :doc:`shared storage <design-shared-storage>`.
964 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
965 GPT-style partition tables (requires `parted
966 <http://www.gnu.org/s/parted/>`_).
967 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
968 - Allowed adding tags on instance creation.
969 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
970 - Allow instances to be started in a paused state, enabling the user to
971 see the complete console output on boot using the console.
972 - Added new hypervisor flag to control default reboot behaviour
973 (``reboot_behavior``).
974 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
975 - Improved out-of-band management support:
977 - Added ``gnt-node health`` command reporting the health status of
979 - Added ``gnt-node power`` command to manage power status of nodes.
980 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
982 - Instance migration can fall back to failover if instance is not
984 - Filters can be used when listing nodes, instances, groups and locks;
985 see :manpage:`ganeti(7)` manpage.
986 - Added post-execution status as variables to :doc:`hooks <hooks>`
988 - Instance tags are exported/imported together with the instance.
989 - When given an explicit job ID, ``gnt-job info`` will work for archived
991 - Jobs can define dependencies on other jobs (not yet supported via
992 RAPI or command line, but used by internal commands and usable via
995 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
998 - Instance failover is now available as a RAPI resource
999 (``/2/instances/[instance_name]/failover``).
1000 - ``gnt-instance info`` defaults to static information if primary node
1002 - Opcodes have a new ``comment`` attribute.
1003 - Added basic SPICE support to KVM hypervisor.
1004 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1006 Node group improvements
1007 ~~~~~~~~~~~~~~~~~~~~~~~
1009 - ``gnt-cluster verify`` has been modified to check groups separately,
1010 thereby improving performance.
1011 - Node group support has been added to ``gnt-cluster verify-disks``,
1012 which now operates per node group.
1013 - Watcher has been changed to work better with node groups.
1015 - One process and state file per node group.
1016 - Slow watcher in one group doesn't block other group's watcher.
1018 - Added new command, ``gnt-group evacuate``, to move all instances in a
1019 node group to other groups.
1020 - Added ``gnt-instance change-group`` to move an instance to another
1022 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1023 per-group operations.
1024 - Node groups can be tagged.
1025 - Some operations switch from an exclusive to a shared lock as soon as
1027 - Instance's primary and secondary nodes' groups are now available as
1028 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1029 and ``snodes.group.uuid``).
1034 - Numerous updates to documentation and manpages.
1036 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1038 - Some opcode/job results are now also documented, see :doc:`RAPI
1041 - A lockset's internal lock is now also visible in lock monitor.
1042 - Log messages from job queue workers now contain information about the
1043 opcode they're processing.
1044 - ``gnt-instance console`` no longer requires the instance lock.
1045 - A short delay when waiting for job changes reduces the number of LUXI
1046 requests significantly.
1047 - DRBD metadata volumes are overwritten with zeros during disk creation.
1048 - Out-of-band commands no longer acquire the cluster lock in exclusive
1050 - ``devel/upload`` now uses correct permissions for directories.
1056 *(Released Fri, 23 Mar 2012)*
1058 This was the sixth release candidate of the 2.5 series.
1064 *(Released Mon, 9 Jan 2012)*
1066 This was the fifth release candidate of the 2.5 series.
1072 *(Released Thu, 27 Oct 2011)*
1074 This was the fourth release candidate of the 2.5 series.
1080 *(Released Wed, 26 Oct 2011)*
1082 This was the third release candidate of the 2.5 series.
1088 *(Released Tue, 18 Oct 2011)*
1090 This was the second release candidate of the 2.5 series.
1096 *(Released Tue, 4 Oct 2011)*
1098 This was the first release candidate of the 2.5 series.
1104 *(Released Wed, 31 Aug 2011)*
1106 This was the third beta release of the 2.5 series.
1112 *(Released Mon, 22 Aug 2011)*
1114 This was the second beta release of the 2.5 series.
1120 *(Released Fri, 12 Aug 2011)*
1122 This was the first beta release of the 2.5 series.
1128 *(Released Thu, 27 Oct 2011)*
1130 - Fixed bug when parsing command line parameter values ending in
1132 - Fixed assertion error after unclean master shutdown
1133 - Disable HTTP client pool for RPC, significantly reducing memory usage
1135 - Fixed queue archive creation with wrong permissions
1141 *(Released Tue, 23 Aug 2011)*
1145 - Fixed documentation for importing with ``--src-dir`` option
1146 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1147 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1153 *(Released Fri, 5 Aug 2011)*
1155 Many bug-fixes and a few small features:
1157 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1158 issues when you tried to add an instance with two MAC addresses in one
1160 - KVM: fixed per-instance stored UID value
1161 - KVM: configure bridged NICs at migration start
1162 - KVM: Fix a bug where instance will not start with never KVM versions
1164 - Added OS search path to ``gnt-cluster info``
1165 - Fixed an issue with ``file_storage_dir`` where you were forced to
1166 provide an absolute path, but the documentation states it is a
1167 relative path, the documentation was right
1168 - Added a new parameter to instance stop/start called ``--no-remember``
1169 that will make the state change to not be remembered
1170 - Implemented ``no_remember`` at RAPI level
1171 - Improved the documentation
1172 - Node evacuation: don't call IAllocator if node is already empty
1173 - Fixed bug in DRBD8 replace disks on current nodes
1174 - Fixed bug in recreate-disks for DRBD instances
1175 - Moved assertion checking locks in ``gnt-instance replace-disks``
1176 causing it to abort with not owning the right locks for some situation
1177 - Job queue: Fixed potential race condition when cancelling queued jobs
1178 - Fixed off-by-one bug in job serial generation
1179 - ``gnt-node volumes``: Fix instance names
1180 - Fixed aliases in bash completion
1181 - Fixed a bug in reopening log files after being sent a SIGHUP
1182 - Added a flag to burnin to allow specifying VCPU count
1183 - Bugfixes to non-root Ganeti configuration
1189 *(Released Thu, 12 May 2011)*
1191 Many bug-fixes and a few new small features:
1193 - Fixed a bug related to log opening failures
1194 - Fixed a bug in instance listing with orphan instances
1195 - Fixed a bug which prevented resetting the cluster-level node parameter
1196 ``oob_program`` to the default
1197 - Many fixes related to the ``cluster-merge`` tool
1198 - Fixed a race condition in the lock monitor, which caused failures
1199 during (at least) creation of many instances in parallel
1200 - Improved output for gnt-job info
1201 - Removed the quiet flag on some ssh calls which prevented debugging
1203 - Improved the N+1 failure messages in cluster verify by actually
1204 showing the memory values (needed and available)
1205 - Increased lock attempt timeouts so that when executing long operations
1206 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1207 too early and thus prevent the use of the 'fair' mechanism
1208 - Changed instance query data (``gnt-instance info``) to not acquire
1209 locks unless needed, thus allowing its use on locked instance if only
1210 static information is asked for
1211 - Improved behaviour with filesystems that do not support rename on an
1213 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1214 kept locks on all nodes during the wipe, which is unneeded
1215 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1216 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1217 10GiB) which caused the wipe to fail right at the end in some cases
1218 - Fixed master IP activation when doing master failover with no-voting
1219 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1220 the master node itself
1221 - Fixed potential data-loss in under disk full conditions, where Ganeti
1222 wouldn't check correctly the return code and would consider
1223 partially-written files 'correct'
1224 - Fixed bug related to multiple VGs and DRBD disk replacing
1225 - Added new disk parameter ``metavg`` that allows placement of the meta
1226 device for DRBD in a different volume group
1227 - Fixed error handling in the node daemon when the system libc doesn't
1228 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1229 - Fixed lock release during replace-disks, which kept cluster-wide locks
1230 when doing disk replaces with an iallocator script
1231 - Added check for missing bridges in cluster verify
1232 - Handle EPIPE errors while writing to the terminal better, so that
1233 piping the output to e.g. ``less`` doesn't cause a backtrace
1234 - Fixed rare case where a ^C during Luxi calls could have been
1235 interpreted as server errors, instead of simply terminating
1236 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1238 - Added a few more parameters to the KVM hypervisor, allowing a second
1239 CDROM, custom disk type for CDROMs and a floppy image
1240 - Removed redundant message in instance rename when the name is given
1242 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1243 disks on new nodes, allowing recreation when the original instance
1244 nodes are completely gone
1245 - Added option when converting disk templates to DRBD to skip waiting
1246 for the resync, in order to make the instance available sooner
1247 - Added two new variables to the OS scripts environment (containing the
1249 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1250 to allow use of ``pvgrub`` as bootloader
1251 - Changed the instance memory modifications to only check out-of-memory
1252 conditions on memory increases, and turned the secondary node warnings
1253 into errors (they can still be overridden via ``--force``)
1254 - Fixed the handling of a corner case when the Python installation gets
1255 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1256 to execute a command that doesn't exist
1257 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1258 node of the instance returned failures during instance shutdown; this
1259 adds the option ``--ignore-consistency`` to gnt-instance move
1261 And as usual, various improvements to the error messages, documentation
1268 *(Released Wed, 09 Mar 2011)*
1270 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1271 the RAPI users file if run twice (even with ``--dry-run``).
