10 Incompatible/important changes
11 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
13 - hroller now also plans for capacity to move non-redundant instances off
14 any node to be rebooted; the old behavior of completely ignoring any
15 non-redundant instances can be restored by adding the --ignore-non-redundant
17 - The cluster option '--no-lvm-storage' was removed in favor of the new option
18 '--enabled-disk-templates'.
19 - On instance creation, disk templates no longer need to be specified
20 with '-t'. The default disk template will be taken from the list of
21 enabled disk templates.
22 - The monitoring daemon is now running as root, in order to be able to collect
23 information only available to root (such as the state of Xen instances).
24 - The ConfD client is now IPv6 compatible.
25 - File and shared file storage is no longer dis/enabled at configure time,
26 but using the option '--enabled-disk-templates' at cluster initialization and
28 - The default directories for file and shared file storage are not anymore
29 specified at configure time, but taken from the cluster's configuration.
30 They can be set at cluster initialization and modification with
31 '--file-storage-dir' and '--shared-file-storage-dir'.
32 - Cluster verification now includes stricter checks regarding the
33 default file and shared file storage directories. It now checks that
34 the directories are explicitely allowed in the 'file-storage-paths' file and
35 that the directories exist on all nodes.
40 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
41 the correct command syntax. It is possible to use different DRBD versions
42 on different nodes as long as they are compatible to each other. This
43 enables rolling upgrades of DRBD with no downtime. As permanent operation
44 of different DRBD versions within a node group is discouraged,
45 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
46 - New "inst-status-xen" data collector for the monitoring daemon, providing
47 information about the state of the xen instances on the nodes.
48 - New "lv" data collector for the monitoring daemon, collecting data about the
49 logical volumes on the nodes, and pairing them with the name of the instances
51 - New "diskstats" data collector, collecting the data from /proc/diskstats and
52 presenting them over the monitoring daemon interface.
53 - The ConfD client is now IPv6 compatible.
57 The following new dependencies have been added.
61 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
62 for the unit tests (and only used for testing).
68 *(Released Fri, 2 Aug 2013)*
70 Incompatible/important changes
71 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
73 - Instance policy can contain multiple instance specs, as described in
74 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
75 <design-partitioned>`. As a consequence, it's not possible to partially change
76 or override instance specs. Bounding specs (min and max) can be specified as a
77 whole using the new option ``--ipolicy-bounds-specs``, while standard
78 specs use the new option ``--ipolicy-std-specs``.
79 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
80 gnt-instance is a valid YAML object.
81 - hail now honors network restrictions when allocating nodes. This led to an
82 update of the IAllocator protocol. See the IAllocator documentation for
84 - confd now only answers static configuration request over the network. luxid
85 was extracted, listens on the local LUXI socket and responds to live queries.
86 This allows finer grained permissions if using separate users.
91 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
92 to always require authentication, ``--require-authentication``. It can
93 be specified in ``$sysconfdir/default/ganeti``.
94 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
95 be used to manage the disk templates to be used by instances in the cluster.
96 Initially, it will be set to a list that includes plain, drbd, if they were
97 enabled by specifying a volume group name, and file and sharedfile, if those
98 were enabled at configure time. Additionally, it will include all disk
99 templates that are currently used by instances. The order of disk templates
100 will be based on Ganeti's history of supporting them. In the future, the
101 first entry of the list will be used as a default disk template on instance
103 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
104 configuration back to the previous stable version.
105 - Disk templates in group ipolicy can be restored to the default value.
106 - Initial support for diskless instances and virtual clusters in QA.
107 - More QA and unit tests for instance policies.
108 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
109 describing why the opcode itself was executed.
110 - The monitoring daemon is now available. It allows users to query the cluster
111 for obtaining information about the status of the system. The daemon is only
112 responsible for providing the information over the network: the actual data
113 gathering is performed by data collectors (currently, only the DRBD status
114 collector is available).
115 - In order to help developers work on Ganeti, a new script
116 (``devel/build_chroot``) is provided, for building a chroot that contains all
117 the required development libraries and tools for compiling Ganeti on a Debian
119 - A new tool, ``harep``, for performing self-repair and recreation of instances
120 in Ganeti has been added.
121 - Split queries are enabled for tags, network, exports, cluster info, groups,
123 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
124 It prints the command line to set the current policies, to ease
126 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
127 devices for KVM virtio-net interfaces will get created with VNET_HDR
128 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
129 virtio-net interfaces, which prevents host kernel tainting and log
130 flooding, when dealing with broken or malicious virtio-net drivers.
131 It's set to true by default.
132 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
137 The following new dependencies have been added:
140 - The ``curl`` library is not optional anymore for compiling the Haskell code.
141 - ``snap-server`` library (if monitoring is enabled).
144 - The minimum Python version needed to run Ganeti is now 2.6.
145 - ``yaml`` library (only for running the QA).
150 - Fix upgrading/downgrading from 2.7
151 - Increase maximum RAPI message size
152 - Documentation updates
153 - Split ``confd`` between ``luxid`` and ``confd``
154 - Merge 2.7 series up to the 2.7.1 release
155 - Allow the ``modify_etc_hosts`` option to be changed
156 - Add better debugging for ``luxid`` queries
157 - Expose bulk parameter for GetJobs in RAPI client
158 - Expose missing ``network`` fields in RAPI
159 - Add some ``cluster verify`` tests
160 - Some unittest fixes
161 - Fix a malfunction in ``hspace``'s tiered allocation
162 - Fix query compatibility between haskell and python implementations
163 - Add the ``vnet_hdr`` HV parameter for KVM
164 - Add ``--cleanup`` to instance failover
165 - Change the connected groups format in ``gnt-network info`` output; it
166 was previously displayed as a raw list by mistake. (Merged from 2.7)
172 *(Released Mon, 24 Jun 2013)*
174 This was the first beta release of the 2.8 series. All important changes
175 are listed in the latest 2.8 entry.
181 *(Released Thu, 25 Jul 2013)*
183 - Add logrotate functionality in daemon-util
184 - Add logrotate example file
185 - Add missing fields to network queries over rapi
186 - Fix network object timestamps
187 - Add support for querying network timestamps
188 - Fix a typo in the example crontab
189 - Fix a documentation typo
195 *(Released Thu, 04 Jul 2013)*
197 Incompatible/important changes
198 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
200 - Instance policies for disk size were documented to be on a per-disk
201 basis, but hail applied them to the sum of all disks. This has been
203 - ``hbal`` will now exit with status 0 if, during job execution over
204 LUXI, early exit has been requested and all jobs are successful;
205 before, exit status 1 was used, which cannot be differentiated from
207 - Compatibility with newer versions of rbd has been fixed
208 - ``gnt-instance batch-create`` has been changed to use the bulk create
209 opcode from Ganeti. This lead to incompatible changes in the format of
210 the JSON file. It's now not a custom dict anymore but a dict
211 compatible with the ``OpInstanceCreate`` opcode.