1273 The release fixes that bug (nothing else changed).
1279 *(Released Mon, 07 Mar 2011)*
1281 Final 2.4.0 release. Just a few small fixes:
1283 - Fixed RAPI node evacuate
1284 - Fixed the kvm-ifup script
1285 - Fixed internal error handling for special job cases
1286 - Updated man page to specify the escaping feature for options
1292 *(Released Mon, 28 Feb 2011)*
1294 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1295 that this feature wiped the disks of the wrong instance, leading to loss
1300 - Fixed title of query field containing instance name
1301 - Expanded the glossary in the documentation
1302 - Fixed one unittest (internal issue)
1308 *(Released Mon, 21 Feb 2011)*
1310 A number of bug fixes plus just a couple functionality changes.
1312 On the user-visible side, the ``gnt-* list`` command output has changed
1313 with respect to "special" field states. The current rc1 style of display
1314 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1315 the default output mode special fields states are displayed as follows:
1317 - Offline resource: ``*``
1318 - Unavailable/not applicable: ``-``
1319 - Data missing (RPC failure): ``?``
1320 - Unknown field: ``??``
1322 Another user-visible change is the addition of ``--force-join`` to
1327 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1328 - Fixed regression in RAPI/instance reinstall where all parameters were
1329 required (instead of optional)
1330 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1331 - Fixed iallocator usage (offline nodes were not considered offline)
1332 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1333 - Fixed hypervisor and OS parameter validation with respect to
1334 non-vm_capable nodes
1335 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1337 - Fixed ``tools/listrunner`` with respect to agent-based usage
1343 *(Released Fri, 4 Feb 2011)*
1345 Many changes and fixes since the beta1 release. While there were some
1346 internal changes, the code has been mostly stabilised for the RC
1349 Note: the dumb allocator was removed in this release, as it was not kept
1350 up-to-date with the IAllocator protocol changes. It is recommended to
1351 use the ``hail`` command from the ganeti-htools package.
1353 Note: the 2.4 and up versions of Ganeti are not compatible with the
1354 0.2.x branch of ganeti-htools. You need to upgrade to
1355 ganeti-htools-0.3.0 (or later).
1357 Regressions fixed from 2.3
1358 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1360 - Fixed the ``gnt-cluster verify-disks`` command
1361 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1363 - Fixed disk adoption breakage
1364 - Fixed wrong headers in instance listing for field aliases
1369 - Fixed corner case in KVM handling of NICs
1370 - Fixed many cases of wrong handling of non-vm_capable nodes
1371 - Fixed a bug where a missing instance symlink was not possible to
1372 recreate with any ``gnt-*`` command (now ``gnt-instance
1373 activate-disks`` does it)
1374 - Fixed the volume group name as reported by ``gnt-cluster
1376 - Increased timeouts for the import-export code, hopefully leading to
1377 fewer aborts due network or instance timeouts
1378 - Fixed bug in ``gnt-node list-storage``
1379 - Fixed bug where not all daemons were started on cluster
1380 initialisation, but only at the first watcher run
1381 - Fixed many bugs in the OOB implementation
1382 - Fixed watcher behaviour in presence of instances with offline
1384 - Fixed instance list output for instances running on the wrong node
1385 - a few fixes to the cluster-merge tool, but it still cannot merge
1386 multi-node groups (currently it is not recommended to use this tool)
1392 - Improved network configuration for the KVM hypervisor
1393 - Added e1000 as a supported NIC for Xen-HVM
1394 - Improved the lvmstrap tool to also be able to use partitions, as
1395 opposed to full disks
1396 - Improved speed of disk wiping (the cluster parameter
1397 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1398 of instance creations
1399 - Added documentation for the OS parameters
1400 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1401 hypervisor is not responding
1402 - Added display of blacklisted and hidden OS information in
1403 ``gnt-cluster info``
1404 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1405 NIC and node parameters, which might create problems with currently
1406 invalid (but undetected) configuration files, but prevents validation
1407 failures when unrelated parameters are modified
1408 - Changed cluster initialisation to wait for the master daemon to become
1410 - Expanded the RAPI interface:
1412 - Added config redistribution resource
1413 - Added activation/deactivation of instance disks
1414 - Added export of console information
1416 - Implemented log file reopening on SIGHUP, which allows using
1417 logrotate(8) for the Ganeti log files
1418 - Added a basic OOB helper script as an example
1424 *(Released Fri, 14 Jan 2011)*
1429 - Fixed timezone issues when formatting timestamps
1430 - Added support for node groups, available via ``gnt-group`` and other
1432 - Added out-of-band framework and management, see :doc:`design
1433 document <design-oob>`
1434 - Removed support for roman numbers from ``gnt-node list`` and
1435 ``gnt-instance list``.