212 - Parent directories for file storage need to be listed in
213 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
214 write the file automatically based on old configuration values, but it
215 can not distribute it across all nodes and the file contents should be
216 verified. Use ``gnt-cluster copyfile
217 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
218 upgraded. The reason for requiring this list of paths now is that
219 before it would have been possible to inject new paths via RPC,
220 allowing files to be created in arbitrary locations. The RPC protocol
221 is protected using SSL/X.509 certificates, but as a design principle
222 Ganeti does not permit arbitrary paths to be passed.
223 - The parsing of the variants file for OSes (see
224 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
225 lines and comment lines (starting with ``#``) are ignored for better
227 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
228 longer available. ``gnt-node add`` now invokes a new tool on the
229 destination node, named ``prepare-node-join``, to configure the SSH
230 daemon. Paramiko is no longer necessary to configure nodes' SSH
231 daemons via ``gnt-node add``.
232 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
233 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
234 the flag is not reset after a master failover.
235 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
236 is recommended. 2.6 will be mandatory from the 2.8 series.
242 - New network management functionality to support automatic allocation
243 of IP addresses and managing of network parameters. See
244 :manpage:`gnt-network(8)` for more details.
245 - New external storage backend, to allow managing arbitrary storage
246 systems external to the cluster. See
247 :manpage:`ganeti-extstorage-interface(7)`.
248 - New ``exclusive-storage`` node parameter added, restricted to
249 nodegroup level. When it's set to true, physical disks are assigned in
250 an exclusive fashion to instances, as documented in :doc:`Partitioned
251 Ganeti <design-partitioned>`. Currently, only instances using the
252 ``plain`` disk template are supported.
253 - The KVM hypervisor has been updated with many new hypervisor
254 parameters, including a generic one for passing arbitrary command line
255 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
256 compatible up to qemu 1.4.
257 - A new tool, called ``mon-collector``, is the stand-alone executor of
258 the data collectors for a monitoring system. As of this version, it
259 just includes the DRBD data collector, that can be executed by calling
260 ``mon-collector`` using the ``drbd`` parameter. See
261 :manpage:`mon-collector(7)`.
262 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
263 for RAPI users. It allows granting permissions to query for
264 information to a specific user without giving
265 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
266 - A new tool named ``node-cleanup`` has been added. It cleans remains of
267 a cluster from a machine by stopping all daemons, removing
268 certificates and ssconf files. Unless the ``--no-backup`` option is
269 given, copies of the certificates are made.
270 - Instance creations now support the use of opportunistic locking,
271 potentially speeding up the (parallel) creation of multiple instances.
272 This feature is currently only available via the :doc:`RAPI
273 <rapi>` interface and when an instance allocator is used. If the
274 ``opportunistic_locking`` parameter is set the opcode will try to
275 acquire as many locks as possible, but will not wait for any locks
276 held by other opcodes. If not enough resources can be found to
277 allocate the instance, the temporary error code
278 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
279 retried thereafter, with or without opportunistic locking.
280 - New experimental linux-ha resource scripts.
281 - Restricted-commands support: ganeti can now be asked (via command line
282 or rapi) to perform commands on a node. These are passed via ganeti
283 RPC rather than ssh. This functionality is restricted to commands
284 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
285 reasons. The file is not copied automatically.
291 - Diskless instances are now externally mirrored (Issue 237). This for
292 now has only been tested in conjunction with explicit target nodes for
294 - Queries not needing locks or RPC access to the node can now be
295 performed by the confd daemon, making them independent from jobs, and
296 thus faster to execute. This is selectable at configure time.
297 - The functionality for allocating multiple instances at once has been
298 overhauled and is now also available through :doc:`RAPI <rapi>`.
300 There are no significant changes from version 2.7.0~rc3.
306 *(Released Tue, 25 Jun 2013)*
308 - Fix permissions on the confd query socket (Issue 477)
309 - Fix permissions on the job archive dir (Issue 498)
310 - Fix handling of an internal exception in replace-disks (Issue 472)
311 - Fix gnt-node info handling of shortened names (Issue 497)
312 - Fix gnt-instance grow-disk when wiping is enabled
313 - Documentation improvements, and support for newer pandoc
314 - Fix hspace honoring ipolicy for disks (Issue 484)
315 - Improve handling of the ``kvm_extra`` HV parameter
321 *(Released Fri, 24 May 2013)*
323 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
325 - Disks added through ``gnt-instance modify`` or created through
326 ``gnt-instance recreate-disks`` are wiped, if the
327 ``prealloc_wipe_disks`` flag is set.
328 - If wiping newly created disks fails, the disks are removed. Also,
329 partial failures in creating disks through ``gnt-instance modify``
330 triggers a cleanup of the partially-created disks.
331 - Removing the master IP address doesn't fail if the address has been
333 - Fix ownership of the OS log dir
334 - Workaround missing SO_PEERCRED constant (Issue 191)
340 *(Released Fri, 3 May 2013)*
342 This was the first release candidate of the 2.7 series. Since beta3:
344 - Fix kvm compatibility with qemu 1.4 (Issue 389)
345 - Documentation updates (admin guide, upgrade notes, install
346 instructions) (Issue 372)
347 - Fix gnt-group list nodes and instances count (Issue 436)
348 - Fix compilation without non-mandatory libraries (Issue 441)
349 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
350 - Make confd logging more verbose at INFO level (Issue 435)
351 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
352 - Fix failure path for instance storage type conversion (Issue 229)
353 - Update htools text backend documentation
354 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
355 - Disable inter-cluster instance move for file-based instances, because
356 it is dependant on instance export, which is not supported for
357 file-based instances. (Issue 414)
358 - Fix gnt-job crashes on non-ascii characters (Issue 427)
359 - Fix volume group checks on non-vm-capable nodes (Issue 432)
365 *(Released Mon, 22 Apr 2013)*
367 This was the third beta release of the 2.7 series. Since beta2:
369 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
370 - Fix data loss on wrong usage of ``gnt-instance move``
371 - Properly export errors in confd-based job queries
372 - Add ``users-setup`` tool
373 - Fix iallocator protocol to report 0 as a disk size for diskless
374 instances. This avoids hail breaking when a diskless instance is
376 - Fix job queue directory permission problem that made confd job queries
377 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
378 for access to archived jobs (Issue 406).
379 - Limit the sizes of networks supported by ``gnt-network`` to something
380 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
381 - Fix bugs in instance disk template conversion
382 - Fix GHC 7 compatibility
383 - Fix ``burnin`` install path (Issue 426).
384 - Allow very small disk grows (Issue 347).
385 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
386 that noded doesn't import masterd code (Issue 419).
387 - Make sure the default metavg at cluster init is the same as the vg, if
388 unspecified (Issue 358).
389 - Fix cleanup of partially created disks (part of Issue 416)
395 *(Released Tue, 2 Apr 2013)*
397 This was the second beta release of the 2.7 series. Since beta1:
399 - Networks no longer have a "type" slot, since this information was
400 unused in Ganeti: instead of it tags should be used.
401 - The rapi client now has a ``target_node`` option to MigrateInstance.
402 - Fix early exit return code for hbal (Issue 386).
403 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
404 - Fix ``rbd showmapped`` output parsing (Issue 312).
405 - Networks are now referenced indexed by UUID, rather than name. This
406 will require running cfgupgrade, from 2.7.0beta1, if networks are in
408 - The OS environment now includes network information.
409 - Deleting of a network is now disallowed if any instance nic is using
410 it, to prevent dangling references.
411 - External storage is now documented in man pages.
412 - The exclusive_storage flag can now only be set at nodegroup level.
413 - Hbal can now submit an explicit priority with its jobs.
414 - Many network related locking fixes.
415 - Bump up the required pylint version to 0.25.1.
416 - Fix the ``no_remember`` option in RAPI client.
417 - Many ipolicy related tests, qa, and fixes.
418 - Many documentation improvements and fixes.
419 - Fix building with ``--disable-file-storage``.
420 - Fix ``-q`` option in htools, which was broken if passed more than
422 - Some haskell/python interaction improvements and fixes.
423 - Fix iallocator in case of missing LVM storage.
424 - Fix confd config load in case of ``--no-lvm-storage``.
425 - The confd/query functionality is now mentioned in the security
432 *(Released Wed, 6 Feb 2013)*
434 This was the first beta release of the 2.7 series. All important changes
435 are listed in the latest 2.7 entry.
441 *(Released Fri, 21 Dec 2012)*
443 Important behaviour change: hbal won't rebalance anymore instances which
444 have the ``auto_balance`` attribute set to false. This was the intention
445 all along, but until now it only skipped those from the N+1 memory
446 reservation (DRBD-specific).
448 A significant number of bug fixes in this release:
450 - Fixed disk adoption interaction with ipolicy checks.
451 - Fixed networking issues when instances are started, stopped or
452 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
453 - Fixed the warning in cluster verify for shared storage instances not
455 - Fixed removal of storage directory on shared file storage (issue 262).
456 - Fixed validation of LVM volume group name in OpClusterSetParams
457 (``gnt-cluster modify``) (issue 285).
458 - Fixed runtime memory increases (``gnt-instance modify -m``).
459 - Fixed live migration under Xen's ``xl`` mode.
460 - Fixed ``gnt-instance console`` with ``xl``.
461 - Fixed building with newer Haskell compiler/libraries.
462 - Fixed PID file writing in Haskell daemons (confd); this prevents
463 restart issues if confd was launched manually (outside of
464 ``daemon-util``) while another copy of it was running
465 - Fixed a type error when doing live migrations with KVM (issue 297) and
466 the error messages for failing migrations have been improved.
467 - Fixed opcode validation for the out-of-band commands (``gnt-node
469 - Fixed a type error when unsetting OS hypervisor parameters (issue
470 311); now it's possible to unset all OS-specific hypervisor
472 - Fixed the ``dry-run`` mode for many operations: verification of
473 results was over-zealous but didn't take into account the ``dry-run``
474 operation, resulting in "wrong" failures.
475 - Fixed bash completion in ``gnt-job list`` when the job queue has
476 hundreds of entries; especially with older ``bash`` versions, this
477 results in significant CPU usage.
479 And lastly, a few other improvements have been made:
481 - Added option to force master-failover without voting (issue 282).
482 - Clarified error message on lock conflict (issue 287).
483 - Logging of newly submitted jobs has been improved (issue 290).
484 - Hostname checks have been made uniform between instance rename and
486 - The ``--submit`` option is now supported by ``gnt-debug delay``.
487 - Shutting down the master daemon by sending SIGTERM now stops it from
488 processing jobs waiting for locks; instead, those jobs will be started
489 once again after the master daemon is started the next time (issue
491 - Support for Xen's ``xl`` program has been improved (besides the fixes
493 - Reduced logging noise in the Haskell confd daemon (only show one log
494 entry for each config reload, instead of two).
495 - Several man page updates and typo fixes.
501 *(Released Fri, 12 Oct 2012)*
503 A small bugfix release. Among the bugs fixed:
505 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
506 made the command unusable.
507 - Commands that issue many jobs don't fail anymore just because some jobs
508 take so long that other jobs are archived.
509 - Failures during ``gnt-instance reinstall`` are reflected by the exit
511 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
513 - When ``always_failover`` is set, ``--allow-failover`` is not required
514 in migrate commands anymore.
515 - ``bash_completion`` works even if extglob is disabled.
516 - Fixed bug with locks that made failover for RDB-based instances fail.
517 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
518 a random node instead of one based on the allocator metric.
519 - Support for newer versions of pylint and pep8.
520 - Hail doesn't fail anymore when trying to add an instance of type
521 ``file``, ``sharedfile`` or ``rbd``.
522 - Added new Makefile target to rebuild the whole distribution, so that
523 all files are included.
529 *(Released Fri, 27 Jul 2012)*
532 .. attention:: The ``LUXI`` protocol has been made more consistent
533 regarding its handling of command arguments. This, however, leads to
534 incompatibility issues with previous versions. Please ensure that you
535 restart Ganeti daemons soon after the upgrade, otherwise most
536 ``LUXI`` calls (job submission, setting/resetting the drain flag,
537 pausing/resuming the watcher, cancelling and archiving jobs, querying
538 the cluster configuration) will fail.
547 The current ``admin_up`` field, which used to denote whether an instance
548 should be running or not, has been removed. Instead, ``admin_state`` is
549 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
551 The rational behind this is that an instance being “down” can have
554 - it could be down during a reboot
555 - it could be temporarily be down for a reinstall
556 - or it could be down because it is deprecated and kept just for its
559 The previous Boolean state was making it difficult to do capacity
560 calculations: should Ganeti reserve memory for a down instance? Now, the
561 tri-state field makes it clear:
563 - in ``up`` and ``down`` state, all resources are reserved for the
564 instance, and it can be at any time brought up if it is down
565 - in ``offline`` state, only disk space is reserved for it, but not
568 The field can have an extra use: since the transition between ``up`` and
569 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
570 transition between ``offline`` and ``down`` is done via ``gnt-instance
571 modify``, it is possible to given different rights to users. For
572 example, owners of an instance could be allowed to start/stop it, but
573 not transition it out of the offline state.
575 Instance policies and specs
576 +++++++++++++++++++++++++++
578 In previous Ganeti versions, an instance creation request was not
579 limited on the minimum size and on the maximum size just by the cluster
580 resources. As such, any policy could be implemented only in third-party
581 clients (RAPI clients, or shell wrappers over ``gnt-*``
582 tools). Furthermore, calculating cluster capacity via ``hspace`` again
583 required external input with regards to instance sizes.
585 In order to improve these workflows and to allow for example better
586 per-node group differentiation, we introduced instance specs, which
589 - minimum instance disk size, disk count, memory size, cpu count
590 - maximum values for the above metrics
591 - and “standard” values (used in ``hspace`` to calculate the standard
594 The minimum/maximum values can be also customised at node-group level,
595 for example allowing more powerful hardware to support bigger instance
598 Beside the instance specs, there are a few other settings belonging to
599 the instance policy framework. It is possible now to customise, per
600 cluster and node-group:
602 - the list of allowed disk templates
603 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
604 - the maximum ratio of instance to spindles (see below for more
605 information) for local storage
607 All these together should allow all tools that talk to Ganeti to know
608 what are the ranges of allowed values for instances and the
609 over-subscription that is allowed.