1436 - Allowed modification of master network interface via ``gnt-cluster
1437 modify --master-netdev``
1438 - Accept offline secondaries while shutting down instance disks
1439 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1440 - Added support for multiple LVM volume groups
1441 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1443 - Added commands to list available fields:
1445 - ``gnt-node list-fields``
1446 - ``gnt-group list-fields``
1447 - ``gnt-instance list-fields``
1449 - Updated documentation and man pages
1454 - Moved ``rapi_users`` file into separate directory, now named
1455 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1457 - Added new tool for running commands on many machines,
1458 ``tools/ganeti-listrunner``
1459 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1460 improving the ``gnt-instance console`` output
1461 - Allowed customisation of disk index separator at ``configure`` time
1462 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1463 - Added support for non-partitioned md disks in ``lvmstrap``
1464 - Added script to gracefully power off KVM instances
1465 - Split ``utils`` module into smaller parts
1466 - Changed query operations to return more detailed information, e.g.
1467 whether an information is unavailable due to an offline node. To use
1468 this new functionality, the LUXI call ``Query`` must be used. Field
1469 information is now stored by the master daemon and can be retrieved
1470 using ``QueryFields``. Instances, nodes and groups can also be queried
1471 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1472 exposed via RAPI). The following commands make use of this
1473 infrastructure change:
1475 - ``gnt-group list``
1476 - ``gnt-group list-fields``
1478 - ``gnt-node list-fields``
1479 - ``gnt-instance list``
1480 - ``gnt-instance list-fields``
1481 - ``gnt-debug locks``
1486 - New RAPI resources (see :doc:`rapi`):
1490 - ``/2/groups/[group_name]``
1491 - ``/2/groups/[group_name]/assign-nodes``
1492 - ``/2/groups/[group_name]/modify``
1493 - ``/2/groups/[group_name]/rename``
1494 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1498 - Implemented ``no_install`` for instance creation
1499 - Implemented OS parameters for instance reinstallation, allowing
1500 use of special settings on reinstallation (e.g. for preserving data)
1505 - Added IPv6 support in import/export
1506 - Pause DRBD synchronization while wiping disks on instance creation
1507 - Updated unittests and QA scripts
1508 - Improved network parameters passed to KVM
1509 - Converted man pages from docbook to reStructuredText
1515 *(Released Mon, 20 Dec 2010)*
1517 Released version 2.3.1~rc1 without any changes.
1523 *(Released Wed, 1 Dec 2010)*
1525 - impexpd: Disable OpenSSL compression in socat if possible (backport
1526 from master, commit e90739d625b, see :doc:`installation guide
1527 <install-quick>` for details)
1528 - Changed unittest coverage report to exclude test scripts
1529 - Added script to check version format
1535 *(Released Wed, 1 Dec 2010)*
1537 Released version 2.3.0~rc1 without any changes.
1543 *(Released Fri, 19 Nov 2010)*
1545 A number of bugfixes and documentation updates:
1547 - Update ganeti-os-interface documentation
1548 - Fixed a bug related to duplicate MACs or similar items which should be
1550 - Fix breakage in OS state modify
1551 - Reinstall instance: disallow offline secondaries (fixes bug related to
1552 OS changing but reinstall failing)
1553 - plus all the other fixes between 2.2.1 and 2.2.2
1559 *(Released Tue, 2 Nov 2010)*
1561 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1562 - Fixed master failover race with watcher
1563 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1565 - Accept previously stopped instance for export with instance removal
1566 - Simplify and extend the environment variables for instance OS scripts
1567 - Added new node flags, ``master_capable`` and ``vm_capable``
1568 - Added optional instance disk wiping prior during allocation. This is a
1569 cluster-wide option and can be set/modified using
1570 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1571 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1572 :doc:`install-quick`
1573 - Added a new watcher option (``--ignore-pause``)
1574 - Added option to ignore offline node on instance start/stop
1575 (``--ignore-offline``)
1576 - Allow overriding OS parameters with ``gnt-instance reinstall``
1577 - Added ability to change node's secondary IP address using ``gnt-node
1579 - Implemented privilege separation for all daemons except
1580 ``ganeti-noded``, see ``configure`` options
1581 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1583 - Implemented job priorities (see ``ganeti(7)`` manpage)
1584 - Ignore failures while shutting down instances during failover from
1586 - Exit daemon's bootstrap process only once daemon is ready
1587 - Export more information via ``LUInstanceQuery``/remote API
1588 - Improved documentation, QA and unittests
1589 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1590 restart if the file was created or changed
1591 - Added LUXI protocol version sent with each request and response,
1592 allowing detection of server/client mismatches
1593 - Moved the Python scripts among gnt-* and ganeti-* into modules
1594 - Moved all code related to setting up SSH to an external script,
1596 - Infrastructure changes for node group support in future versions
1602 *(Released Fri, 19 Nov 2010)*
1604 A few small bugs fixed, and some improvements to the build system:
1606 - Fix documentation regarding conversion to drbd
1607 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1609 - Fix error handling in node modify with multiple changes
1610 - Allow remote imports without checked names
1616 *(Released Tue, 19 Oct 2010)*
1618 - Disable SSL session ID cache in RPC client
1624 *(Released Thu, 14 Oct 2010)*
1626 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1627 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1629 - Fix problems with interaction between Curl and Python's HTTP server,
1630 resulting in increased speed in many RPC calls
1631 - Improve our release script to prevent breakage with older aclocal and
1638 *(Released Thu, 7 Oct 2010)*
1640 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1641 value retrieved from master
1642 - Added support for blacklisted or hidden OS definitions
1643 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1644 - Added support for -mem-path in KVM hypervisor abstraction layer
1645 - Allow overriding instance parameters in tool for inter-cluster
1646 instance moves (``tools/move-instance``)
1647 - Improved opcode summaries (e.g. in ``gnt-job list``)
1648 - Improve consistency of OS listing by sorting it
1649 - Documentation updates
1655 *(Released Fri, 8 Oct 2010)*
1657 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1663 *(Released Mon, 4 Oct 2010)*
1665 - Fixed regression in ``gnt-instance rename``
1671 *(Released Wed, 22 Sep 2010)*
1673 - Fixed OS_VARIANT variable for OS scripts
1674 - Fixed cluster tag operations via RAPI
1675 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1676 - Disabled RAPI CA checks in watcher
1682 *(Released Mon, 23 Aug 2010)*
1684 - Support DRBD versions of the format "a.b.c.d"
1686 - Re-introduce support for usage from multiple threads in RAPI client
1687 - Instance renames and modify via RAPI
1688 - Work around race condition between processing and archival in job
1690 - Mark opcodes following failed one as failed, too
1691 - Job field ``lock_status`` was removed due to difficulties making it
1692 work with the changed job queue in Ganeti 2.2; a better way to monitor
1693 locks is expected for a later 2.2.x release
1694 - Fixed dry-run behaviour with many commands
1695 - Support ``ssh-agent`` again when adding nodes
1696 - Many additional bugfixes
1702 *(Released Fri, 30 Jul 2010)*
1704 Important change: the internal RPC mechanism between Ganeti nodes has
1705 changed from using a home-grown http library (based on the Python base
1706 libraries) to use the PycURL library. This requires that PycURL is
1707 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1708 against GnuTLS by default. cURL's support for GnuTLS had known issues
1709 before cURL 7.21.0 and we recommend using the latest cURL release or
1710 linking against OpenSSL. Most other distributions already link PycURL
1711 and cURL against OpenSSL. The command::
1713 python -c 'import pycurl; print pycurl.version'
1715 can be used to determine the libraries PycURL and cURL are linked
1718 Other significant changes:
1720 - Rewrote much of the internals of the job queue, in order to achieve
1721 better parallelism; this decouples job query operations from the job
1722 processing, and it should allow much nicer behaviour of the master
1723 daemon under load, and it also has uncovered some long-standing bugs
1724 related to the job serialisation (now fixed)
1725 - Added a default iallocator setting to the cluster parameters,
1726 eliminating the need to always pass nodes or an iallocator for
1727 operations that require selection of new node(s)
1728 - Added experimental support for the LXC virtualization method
1729 - Added support for OS parameters, which allows the installation of
1730 instances to pass parameter to OS scripts in order to customise the
1732 - Added a hypervisor parameter controlling the migration type (live or
1733 non-live), since hypervisors have various levels of reliability; this
1734 has renamed the 'live' parameter to 'mode'
1735 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1736 logical volumes, meaning that cluster verify will ignore them and not
1737 flag their presence as errors
1738 - The watcher will now reset the error count for failed instances after
1739 8 hours, thus allowing self-healing if the problem that caused the
1740 instances to be down/fail to start has cleared in the meantime
1741 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1742 check for, and warn, if the drbd module parameter ``usermode_helper``
1743 is not consistent with the cluster-wide setting; this is needed to
1744 make diagnose easier of failed drbd creations
1745 - Started adding base IPv6 support, but this is not yet
1746 enabled/available for use
1747 - Rename operations (cluster, instance) will now return the new name,
1748 which is especially useful if a short name was passed in
1749 - Added support for instance migration in RAPI
1750 - Added a tool to pre-configure nodes for the SSH setup, before joining
1751 them to the cluster; this will allow in the future a simplified model
1752 for node joining (but not yet fully enabled in 2.2); this needs the
1753 paramiko python library
1754 - Fixed handling of name-resolving errors
1755 - Fixed consistency of job results on the error path
1756 - Fixed master-failover race condition when executed multiple times in
1758 - Fixed many bugs related to the job queue (mostly introduced during the
1759 2.2 development cycle, so not all are impacting 2.1)
1760 - Fixed instance migration with missing disk symlinks
1761 - Fixed handling of unknown jobs in ``gnt-job archive``
1762 - And many other small fixes/improvements
1766 - Enhanced both the unittest and the QA coverage
1767 - Switched the opcode validation to a generic model, and extended the
1768 validation to all opcode parameters
1769 - Changed more parts of the code that write shell scripts to use the
1771 - Switched the master daemon to use the asyncore library for the Luxi
1778 *(Released Thu, 17 Jun 2010)*
1780 - Added tool (``move-instance``) and infrastructure to move instances
1781 between separate clusters (see :doc:`separate documentation
1782 <move-instance>` and :doc:`design document <design-2.2>`)
1783 - Added per-request RPC timeout
1784 - RAPI now requires a Content-Type header for requests with a body (e.g.
1785 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1786 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1787 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1789 - Implemented initial support for running Ganeti daemons as separate
1790 users, see configure-time flags ``--with-user-prefix`` and
1791 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1793 - Instances can be removed after export (``gnt-backup export
1794 --remove-instance``)
1795 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1796 key (instead of 1024 bit)
1797 - Added new cluster configuration file for cluster domain secret
1798 - Import/export now use SSL instead of SSH
1799 - Added support for showing estimated time when exporting an instance,
1800 see the ``ganeti-os-interface(7)`` manpage and look for
1807 *(Released Tue, 16 Nov 2010)*
1809 Some more bugfixes. Unless critical bugs occur, this will be the last
1812 - Fix case of MAC special-values
1813 - Fix mac checker regex
1814 - backend: Fix typo causing "out of range" error
1815 - Add missing --units in gnt-instance list man page
1821 *(Released Tue, 24 Aug 2010)*
1824 - Don't ignore secondary node silently on non-mirrored disk templates
1826 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1827 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1828 - Properly document the usb_mouse parameter
1829 - Fix path in ganeti-rapi(8) (issue 116)
1830 - Adjust error message when the ganeti user's .ssh directory is
1832 - Add same-node-check when changing the disk template to drbd
1838 *(Released Fri, 16 Jul 2010)*
1841 - Add an option to only select some reboot types during qa/burnin.
1842 (on some hypervisors consequent reboots are not supported)
1843 - Fix infrequent race condition in master failover. Sometimes the old
1844 master ip address would be still detected as up for a short time
1845 after it was removed, causing failover to fail.
1846 - Decrease mlockall warnings when the ctypes module is missing. On
1847 Python 2.4 we support running even if no ctypes module is installed,
1848 but we were too verbose about this issue.
1849 - Fix building on old distributions, on which man doesn't have a
1851 - Fix RAPI not to ignore the MAC address on instance creation
1852 - Implement the old instance creation format in the RAPI client.
1858 *(Released Thu, 01 Jul 2010)*
1860 A small bugfix release:
1861 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1862 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1863 a non-existing option
1864 - Fix instance create: the --force-variant option was ignored
1865 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1866 - Fix modify node storage with non-FQDN arguments
1867 - Fix RAPI client to authenticate under Python 2.6 when used
1868 for more than 5 requests needing authentication
1869 - Fix gnt-instance modify -t (storage) giving a wrong error message
1870 when converting a non-shutdown drbd instance to plain
1876 *(Released Fri, 18 Jun 2010)*
1878 A small bugfix release:
1880 - Fix live migration of KVM instances started with older Ganeti
1881 versions which had fewer hypervisor parameters
1882 - Fix gnt-instance grow-disk on down instances
1883 - Fix an error-reporting bug during instance migration
1884 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1885 silently ignoring broken ones
1886 - Fix an RPC error reporting bug affecting, for example, RAPI client
1888 - Fix bug triggered by different API version os-es on different nodes
1889 - Fix a bug in instance startup with custom hvparams: OS level
1890 parameters would fail to be applied.
1891 - Fix the RAPI client under Python 2.6 (but more work is needed to
1892 make it work completely well with OpenSSL)
1893 - Fix handling of errors when resolving names from DNS
1899 *(Released Thu, 3 Jun 2010)*
1901 A medium sized development cycle. Some new features, and some
1902 fixes/small improvements/cleanups.
1904 Significant features
1905 ~~~~~~~~~~~~~~~~~~~~
1907 The node deamon now tries to mlock itself into memory, unless the
1908 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1909 its logs, and falls back to console logging. This allows emergency
1910 features such as ``gnt-node powercycle`` to work even in the event of a
1911 broken node disk (tested offlining the disk hosting the node's
1912 filesystem and dropping its memory caches; don't try this at home)
1914 KVM: add vhost-net acceleration support. It can be tested with a new
1915 enough version of the kernel and of qemu-kvm.
1917 KVM: Add instance chrooting feature. If you use privilege dropping for
1918 your VMs you can also now force them to chroot to an empty directory,
1919 before starting the emulated guest.
1921 KVM: Add maximum migration bandwith and maximum downtime tweaking
1922 support (requires a new-enough version of qemu-kvm).
1924 Cluster verify will now warn if the master node doesn't have the master
1925 ip configured on it.