611 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
612 instance configuration, and the physical CPU configuration from the
613 node. For the spindle ratios however, we didn't track before these
614 values, so new parameters have been added:
616 - a new node parameter ``spindle_count``, defaults to 1, customisable at
617 node group or node level
618 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
620 Note that spindles in this context doesn't need to mean actual
621 mechanical hard-drives; it's just a relative number for both the node
622 I/O capacity and instance I/O consumption.
624 Instance migration behaviour
625 ++++++++++++++++++++++++++++
627 While live-migration is in general desirable over failover, it is
628 possible that for some workloads it is actually worse, due to the
629 variable time of the “suspend” phase during live migration.
631 To allow the tools to work consistently over such instances (without
632 having to hard-code instance names), a new backend parameter
633 ``always_failover`` has been added to control the migration/failover
634 behaviour. When set to True, all migration requests for an instance will
635 instead fall-back to failover.
637 Instance memory ballooning
638 ++++++++++++++++++++++++++
640 Initial support for memory ballooning has been added. The memory for an
641 instance is no longer fixed (backend parameter ``memory``), but instead
642 can vary between minimum and maximum values (backend parameters
643 ``minmem`` and ``maxmem``). Currently we only change an instance's
646 - live migrating or failing over and instance and the target node
647 doesn't have enough memory
648 - user requests changing the memory via ``gnt-instance modify
654 In order to control the use of specific CPUs by instance, support for
655 controlling CPU pinning has been added for the Xen, HVM and LXC
656 hypervisors. This is controlled by a new hypervisor parameter
657 ``cpu_mask``; details about possible values for this are in the
658 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
659 VCPU-to-CPU mapping) form will work well only when all nodes in your
660 cluster have the same amount of CPUs.
665 Another area in which Ganeti was not customisable were the parameters
666 used for storage configuration, e.g. how many stripes to use for LVM,
667 DRBD resync configuration, etc.
669 To improve this area, we've added disks parameters, which are
670 customisable at cluster and node group level, and which allow to
671 specify various parameters for disks (DRBD has the most parameters
672 currently), for example:
674 - DRBD resync algorithm and parameters (e.g. speed)
675 - the default VG for meta-data volumes for DRBD
676 - number of stripes for LVM (plain disk template)
679 These parameters can be modified via ``gnt-cluster modify -D …`` and
680 ``gnt-group modify -D …``, and are used at either instance creation (in
681 case of LVM stripes, for example) or at disk “activation” time
684 Rados block device support
685 ++++++++++++++++++++++++++
687 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
688 denoted by the ``rbd`` disk template type. This is considered
689 experimental, feedback is welcome. For details on configuring it, see
690 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
695 The existing master IP functionality works well only in simple setups (a
696 single network shared by all nodes); however, if nodes belong to
697 different networks, then the ``/32`` setup and lack of routing
698 information is not enough.
700 To allow the master IP to function well in more complex cases, the
701 system was reworked as follows:
703 - a master IP netmask setting has been added
704 - the master IP activation/turn-down code was moved from the node daemon
706 - whether to run the Ganeti-supplied master IP script or a user-supplied
707 on is a ``gnt-cluster init`` setting
709 Details about the location of the standard and custom setup scripts are
710 in the man page :manpage:`gnt-cluster(8)`; for information about the
711 setup script protocol, look at the Ganeti-supplied script.
716 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
719 It is now possible to use TLS-protected connections, and when renewing
720 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
721 it is now possible to specify spice or spice CA certificates. Also, it
722 is possible to configure a password for SPICE sessions via the
723 hypervisor parameter ``spice_password_file``.
725 There are also new parameters to control the compression and streaming
726 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
727 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
728 for the spice parameters.
730 Lastly, it is now possible to see the SPICE connection information via
731 ``gnt-instance console``.
736 A new tool (``tools/ovfconverter``) has been added that supports
737 conversion between Ganeti and the `Open Virtualization Format
738 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
741 This relies on the ``qemu-img`` tool to convert the disk formats, so the
742 actual compatibility with other virtualization solutions depends on it.
747 The configuration query daemon (``ganeti-confd``) is now optional, and
748 has been rewritten in Haskell; whether to use the daemon at all, use the
749 Python (default) or the Haskell version is selectable at configure time
750 via the ``--enable-confd`` parameter, which can take one of the
751 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
752 daemon will result in a smaller footprint; for larger systems, we
753 welcome feedback on the Haskell version which might become the default
756 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
757 daemon running. The Python version doesn't implement the new call.
760 User interface changes
761 ~~~~~~~~~~~~~~~~~~~~~~
763 We have replaced the ``--disks`` option of ``gnt-instance
764 replace-disks`` with a more flexible ``--disk`` option, which allows
765 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
766 disk size and mode can be changed upon recreation (via ``gnt-instance
767 recreate-disks``, which accepts the same ``--disk`` option).
769 As many people are used to a ``show`` command, we have added that as an
770 alias to ``info`` on all ``gnt-*`` commands.
772 The ``gnt-instance grow-disk`` command has a new mode in which it can
773 accept the target size of the disk, instead of the delta; this can be
774 more safe since two runs in absolute mode will be idempotent, and
775 sometimes it's also easier to specify the desired size directly.
777 Also the handling of instances with regard to offline secondaries has
778 been improved. Instance operations should not fail because one of it's
779 secondary nodes is offline, even though it's safe to proceed.
781 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
782 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
783 minors to instance name.
788 RAPI coverage has improved, with (for example) new resources for
789 recreate-disks, node power-cycle, etc.
794 There is partial support for ``xl`` in the Xen hypervisor; feedback is
797 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
798 whether to still support Python 2.4 or move to Python 2.6 as minimum
801 Support for Fedora has been slightly improved; the provided example
802 init.d script should work better on it and the INSTALL file should
803 document the needed dependencies.
808 The deprecated ``QueryLocks`` LUXI request has been removed. Use
809 ``Query(what=QR_LOCK, ...)`` instead.
811 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
812 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
813 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
814 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
815 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
817 RAPI client: ``CertificateError`` now derives from
818 ``GanetiApiError``. This should make it more easy to handle Ganeti
821 Deprecation warnings due to PyCrypto/paramiko import in
822 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
823 make sure to run an up-to-date paramiko version, if you use this tool.
825 The QA scripts now depend on Python 2.5 or above (the main code base
826 still works with Python 2.4).
828 The configuration file (``config.data``) is now written without
829 indentation for performance reasons; if you want to edit it, it can be
830 re-formatted via ``tools/fmtjson``.