1927 Add a new (incompatible) instance creation request format to RAPI which
1928 supports all parameters (previously only a subset was supported, and it
1929 wasn't possible to extend the old format to accomodate all the new
1930 features. The old format is still supported, and a client can check for
1931 this feature, before using it, by checking for its presence in the
1932 ``features`` RAPI resource.
1934 Now with ancient latin support. Try it passing the ``--roman`` option to
1935 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1936 (requires the python-roman module to be installed, in order to work).
1941 As usual many internal code refactorings, documentation updates, and
1944 - Lots of improvements and cleanups to the experimental Remote API
1945 (RAPI) client library.
1946 - A new unit test suite for the core daemon libraries.
1947 - A fix to creating missing directories makes sure the umask is not
1948 applied anymore. This enforces the same directory permissions
1950 - Better handling terminating daemons with ctrl+c (used when running
1951 them in debugging mode).
1952 - Fix a race condition in live migrating a KVM instance, when stat()
1953 on the old proc status file returned EINVAL, which is an unexpected
1955 - Fixed manpage checking with newer man and utf-8 charachters. But now
1956 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1962 *(Released Fri, 7 May 2010)*
1964 Fix a bug which prevented untagged KVM instances from starting.
1970 *(Released Fri, 7 May 2010)*
1972 Another release with a long development cycle, during which many
1973 different features were added.
1975 Significant features
1976 ~~~~~~~~~~~~~~~~~~~~
1978 The KVM hypervisor now can run the individual instances as non-root, to
1979 reduce the impact of a VM being hijacked due to bugs in the
1980 hypervisor. It is possible to run all instances as a single (non-root)
1981 user, to manually specify a user for each instance, or to dynamically
1982 allocate a user out of a cluster-wide pool to each instance, with the
1983 guarantee that no two instances will run under the same user ID on any
1986 An experimental RAPI client library, that can be used standalone
1987 (without the other Ganeti libraries), is provided in the source tree as
1988 ``lib/rapi/client.py``. Note this client might change its interface in
1989 the future, as we iterate on its capabilities.
1991 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1992 replace the cluster's certificates and crypto keys. This might help in
1993 case they have been compromised, or have simply expired.
1995 A new disk option for instance creation has been added that allows one
1996 to "adopt" currently existing logical volumes, with data
1997 preservation. This should allow easier migration to Ganeti from
1998 unmanaged (or managed via other software) instances.
2000 Another disk improvement is the possibility to convert between redundant
2001 (DRBD) and plain (LVM) disk configuration for an instance. This should
2002 allow better scalability (starting with one node and growing the
2003 cluster, or shrinking a two-node cluster to one node).
2005 A new feature that could help with automated node failovers has been
2006 implemented: if a node sees itself as offline (by querying the master
2007 candidates), it will try to shutdown (hard) all instances and any active
2008 DRBD devices. This reduces the risk of duplicate instances if an
2009 external script automatically failovers the instances on such nodes. To
2010 enable this, the cluster parameter ``maintain_node_health`` should be
2011 enabled; in the future this option (per the name) will enable other
2012 automatic maintenance features.
2014 Instance export/import now will reuse the original instance
2015 specifications for all parameters; that means exporting an instance,
2016 deleting it and the importing it back should give an almost identical
2017 instance. Note that the default import behaviour has changed from
2018 before, where it created only one NIC; now it recreates the original
2021 Cluster verify has added a few new checks: SSL certificates validity,
2022 /etc/hosts consistency across the cluster, etc.
2027 As usual, many internal changes were done, documentation fixes,
2030 - Fixed cluster initialization with disabled cluster storage (regression
2031 introduced in 2.1.1)
2032 - File-based storage supports growing the disks
2033 - Fixed behaviour of node role changes
2034 - Fixed cluster verify for some corner cases, plus a general rewrite of
2035 cluster verify to allow future extension with more checks
2036 - Fixed log spamming by watcher and node daemon (regression introduced
2038 - Fixed possible validation issues when changing the list of enabled
2040 - Fixed cleanup of /etc/hosts during node removal
2041 - Fixed RAPI response for invalid methods
2042 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2043 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2044 from ide disks, etc.)
2045 - Allow OS changes without re-installation (to record a changed OS
2046 outside of Ganeti, or to allow OS renames)
2047 - Allow instance creation without OS installation (useful for example if
2048 the OS will be installed manually, or restored from a backup not in
2050 - Implemented option to make cluster ``copyfile`` use the replication
2052 - Added list of enabled hypervisors to ssconf (possibly useful for
2054 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2056 - A partial form of node re-IP is possible via node readd, which now
2057 allows changed node primary IP
2058 - Command line utilities now show an informational message if the job is
2060 - The logs of the master daemon now show the PID/UID/GID of the
2067 *(Released Fri, 12 Mar 2010)*
2069 During the 2.1.0 long release candidate cycle, a lot of improvements and
2070 changes have accumulated with were released later as 2.1.1.
2075 The node evacuate command (``gnt-node evacuate``) was significantly
2076 rewritten, and as such the IAllocator protocol was changed - a new
2077 request type has been added. This unfortunate change during a stable
2078 series is designed to improve performance of node evacuations; on
2079 clusters with more than about five nodes and which are well-balanced,
2080 evacuation should proceed in parallel for all instances of the node
2081 being evacuated. As such, any existing IAllocator scripts need to be
2082 updated, otherwise the above command will fail due to the unknown
2083 request. The provided "dumb" allocator has not been updated; but the
2084 ganeti-htools package supports the new protocol since version 0.2.4.
2086 Another important change is increased validation of node and instance
2087 names. This might create problems in special cases, if invalid host
2088 names are being used.
2090 Also, a new layer of hypervisor parameters has been added, that sits at
2091 OS level between the cluster defaults and the instance ones. This allows
2092 customisation of virtualization parameters depending on the installed
2093 OS. For example instances with OS 'X' may have a different KVM kernel
2094 (or any other parameter) than the cluster defaults. This is intended to
2095 help managing a multiple OSes on the same cluster, without manual
2096 modification of each instance's parameters.
2098 A tool for merging clusters, ``cluster-merge``, has been added in the
2099 tools sub-directory.