832 A number of bugs has been fixed in the cluster merge tool.
834 ``x509`` certification verification (used in import-export) has been
835 changed to allow the same clock skew as permitted by the cluster
836 verification. This will remove some rare but hard to diagnose errors in
843 *(Released Thu, 19 Jul 2012)*
845 Very few changes from rc4 to the final release, only bugfixes:
847 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
848 instance, fix CDROM booting for KVM instances)
849 - fixed node group modification of node parameters
850 - fixed issue in LUClusterVerifyGroup with multi-group clusters
851 - fixed generation of bash completion to ensure a stable ordering
858 *(Released Fri, 13 Jul 2012)*
860 Third release candidate for 2.6. The following changes were done from
863 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
864 - Fixed an inconsistency in the LUXI protocol with the provided
865 arguments (NOT backwards compatible)
866 - Fixed a bug with node groups ipolicy where ``min`` was greater than
867 the cluster ``std`` value
868 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
869 easier instance debugging on nodes (requires ``hconfd`` to work)
875 *(Released Tue, 03 Jul 2012)*
877 Second release candidate for 2.6. The following changes were done from
880 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
882 - Fixed a RAPI regression on missing beparams/memory
883 - Fixed redistribution of files on offline nodes
884 - Added possibility to run activate-disks even though secondaries are
885 offline. With this change it relaxes also the strictness on some other
886 commands which use activate disks internally:
887 * ``gnt-instance start|reboot|rename|backup|export``
888 - Made it possible to remove safely an instance if its secondaries are
890 - Made it possible to reinstall even though secondaries are offline
896 *(Released Mon, 25 Jun 2012)*
898 First release candidate for 2.6. The following changes were done from
901 - Fixed bugs with disk parameters and ``rbd`` templates as well as
903 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
904 behaviour. It supports now the modify operation
905 - ``hcheck`` implemented to analyze cluster health and possibility of
906 improving health by rebalance
907 - ``hbal`` has been improved in dealing with split instances
913 *(Released Mon, 11 Jun 2012)*
915 Second beta release of 2.6. The following changes were done from beta2
918 - Fixed ``daemon-util`` with non-root user models
919 - Fixed creation of plain instances with ``--no-wait-for-sync``
920 - Fix wrong iv_names when running ``cfgupgrade``
921 - Export more information in RAPI group queries
922 - Fixed bug when changing instance network interfaces
923 - Extended burnin to do NIC changes
924 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
925 - Changed default for DRBD barriers
926 - Fixed DRBD error reporting for syncer rate
927 - Verify the options on disk parameters
929 And of course various fixes to documentation and improved unittests and
936 *(Released Wed, 23 May 2012)*
938 First beta release of 2.6. The following changes were done from beta1 to
941 - integrated patch for distributions without ``start-stop-daemon``
942 - adapted example init.d script to work on Fedora
943 - fixed log handling in Haskell daemons
944 - adapted checks in the watcher for pycurl linked against libnss
945 - add partial support for ``xl`` instead of ``xm`` for Xen
946 - fixed a type issue in cluster verification
947 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
950 Plus integrated fixes from the 2.5 branch:
952 - fixed ``kvm-ifup`` to use ``/bin/bash``
953 - fixed parallel build failures
954 - KVM live migration when using a custom keymap
960 *(Released Tue, 24 Jul 2012)*
962 A small bugfix release, with no new features:
964 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
965 different default shell (e.g. Debian, Ubuntu)
966 - fixed KVM startup and live migration with a custom keymap (fixes Issue
967 243 and Debian bug #650664)
968 - fixed compatibility with KVM versions that don't support multiple boot
969 devices (fixes Issue 230 and Debian bug #624256)
971 Additionally, a few fixes were done to the build system (fixed parallel
972 build failures) and to the unittests (fixed race condition in test for
973 FileID functions, and the default enable/disable mode for QA test is now
980 *(Released Fri, 11 May 2012)*
982 A small bugfix release.
984 The main issues solved are on the topic of compatibility with newer LVM
987 - fixed parsing of ``lv_attr`` field
988 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
989 the ``--force`` flag is needed
991 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
992 to accept kernel 3.x too (was hardcoded to 2.6.*).
994 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
995 of hook results and that also made display of other errors suboptimal
996 was fixed; the code behaves now like 2.4 and earlier.
998 Another change in 2.5, the cleanup of the OS scripts environment, is too
999 aggressive: it removed even the ``PATH`` variable, which requires the OS
1000 scripts to *always* need to export it. Since this is a bit too strict,
1001 we now export a minimal PATH, the same that we export for hooks.
1003 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1004 integrated into this release.
1006 Finally, a few other miscellaneous changes were done (no new features,
1007 just small improvements):
1009 - Fix ``gnt-group --help`` display
1010 - Fix hardcoded Xen kernel path
1011 - Fix grow-disk handling of invalid units
1012 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1013 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1019 *(Released Thu, 12 Apr 2012)*
1021 Incompatible/important changes and bugfixes
1022 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1024 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1025 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1026 to match the underlying LUXI interface.
1027 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1028 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1029 not maintain backwards-compatibility as the underlying operation
1030 changed in an incompatible way. The RAPI client can talk to old
1031 servers, but it needs to be told so as the return value changed.
1032 - When creating file-based instances via RAPI, the ``file_driver``
1033 parameter no longer defaults to ``loop`` and must be specified.
1034 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1036 - Support for the undocumented and deprecated RAPI instance creation
1037 request format version 0 has been dropped. Use version 1, supported
1038 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1039 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1040 documentation <install>`.
1041 - The "cluster-verify" hooks are now executed per group by the
1042 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1043 if you just run ``gnt-cluster verify``, which generates one opcode per
1045 - The environment as passed to the OS scripts is cleared, and thus no
1046 environment variables defined in the node daemon's environment will be
1047 inherited by the scripts.
1048 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1050 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1051 support operations involving multiple node groups.
1052 - Offline nodes are ignored when failing over an instance.
1053 - Support for KVM version 1.0, which changed the version reporting format
1055 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1057 - ``Makefile`` is now compatible with Automake 1.11.2
1058 - Includes all bugfixes made in the 2.4 series
1063 - The ganeti-htools project has been merged into the ganeti-core source
1064 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1065 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1066 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1067 GPT-style partition tables (requires `parted
1068 <http://www.gnu.org/s/parted/>`_).
1069 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1070 - Allowed adding tags on instance creation.
1071 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1072 - Allow instances to be started in a paused state, enabling the user to
1073 see the complete console output on boot using the console.
1074 - Added new hypervisor flag to control default reboot behaviour
1075 (``reboot_behavior``).
1076 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1077 - Improved out-of-band management support:
1079 - Added ``gnt-node health`` command reporting the health status of
1081 - Added ``gnt-node power`` command to manage power status of nodes.
1082 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1084 - Instance migration can fall back to failover if instance is not
1086 - Filters can be used when listing nodes, instances, groups and locks;
1087 see :manpage:`ganeti(7)` manpage.
1088 - Added post-execution status as variables to :doc:`hooks <hooks>`
1090 - Instance tags are exported/imported together with the instance.
1091 - When given an explicit job ID, ``gnt-job info`` will work for archived
1093 - Jobs can define dependencies on other jobs (not yet supported via
1094 RAPI or command line, but used by internal commands and usable via
1097 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1100 - Instance failover is now available as a RAPI resource
1101 (``/2/instances/[instance_name]/failover``).
1102 - ``gnt-instance info`` defaults to static information if primary node
1104 - Opcodes have a new ``comment`` attribute.
1105 - Added basic SPICE support to KVM hypervisor.
1106 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1108 Node group improvements
1109 ~~~~~~~~~~~~~~~~~~~~~~~
1111 - ``gnt-cluster verify`` has been modified to check groups separately,
1112 thereby improving performance.
1113 - Node group support has been added to ``gnt-cluster verify-disks``,
1114 which now operates per node group.
1115 - Watcher has been changed to work better with node groups.
1117 - One process and state file per node group.
1118 - Slow watcher in one group doesn't block other group's watcher.
1120 - Added new command, ``gnt-group evacuate``, to move all instances in a
1121 node group to other groups.
1122 - Added ``gnt-instance change-group`` to move an instance to another
1124 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1125 per-group operations.
1126 - Node groups can be tagged.
1127 - Some operations switch from an exclusive to a shared lock as soon as
1129 - Instance's primary and secondary nodes' groups are now available as
1130 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1131 and ``snodes.group.uuid``).
1136 - Numerous updates to documentation and manpages.
1138 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1140 - Some opcode/job results are now also documented, see :doc:`RAPI
1143 - A lockset's internal lock is now also visible in lock monitor.
1144 - Log messages from job queue workers now contain information about the
1145 opcode they're processing.
1146 - ``gnt-instance console`` no longer requires the instance lock.
1147 - A short delay when waiting for job changes reduces the number of LUXI
1148 requests significantly.
1149 - DRBD metadata volumes are overwritten with zeros during disk creation.
1150 - Out-of-band commands no longer acquire the cluster lock in exclusive
1152 - ``devel/upload`` now uses correct permissions for directories.
1158 *(Released Fri, 23 Mar 2012)*
1160 This was the sixth release candidate of the 2.5 series.
1166 *(Released Mon, 9 Jan 2012)*
1168 This was the fifth release candidate of the 2.5 series.
1174 *(Released Thu, 27 Oct 2011)*
1176 This was the fourth release candidate of the 2.5 series.
1182 *(Released Wed, 26 Oct 2011)*
1184 This was the third release candidate of the 2.5 series.
1190 *(Released Tue, 18 Oct 2011)*
1192 This was the second release candidate of the 2.5 series.
1198 *(Released Tue, 4 Oct 2011)*
1200 This was the first release candidate of the 2.5 series.
1206 *(Released Wed, 31 Aug 2011)*
1208 This was the third beta release of the 2.5 series.
1214 *(Released Mon, 22 Aug 2011)*
1216 This was the second beta release of the 2.5 series.
1222 *(Released Fri, 12 Aug 2011)*
1224 This was the first beta release of the 2.5 series.
1230 *(Released Thu, 27 Oct 2011)*
1232 - Fixed bug when parsing command line parameter values ending in
1234 - Fixed assertion error after unclean master shutdown
1235 - Disable HTTP client pool for RPC, significantly reducing memory usage
1237 - Fixed queue archive creation with wrong permissions
1243 *(Released Tue, 23 Aug 2011)*
1247 - Fixed documentation for importing with ``--src-dir`` option
1248 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1249 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1255 *(Released Fri, 5 Aug 2011)*
1257 Many bug-fixes and a few small features:
1259 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1260 issues when you tried to add an instance with two MAC addresses in one
1262 - KVM: fixed per-instance stored UID value
1263 - KVM: configure bridged NICs at migration start
1264 - KVM: Fix a bug where instance will not start with never KVM versions
1266 - Added OS search path to ``gnt-cluster info``
1267 - Fixed an issue with ``file_storage_dir`` where you were forced to
1268 provide an absolute path, but the documentation states it is a
1269 relative path, the documentation was right
1270 - Added a new parameter to instance stop/start called ``--no-remember``
1271 that will make the state change to not be remembered
1272 - Implemented ``no_remember`` at RAPI level
1273 - Improved the documentation
1274 - Node evacuation: don't call IAllocator if node is already empty
1275 - Fixed bug in DRBD8 replace disks on current nodes
1276 - Fixed bug in recreate-disks for DRBD instances
1277 - Moved assertion checking locks in ``gnt-instance replace-disks``
1278 causing it to abort with not owning the right locks for some situation
1279 - Job queue: Fixed potential race condition when cancelling queued jobs
1280 - Fixed off-by-one bug in job serial generation
1281 - ``gnt-node volumes``: Fix instance names
1282 - Fixed aliases in bash completion
1283 - Fixed a bug in reopening log files after being sent a SIGHUP
1284 - Added a flag to burnin to allow specifying VCPU count
1285 - Bugfixes to non-root Ganeti configuration
1291 *(Released Thu, 12 May 2011)*
1293 Many bug-fixes and a few new small features:
1295 - Fixed a bug related to log opening failures
1296 - Fixed a bug in instance listing with orphan instances
1297 - Fixed a bug which prevented resetting the cluster-level node parameter
1298 ``oob_program`` to the default
1299 - Many fixes related to the ``cluster-merge`` tool
1300 - Fixed a race condition in the lock monitor, which caused failures
1301 during (at least) creation of many instances in parallel
1302 - Improved output for gnt-job info
1303 - Removed the quiet flag on some ssh calls which prevented debugging
1305 - Improved the N+1 failure messages in cluster verify by actually
1306 showing the memory values (needed and available)
1307 - Increased lock attempt timeouts so that when executing long operations
1308 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1309 too early and thus prevent the use of the 'fair' mechanism
1310 - Changed instance query data (``gnt-instance info``) to not acquire
1311 locks unless needed, thus allowing its use on locked instance if only
1312 static information is asked for
1313 - Improved behaviour with filesystems that do not support rename on an
1315 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1316 kept locks on all nodes during the wipe, which is unneeded
1317 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1318 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1319 10GiB) which caused the wipe to fail right at the end in some cases
1320 - Fixed master IP activation when doing master failover with no-voting
1321 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1322 the master node itself
1323 - Fixed potential data-loss in under disk full conditions, where Ganeti
1324 wouldn't check correctly the return code and would consider
1325 partially-written files 'correct'
1326 - Fixed bug related to multiple VGs and DRBD disk replacing
1327 - Added new disk parameter ``metavg`` that allows placement of the meta
1328 device for DRBD in a different volume group
1329 - Fixed error handling in the node daemon when the system libc doesn't
1330 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1331 - Fixed lock release during replace-disks, which kept cluster-wide locks
1332 when doing disk replaces with an iallocator script
1333 - Added check for missing bridges in cluster verify
1334 - Handle EPIPE errors while writing to the terminal better, so that
1335 piping the output to e.g. ``less`` doesn't cause a backtrace
1336 - Fixed rare case where a ^C during Luxi calls could have been
1337 interpreted as server errors, instead of simply terminating
1338 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1340 - Added a few more parameters to the KVM hypervisor, allowing a second
1341 CDROM, custom disk type for CDROMs and a floppy image
1342 - Removed redundant message in instance rename when the name is given
1344 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1345 disks on new nodes, allowing recreation when the original instance
1346 nodes are completely gone
1347 - Added option when converting disk templates to DRBD to skip waiting
1348 for the resync, in order to make the instance available sooner
1349 - Added two new variables to the OS scripts environment (containing the
1351 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1352 to allow use of ``pvgrub`` as bootloader
1353 - Changed the instance memory modifications to only check out-of-memory
1354 conditions on memory increases, and turned the secondary node warnings
1355 into errors (they can still be overridden via ``--force``)
1356 - Fixed the handling of a corner case when the Python installation gets
1357 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1358 to execute a command that doesn't exist
1359 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1360 node of the instance returned failures during instance shutdown; this
1361 adds the option ``--ignore-consistency`` to gnt-instance move
1363 And as usual, various improvements to the error messages, documentation
1370 *(Released Wed, 09 Mar 2011)*
1372 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1373 the RAPI users file if run twice (even with ``--dry-run``).