2104 - Improved the int/float conversions that should make the code more
2105 robust in face of errors from the node daemons
2106 - Fixed the remove node code in case of internal configuration errors
2107 - Fixed the node daemon behaviour in face of inconsistent queue
2108 directory (e.g. read-only file-system where we can't open the files
2110 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2111 now it either aborts cleanly or, if given the new "auto_promote"
2112 parameter, will automatically promote other nodes as needed
2113 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2114 completely prevent Ganeti from working
2115 - Fixed bug for instance export when not all disks were successfully
2117 - Fixed behaviour of node add when the new node is slow in starting up
2119 - Fixed handling of signals in the LUXI client, which should improve
2120 behaviour of command-line scripts
2121 - Added checks for invalid node/instance names in the configuration (now
2122 flagged during cluster verify)
2123 - Fixed watcher behaviour for disk activation errors
2124 - Fixed two potentially endless loops in http library, which led to the
2125 RAPI daemon hanging and consuming 100% CPU in some cases
2126 - Fixed bug in RAPI daemon related to hashed passwords
2127 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2129 - Enhanced compatibility with non-Debian OSes, but not using absolute
2130 path in some commands and allowing customisation of the ssh
2131 configuration directory
2132 - Fixed possible future issue with new Python versions by abiding to the
2133 proper use of ``__slots__`` attribute on classes
2134 - Added checks that should prevent directory traversal attacks
2135 - Many documentation fixes based on feedback from users
2140 - Added an "early_release" more for instance replace disks and node
2141 evacuate, where we release locks earlier and thus allow higher
2142 parallelism within the cluster
2143 - Added watcher hooks, intended to allow the watcher to restart other
2144 daemons (e.g. from the ganeti-nbma project), but they can be used of
2145 course for any other purpose
2146 - Added a compile-time disable for DRBD barriers, to increase
2147 performance if the administrator trusts the power supply or the
2148 storage system to not lose writes
2149 - Added the option of using syslog for logging instead of, or in
2150 addition to, Ganeti's own log files
2151 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2152 can indeed boot from a paravirtual NIC
2153 - Added a generic debug level for many operations; while this is not
2154 used widely yet, it allows one to pass the debug value all the way to
2156 - Enhanced the hooks environment for instance moves (failovers,
2157 migrations) where the primary/secondary nodes changed during the
2158 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2159 - Enhanced data validations for many user-supplied values; one important
2160 item is the restrictions imposed on instance and node names, which
2161 might reject some (invalid) host names
2162 - Add a configure-time option to disable file-based storage, if it's not
2163 needed; this allows greater security separation between the master
2164 node and the other nodes from the point of view of the inter-node RPC
2166 - Added user notification in interactive tools if job is waiting in the
2167 job queue or trying to acquire locks
2168 - Added log messages when a job is waiting for locks
2169 - Added filtering by node tags in instance operations which admit
2170 multiple instances (start, stop, reboot, reinstall)
2171 - Added a new tool for cluster mergers, ``cluster-merge``
2172 - Parameters from command line which are of the form ``a=b,c=d`` can now
2173 use backslash escapes to pass in values which contain commas,
2174 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2176 - For KVM, the instance name is the first parameter passed to KVM, so
2177 that it's more visible in the process list
2183 *(Released Tue, 2 Mar 2010)*
2185 Ganeti 2.1 brings many improvements with it. Major changes:
2187 - Added infrastructure to ease automated disk repairs
2188 - Added new daemon to export configuration data in a cheaper way than
2189 using the remote API
2190 - Instance NICs can now be routed instead of being associated with a
2192 - Improved job locking logic to reduce impact of jobs acquiring multiple
2193 locks waiting for other long-running jobs
2195 In-depth implementation details can be found in the Ganeti 2.1 design
2201 - Added chroot hypervisor
2202 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2204 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2205 ``bootloader_path`` and ``bootloader_args``)
2206 - Added the ``use_localtime`` option for the xen-hvm and kvm
2207 hypervisors, and the default value for this has changed to false (in
2208 2.0 xen-hvm always enabled it)
2209 - Added luxi call to submit multiple jobs in one go
2210 - Added cluster initialization option to not modify ``/etc/hosts``
2212 - Added network interface parameters
2213 - Added dry run mode to some LUs
2214 - Added RAPI resources:
2216 - ``/2/instances/[instance_name]/info``
2217 - ``/2/instances/[instance_name]/replace-disks``
2218 - ``/2/nodes/[node_name]/evacuate``
2219 - ``/2/nodes/[node_name]/migrate``
2220 - ``/2/nodes/[node_name]/role``
2221 - ``/2/nodes/[node_name]/storage``
2222 - ``/2/nodes/[node_name]/storage/modify``
2223 - ``/2/nodes/[node_name]/storage/repair``
2225 - Added OpCodes to evacuate or migrate all instances on a node
2226 - Added new command to list storage elements on nodes (``gnt-node
2227 list-storage``) and modify them (``gnt-node modify-storage``)
2228 - Added new ssconf files with master candidate IP address
2229 (``ssconf_master_candidates_ips``), node primary IP address
2230 (``ssconf_node_primary_ips``) and node secondary IP address
2231 (``ssconf_node_secondary_ips``)
2232 - Added ``ganeti-confd`` and a client library to query the Ganeti
2233 configuration via UDP
2234 - Added ability to run hooks after cluster initialization and before
2236 - Added automatic mode for disk replace (``gnt-instance replace-disks
2238 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2239 after catastrophic data-loss
2240 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2242 - Added ``gnt-instance move`` command to move instances
2243 - Added ``gnt-cluster watcher`` command to control watcher
2244 - Added ``gnt-node powercycle`` command to powercycle nodes
2245 - Added new job status field ``lock_status``
2246 - Added parseable error codes to cluster verification (``gnt-cluster
2247 verify --error-codes``) and made output less verbose (use
2248 ``--verbose`` to restore previous behaviour)
2249 - Added UUIDs to the main config entities (cluster, nodes, instances)
2250 - Added support for OS variants
2251 - Added support for hashed passwords in the Ganeti remote API users file
2253 - Added option to specify maximum timeout on instance shutdown
2254 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2255 - Added new helper script to start and stop Ganeti daemons
2256 (``daemon-util``), with the intent to reduce the work necessary to
2257 adjust Ganeti for non-Debian distributions and to start/stop daemons
2259 - Added more unittests
2260 - Fixed critical bug in ganeti-masterd startup
2261 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2262 now customisable at the cluster level for both the KVM and Xen
2263 hypervisors using the new ``migration_port`` parameter
2264 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2265 reinstalling an instance
2266 - Allowed ``@`` in tag names
2267 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2268 - Many documentation updates
2269 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2270 - ``gnt-instance reinstall`` can now reinstall multiple instances
2271 - Updated many command line parameters
2272 - Introduced new OS API version 15
2273 - No longer support a default hypervisor
2274 - Treat virtual LVs as inexistent
2275 - Improved job locking logic to reduce lock contention
2276 - Match instance and node names case insensitively
2277 - Reimplemented bash completion script to be more complete
2284 *(Released Thu, 4 Feb 2010)*
2286 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2287 - Fix a string formatting bug
2288 - Improve safety of the code in some error paths
2289 - Improve data validation in the master of values returned from nodes
2295 *(Released Thu, 17 Dec 2009)*
2297 - Fix security issue due to missing validation of iallocator names; this
2298 allows local and remote execution of arbitrary executables
2299 - Fix failure of gnt-node list during instance removal
2300 - Ship the RAPI documentation in the archive
2306 *(Released Wed, 30 Sep 2009)*
2308 - Fixed many wrong messages
2309 - Fixed a few bugs related to the locking library
2310 - Fixed MAC checking at instance creation time
2311 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2312 - Fixed a few issues related to signal handling in both daemons and
2314 - Fixed the example startup script provided
2315 - Fixed insserv dependencies in the example startup script (patch from
2317 - Fixed handling of drained nodes in the iallocator framework