1375 The release fixes that bug (nothing else changed).
1381 *(Released Mon, 07 Mar 2011)*
1383 Final 2.4.0 release. Just a few small fixes:
1385 - Fixed RAPI node evacuate
1386 - Fixed the kvm-ifup script
1387 - Fixed internal error handling for special job cases
1388 - Updated man page to specify the escaping feature for options
1394 *(Released Mon, 28 Feb 2011)*
1396 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1397 that this feature wiped the disks of the wrong instance, leading to loss
1402 - Fixed title of query field containing instance name
1403 - Expanded the glossary in the documentation
1404 - Fixed one unittest (internal issue)
1410 *(Released Mon, 21 Feb 2011)*
1412 A number of bug fixes plus just a couple functionality changes.
1414 On the user-visible side, the ``gnt-* list`` command output has changed
1415 with respect to "special" field states. The current rc1 style of display
1416 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1417 the default output mode special fields states are displayed as follows:
1419 - Offline resource: ``*``
1420 - Unavailable/not applicable: ``-``
1421 - Data missing (RPC failure): ``?``
1422 - Unknown field: ``??``
1424 Another user-visible change is the addition of ``--force-join`` to
1429 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1430 - Fixed regression in RAPI/instance reinstall where all parameters were
1431 required (instead of optional)
1432 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1433 - Fixed iallocator usage (offline nodes were not considered offline)
1434 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1435 - Fixed hypervisor and OS parameter validation with respect to
1436 non-vm_capable nodes
1437 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1439 - Fixed ``tools/listrunner`` with respect to agent-based usage
1445 *(Released Fri, 4 Feb 2011)*
1447 Many changes and fixes since the beta1 release. While there were some
1448 internal changes, the code has been mostly stabilised for the RC
1451 Note: the dumb allocator was removed in this release, as it was not kept
1452 up-to-date with the IAllocator protocol changes. It is recommended to
1453 use the ``hail`` command from the ganeti-htools package.
1455 Note: the 2.4 and up versions of Ganeti are not compatible with the
1456 0.2.x branch of ganeti-htools. You need to upgrade to
1457 ganeti-htools-0.3.0 (or later).
1459 Regressions fixed from 2.3
1460 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1462 - Fixed the ``gnt-cluster verify-disks`` command
1463 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1465 - Fixed disk adoption breakage
1466 - Fixed wrong headers in instance listing for field aliases
1471 - Fixed corner case in KVM handling of NICs
1472 - Fixed many cases of wrong handling of non-vm_capable nodes
1473 - Fixed a bug where a missing instance symlink was not possible to
1474 recreate with any ``gnt-*`` command (now ``gnt-instance
1475 activate-disks`` does it)
1476 - Fixed the volume group name as reported by ``gnt-cluster
1478 - Increased timeouts for the import-export code, hopefully leading to
1479 fewer aborts due network or instance timeouts
1480 - Fixed bug in ``gnt-node list-storage``
1481 - Fixed bug where not all daemons were started on cluster
1482 initialisation, but only at the first watcher run
1483 - Fixed many bugs in the OOB implementation
1484 - Fixed watcher behaviour in presence of instances with offline
1486 - Fixed instance list output for instances running on the wrong node
1487 - a few fixes to the cluster-merge tool, but it still cannot merge
1488 multi-node groups (currently it is not recommended to use this tool)
1494 - Improved network configuration for the KVM hypervisor
1495 - Added e1000 as a supported NIC for Xen-HVM
1496 - Improved the lvmstrap tool to also be able to use partitions, as
1497 opposed to full disks
1498 - Improved speed of disk wiping (the cluster parameter
1499 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1500 of instance creations
1501 - Added documentation for the OS parameters
1502 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1503 hypervisor is not responding
1504 - Added display of blacklisted and hidden OS information in
1505 ``gnt-cluster info``
1506 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1507 NIC and node parameters, which might create problems with currently
1508 invalid (but undetected) configuration files, but prevents validation
1509 failures when unrelated parameters are modified
1510 - Changed cluster initialisation to wait for the master daemon to become
1512 - Expanded the RAPI interface:
1514 - Added config redistribution resource
1515 - Added activation/deactivation of instance disks
1516 - Added export of console information
1518 - Implemented log file reopening on SIGHUP, which allows using
1519 logrotate(8) for the Ganeti log files
1520 - Added a basic OOB helper script as an example
1526 *(Released Fri, 14 Jan 2011)*
1531 - Fixed timezone issues when formatting timestamps
1532 - Added support for node groups, available via ``gnt-group`` and other
1534 - Added out-of-band framework and management, see :doc:`design
1535 document <design-oob>`
1536 - Removed support for roman numbers from ``gnt-node list`` and
1537 ``gnt-instance list``.
1538 - Allowed modification of master network interface via ``gnt-cluster
1539 modify --master-netdev``
1540 - Accept offline secondaries while shutting down instance disks
1541 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1542 - Added support for multiple LVM volume groups
1543 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1545 - Added commands to list available fields:
1547 - ``gnt-node list-fields``
1548 - ``gnt-group list-fields``
1549 - ``gnt-instance list-fields``
1551 - Updated documentation and man pages
1556 - Moved ``rapi_users`` file into separate directory, now named
1557 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1559 - Added new tool for running commands on many machines,
1560 ``tools/ganeti-listrunner``
1561 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1562 improving the ``gnt-instance console`` output
1563 - Allowed customisation of disk index separator at ``configure`` time
1564 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1565 - Added support for non-partitioned md disks in ``lvmstrap``
1566 - Added script to gracefully power off KVM instances
1567 - Split ``utils`` module into smaller parts
1568 - Changed query operations to return more detailed information, e.g.
1569 whether an information is unavailable due to an offline node. To use
1570 this new functionality, the LUXI call ``Query`` must be used. Field
1571 information is now stored by the master daemon and can be retrieved
1572 using ``QueryFields``. Instances, nodes and groups can also be queried
1573 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1574 exposed via RAPI). The following commands make use of this
1575 infrastructure change:
1577 - ``gnt-group list``
1578 - ``gnt-group list-fields``
1580 - ``gnt-node list-fields``
1581 - ``gnt-instance list``
1582 - ``gnt-instance list-fields``
1583 - ``gnt-debug locks``
1588 - New RAPI resources (see :doc:`rapi`):
1592 - ``/2/groups/[group_name]``
1593 - ``/2/groups/[group_name]/assign-nodes``
1594 - ``/2/groups/[group_name]/modify``
1595 - ``/2/groups/[group_name]/rename``
1596 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1600 - Implemented ``no_install`` for instance creation
1601 - Implemented OS parameters for instance reinstallation, allowing
1602 use of special settings on reinstallation (e.g. for preserving data)