2318 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2320 - Fixed error related to invalid job IDs in job polling
2321 - Fixed job/opcode persistence on unclean master shutdown
2322 - Fixed handling of partial job processing after unclean master
2324 - Fixed error reporting from LUs, previously all errors were converted
2325 into execution errors
2326 - Fixed error reporting from burnin
2327 - Decreased significantly the memory usage of the job queue
2328 - Optimised slightly multi-job submission
2329 - Optimised slightly opcode loading
2330 - Backported the multi-job submit framework from the development
2331 branch; multi-instance start and stop should be faster
2332 - Added script to clean archived jobs after 21 days; this will reduce
2333 the size of the queue directory
2334 - Added some extra checks in disk size tracking
2335 - Added an example ethers hook script
2336 - Added a cluster parameter that prevents Ganeti from modifying of
2338 - Added more node information to RAPI responses
2339 - Added a ``gnt-job watch`` command that allows following the ouput of a
2341 - Added a bind-address option to ganeti-rapi
2342 - Added more checks to the configuration verify
2343 - Enhanced the burnin script such that some operations can be retried
2345 - Converted instance reinstall to multi-instance model
2351 *(Released Fri, 7 Aug 2009)*
2353 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2354 to allow using the pre-2.0.2 behaviour in activation, if any existing
2355 instances have mismatched disk sizes in the configuration
2356 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2357 any configuration mismatches for disk sizes
2358 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2359 failover to work on two-node clusters
2360 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2362 - Fixed detection of OS script errors in ``gnt-backup export``
2363 - Fixed exit code of ``gnt-backup export``
2369 *(Released Fri, 17 Jul 2009)*
2371 - Added experimental support for stripped logical volumes; this should
2372 enhance performance but comes with a higher complexity in the block
2373 device handling; stripping is only enabled when passing
2374 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2375 affected even in the non-stripped mode
2376 - Improved resiliency against transient failures at the end of DRBD
2377 resyncs, and in general of DRBD resync checks
2378 - Fixed a couple of issues with exports and snapshot errors
2379 - Fixed a couple of issues in instance listing
2380 - Added display of the disk size in ``gnt-instance info``
2381 - Fixed checking for valid OSes in instance creation
2382 - Fixed handling of the "vcpus" parameter in instance listing and in
2383 general of invalid parameters
2384 - Fixed http server library, and thus RAPI, to handle invalid
2385 username/password combinations correctly; this means that now they
2386 report unauthorized for queries too, not only for modifications,
2387 allowing earlier detect of configuration problems
2388 - Added a new "role" node list field, equivalent to the master/master
2389 candidate/drained/offline flags combinations
2390 - Fixed cluster modify and changes of candidate pool size
2391 - Fixed cluster verify error messages for wrong files on regular nodes
2392 - Fixed a couple of issues with node demotion from master candidate role
2393 - Fixed node readd issues
2394 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2395 - Added a new ``--no-voting`` option for masterfailover to fix failover
2396 on two-nodes clusters when the former master node is unreachable
2397 - Added instance reinstall over RAPI
2403 *(Released Tue, 16 Jun 2009)*
2405 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2406 allow re-adding the start in single-user option (regression from 1.2)
2407 - the watcher writes the instance status to a file, to allow monitoring
2408 to report the instance status (from the master) based on cached
2409 results of the watcher's queries; while this can get stale if the
2410 watcher is being locked due to other work on the cluster, this is
2411 still an improvement
2412 - the watcher now also restarts the node daemon and the rapi daemon if
2414 - fixed the watcher to handle full and drained queue cases
2415 - hooks export more instance data in the environment, which helps if
2416 hook scripts need to take action based on the instance's properties
2417 (no longer need to query back into ganeti)
2418 - instance failovers when the instance is stopped do not check for free
2419 RAM, so that failing over a stopped instance is possible in low memory
2421 - rapi uses queries for tags instead of jobs (for less job traffic), and
2422 for cluster tags it won't talk to masterd at all but read them from
2424 - a couple of error handling fixes in RAPI
2425 - drbd handling: improved the error handling of inconsistent disks after
2426 resync to reduce the frequency of "there are some degraded disks for
2427 this instance" messages
2428 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2429 error handling path called a wrong function name)
2435 *(Released Wed, 27 May 2009)*
2437 - no changes from rc5
2443 *(Released Wed, 20 May 2009)*
2445 - fix a couple of bugs (validation, argument checks)
2446 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2447 - some small improvements to RAPI and IAllocator
2448 - make watcher automatically start the master daemon if down
2454 *(Released Mon, 27 Apr 2009)*
2456 - change the OS list to not require locks; this helps with big clusters
2457 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2458 volume group is broken
2459 - ``gnt-instance info``, without any arguments, doesn't run for all
2460 instances anymore; either pass ``--all`` or pass the desired
2461 instances; this helps against mistakes on big clusters where listing
2462 the information for all instances takes a long time
2463 - miscellaneous doc and man pages fixes
2469 *(Released Wed, 8 Apr 2009)*
2471 - Change the internal locking model of some ``gnt-node`` commands, in
2472 order to reduce contention (and blocking of master daemon) when
2473 batching many creation/reinstall jobs
2474 - Fixes to Xen soft reboot
2475 - No longer build documentation at build time, instead distribute it in
2476 the archive, in order to reduce the need for the whole docbook/rst
2483 *(Released Fri, 27 Mar 2009)*
2485 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2486 - Fix watcher startup sequence, improves the behaviour of busy clusters
2487 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2488 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2489 bind address and other places
2490 - Some documentation fixes and updates
2496 *(Released Mon, 2 Mar 2009)*
2498 - More documentation updates, now all docs should be more-or-less
2500 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2502 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2504 - Fix an issue related to $libdir/run/ganeti and cluster creation
2510 *(Released Thu, 19 Feb 2009)*
2512 - Xen PVM and KVM have switched the default value for the instance root
2513 disk to the first partition on the first drive, instead of the whole
2514 drive; this means that the OS installation scripts must be changed
2516 - Man pages have been updated
2517 - RAPI has been switched by default to HTTPS, and the exported functions
2518 should all work correctly
2519 - RAPI v1 has been removed
2520 - Many improvements to the KVM hypervisor
2521 - Block device errors are now better reported
2522 - Many other bugfixes and small improvements
2528 *(Released Mon, 26 Jan 2009)*
2530 - Version 2 is a general rewrite of the code and therefore the
2531 differences are too many to list, see the design document for 2.0 in
2532 the ``doc/`` subdirectory for more details
2533 - In this beta version there is not yet a migration path from 1.2 (there
2534 will be one in the final 2.0 release)
2535 - A few significant changes are:
2537 - all commands are executed by a daemon (``ganeti-masterd``) and the
2538 various ``gnt-*`` commands are just front-ends to it
2539 - all the commands are entered into, and executed from a job queue,
2540 see the ``gnt-job(8)`` manpage
2541 - the RAPI daemon supports read-write operations, secured by basic
2542 HTTP authentication on top of HTTPS
2543 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2544 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2545 to migrate to DRBD 8 first while still running Ganeti 1.2)
2546 - DRBD devices are using statically allocated minor numbers, which
2547 will be assigned to existing instances during the migration process
2548 - there is support for both Xen PVM and Xen HVM instances running on
2550 - KVM virtualization is supported too
2551 - file-based storage has been implemented, which means that it is
2552 possible to run the cluster without LVM and DRBD storage, for
2553 example using a shared filesystem exported from shared storage (and
2554 still have live migration)
2560 *(Released Tue, 13 Jan 2009)*
2562 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2563 - Reuse the old instance mac address by default on instance import, if
2564 the instance name is the same.
2565 - Handle situations in which the node info rpc returns incomplete
2567 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2568 - Improved version of batcher:
2570 - state file support
2571 - instance mac address support
2572 - support for HVM clusters/instances
2574 - Add an option to show the number of cpu sockets and nodes in
2576 - Support OSes that handle more than one version of the OS api (but do
2577 not change the current API in any other way)
2578 - Fix ``gnt-node migrate``
2579 - ``gnt-debug`` man page
2580 - Fixes various more typos and small issues
2581 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2587 *(Released Wed, 24 Sep 2008)*
2589 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2590 type of disk exported to fully virtualized instances.
2591 - provide access to the serial console of HVM instances
2592 - instance auto_balance flag, set by default. If turned off it will
2593 avoid warnings on cluster verify if there is not enough memory to fail
2594 over an instance. in the future it will prevent automatically failing
2595 it over when we will support that.
2596 - batcher tool for instance creation, see ``tools/README.batcher``
2597 - ``gnt-instance reinstall --select-os`` to interactively select a new
2598 operating system when reinstalling an instance.
2599 - when changing the memory amount on instance modify a check has been
2600 added that the instance will be able to start. also warnings are
2601 emitted if the instance will not be able to fail over, if auto_balance
2603 - documentation fixes
2604 - sync fields between ``gnt-instance list/modify/add/import``
2605 - fix a race condition in drbd when the sync speed was set after giving
2606 the device a remote peer.
2612 *(Released Tue, 22 Jul 2008)*
2614 - note: the allowed size and number of tags per object were reduced
2615 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2616 - fixed twisted 8.x compatibility
2617 - fixed ``gnt-instance replace-disks`` with iallocator
2618 - add TCP keepalives on twisted connections to detect restarted nodes
2619 - disk increase support, see ``gnt-instance grow-disk``
2620 - implement bulk node/instance query for RAPI
2621 - add tags in node/instance listing (optional)
2622 - experimental migration (and live migration) support, read the man page
2623 for ``gnt-instance migrate``
2624 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2625 has some small improvements in handling its state file
2631 *(Released Fri, 13 Jun 2008)*
2633 - Experimental readonly, REST-based remote API implementation;
2634 automatically started on master node, TCP port 5080, if enabled by
2635 ``--enable-rapi`` parameter to configure script.
2636 - Instance allocator support. Add and import instance accept a
2637 ``--iallocator`` parameter, and call that instance allocator to decide
2638 which node to use for the instance. The iallocator document describes
2639 what's expected from an allocator script.
2640 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2641 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2642 node is lost there is still enough memory to fail over the instances
2644 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2645 ``gnt-cluster verify``, to check for site-specific compliance. All the
2646 hooks will run, and their output, if any, will be displayed. Any
2647 failing hook will make the verification return an error value.
2648 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2649 primary and secondary interfaces
2650 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2651 that is still declared as part of the cluster and has failed.
2652 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2653 specifying output fields, that just adds the chosen fields to the
2655 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2656 export from the filesystem.
2657 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2658 have been added. Using them you can enable/disable acpi and pae
2659 support, and specify a path for a cd image to be exported to the
2660 instance. These parameters as the name suggest only work on HVM
2662 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2663 PAE support will be set to the previously hardcoded values, but the
2664 (previously hardcoded) path to the CDROM ISO image will be unset and
2665 if required, needs to be set manually with ``gnt-instance modify``
2667 - The address to which an instance's VNC console is bound is now
2668 selectable per-instance, rather than being cluster wide. Of course
2669 this only applies to instances controlled via VNC, so currently just
2670 applies to HVM clusters.
2676 *(Released Mon, 18 Feb 2008)*
2678 - more tweaks to the disk activation code (especially helpful for DRBD)
2679 - change the default ``gnt-instance list`` output format, now there is
2680 one combined status field (see the manpage for the exact values this
2682 - some more fixes for the mac export to hooks change
2683 - make Ganeti not break with DRBD 8.2.x (which changed the version
2684 format in ``/proc/drbd``) (issue 24)
2685 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2686 template, allowing migration from DRBD0.7+MD to DRBD8
2692 *(Released Wed, 30 Jan 2008)*
2694 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2696 - add command aliases infrastructure and a few aliases
2697 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2698 man pages and the ``--help`` option of ``gnt-node
2699 list``/``gnt-instance list``
2700 - fix ``gnt-backup list`` with down nodes (issue 21)
2701 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2702 - fix the dist archive and add a check for including svn/git files in
2704 - some developer-related changes: improve the burnin and the QA suite,
2705 add an upload script for testing during development
2711 *(Released Wed, 16 Jan 2008)*
2713 - experimental HVM support, read the install document, section
2714 "Initializing the cluster"
2715 - allow for the PVM hypervisor per-instance kernel and initrd paths
2716 - add a new command ``gnt-cluster verify-disks`` which uses a new
2717 algorithm to improve the reconnection of the DRBD pairs if the device
2718 on the secondary node has gone away
2719 - make logical volume code auto-activate LVs at disk activation time
2720 - slightly improve the speed of activating disks
2721 - allow specification of the MAC address at instance creation time, and
2722 changing it later via ``gnt-instance modify``
2723 - fix handling of external commands that generate lots of output on
2725 - update documentation with regard to minimum version of DRBD8 supported
2731 *(Released Tue, 4 Dec 2007)*
2733 - Log the ``xm create`` output to the node daemon log on failure (to
2734 help diagnosing the error)
2735 - In debug mode, log all external commands output if failed to the logs
2736 - Change parsing of lvm commands to ignore stderr
2742 *(Released Wed, 28 Nov 2007)*
2744 - Another round of updates to the DRBD 8 code to deal with more failures
2745 in the replace secondary node operation
2746 - Some more logging of failures in disk operations (lvm, drbd)
2747 - A few documentation updates
2754 *(Released Tue, 13 Nov 2007)*
2756 - Change configuration file format from Python's Pickle to JSON.
2757 Upgrading is possible using the cfgupgrade utility.
2758 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2759 faster replace disks and is more stable (DRBD 8 has many improvements
2760 compared to DRBD 0.7)
2761 - Added command line tags support (see man pages for ``gnt-instance``,
2762 ``gnt-node``, ``gnt-cluster``)
2763 - Added instance rename support
2764 - Added multi-instance startup/shutdown
2765 - Added cluster rename support
2766 - Added ``gnt-node evacuate`` to simplify some node operations
2767 - Added instance reboot operation that can speedup reboot as compared to
2769 - Soften the requirement that hostnames are in FQDN format
2770 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2772 - Removed dependency on debian's patched fping that uses the
2773 non-standard ``-S`` option
2774 - Now the OS definitions are searched for in multiple, configurable
2775 paths (easier for distros to package)
2776 - Some changes to the hooks infrastructure (especially the new
2777 post-configuration update hook)
2778 - Other small bugfixes
2780 .. vim: set textwidth=72 syntax=rst :