1607 - Added IPv6 support in import/export
1608 - Pause DRBD synchronization while wiping disks on instance creation
1609 - Updated unittests and QA scripts
1610 - Improved network parameters passed to KVM
1611 - Converted man pages from docbook to reStructuredText
1617 *(Released Mon, 20 Dec 2010)*
1619 Released version 2.3.1~rc1 without any changes.
1625 *(Released Wed, 1 Dec 2010)*
1627 - impexpd: Disable OpenSSL compression in socat if possible (backport
1628 from master, commit e90739d625b, see :doc:`installation guide
1629 <install-quick>` for details)
1630 - Changed unittest coverage report to exclude test scripts
1631 - Added script to check version format
1637 *(Released Wed, 1 Dec 2010)*
1639 Released version 2.3.0~rc1 without any changes.
1645 *(Released Fri, 19 Nov 2010)*
1647 A number of bugfixes and documentation updates:
1649 - Update ganeti-os-interface documentation
1650 - Fixed a bug related to duplicate MACs or similar items which should be
1652 - Fix breakage in OS state modify
1653 - Reinstall instance: disallow offline secondaries (fixes bug related to
1654 OS changing but reinstall failing)
1655 - plus all the other fixes between 2.2.1 and 2.2.2
1661 *(Released Tue, 2 Nov 2010)*
1663 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1664 - Fixed master failover race with watcher
1665 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1667 - Accept previously stopped instance for export with instance removal
1668 - Simplify and extend the environment variables for instance OS scripts
1669 - Added new node flags, ``master_capable`` and ``vm_capable``
1670 - Added optional instance disk wiping prior during allocation. This is a
1671 cluster-wide option and can be set/modified using
1672 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1673 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1674 :doc:`install-quick`
1675 - Added a new watcher option (``--ignore-pause``)
1676 - Added option to ignore offline node on instance start/stop
1677 (``--ignore-offline``)
1678 - Allow overriding OS parameters with ``gnt-instance reinstall``
1679 - Added ability to change node's secondary IP address using ``gnt-node
1681 - Implemented privilege separation for all daemons except
1682 ``ganeti-noded``, see ``configure`` options
1683 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1685 - Implemented job priorities (see ``ganeti(7)`` manpage)
1686 - Ignore failures while shutting down instances during failover from
1688 - Exit daemon's bootstrap process only once daemon is ready
1689 - Export more information via ``LUInstanceQuery``/remote API
1690 - Improved documentation, QA and unittests
1691 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1692 restart if the file was created or changed
1693 - Added LUXI protocol version sent with each request and response,
1694 allowing detection of server/client mismatches
1695 - Moved the Python scripts among gnt-* and ganeti-* into modules
1696 - Moved all code related to setting up SSH to an external script,
1698 - Infrastructure changes for node group support in future versions
1704 *(Released Fri, 19 Nov 2010)*
1706 A few small bugs fixed, and some improvements to the build system:
1708 - Fix documentation regarding conversion to drbd
1709 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1711 - Fix error handling in node modify with multiple changes
1712 - Allow remote imports without checked names
1718 *(Released Tue, 19 Oct 2010)*
1720 - Disable SSL session ID cache in RPC client
1726 *(Released Thu, 14 Oct 2010)*
1728 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1729 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1731 - Fix problems with interaction between Curl and Python's HTTP server,
1732 resulting in increased speed in many RPC calls
1733 - Improve our release script to prevent breakage with older aclocal and
1740 *(Released Thu, 7 Oct 2010)*
1742 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1743 value retrieved from master
1744 - Added support for blacklisted or hidden OS definitions
1745 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1746 - Added support for -mem-path in KVM hypervisor abstraction layer
1747 - Allow overriding instance parameters in tool for inter-cluster
1748 instance moves (``tools/move-instance``)
1749 - Improved opcode summaries (e.g. in ``gnt-job list``)
1750 - Improve consistency of OS listing by sorting it
1751 - Documentation updates
1757 *(Released Fri, 8 Oct 2010)*
1759 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1765 *(Released Mon, 4 Oct 2010)*
1767 - Fixed regression in ``gnt-instance rename``
1773 *(Released Wed, 22 Sep 2010)*
1775 - Fixed OS_VARIANT variable for OS scripts
1776 - Fixed cluster tag operations via RAPI
1777 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1778 - Disabled RAPI CA checks in watcher
1784 *(Released Mon, 23 Aug 2010)*
1786 - Support DRBD versions of the format "a.b.c.d"
1788 - Re-introduce support for usage from multiple threads in RAPI client
1789 - Instance renames and modify via RAPI
1790 - Work around race condition between processing and archival in job
1792 - Mark opcodes following failed one as failed, too
1793 - Job field ``lock_status`` was removed due to difficulties making it
1794 work with the changed job queue in Ganeti 2.2; a better way to monitor
1795 locks is expected for a later 2.2.x release
1796 - Fixed dry-run behaviour with many commands
1797 - Support ``ssh-agent`` again when adding nodes
1798 - Many additional bugfixes
1804 *(Released Fri, 30 Jul 2010)*
1806 Important change: the internal RPC mechanism between Ganeti nodes has
1807 changed from using a home-grown http library (based on the Python base
1808 libraries) to use the PycURL library. This requires that PycURL is
1809 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1810 against GnuTLS by default. cURL's support for GnuTLS had known issues
1811 before cURL 7.21.0 and we recommend using the latest cURL release or
1812 linking against OpenSSL. Most other distributions already link PycURL
1813 and cURL against OpenSSL. The command::
1815 python -c 'import pycurl; print pycurl.version'
1817 can be used to determine the libraries PycURL and cURL are linked
1820 Other significant changes:
1822 - Rewrote much of the internals of the job queue, in order to achieve
1823 better parallelism; this decouples job query operations from the job
1824 processing, and it should allow much nicer behaviour of the master
1825 daemon under load, and it also has uncovered some long-standing bugs
1826 related to the job serialisation (now fixed)
1827 - Added a default iallocator setting to the cluster parameters,
1828 eliminating the need to always pass nodes or an iallocator for
1829 operations that require selection of new node(s)
1830 - Added experimental support for the LXC virtualization method
1831 - Added support for OS parameters, which allows the installation of
1832 instances to pass parameter to OS scripts in order to customise the
1834 - Added a hypervisor parameter controlling the migration type (live or
1835 non-live), since hypervisors have various levels of reliability; this
1836 has renamed the 'live' parameter to 'mode'
1837 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1838 logical volumes, meaning that cluster verify will ignore them and not
1839 flag their presence as errors
1840 - The watcher will now reset the error count for failed instances after
1841 8 hours, thus allowing self-healing if the problem that caused the
1842 instances to be down/fail to start has cleared in the meantime
1843 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1844 check for, and warn, if the drbd module parameter ``usermode_helper``
1845 is not consistent with the cluster-wide setting; this is needed to
1846 make diagnose easier of failed drbd creations
1847 - Started adding base IPv6 support, but this is not yet
1848 enabled/available for use
1849 - Rename operations (cluster, instance) will now return the new name,
1850 which is especially useful if a short name was passed in
1851 - Added support for instance migration in RAPI
1852 - Added a tool to pre-configure nodes for the SSH setup, before joining
1853 them to the cluster; this will allow in the future a simplified model
1854 for node joining (but not yet fully enabled in 2.2); this needs the
1855 paramiko python library
1856 - Fixed handling of name-resolving errors
1857 - Fixed consistency of job results on the error path
1858 - Fixed master-failover race condition when executed multiple times in
1860 - Fixed many bugs related to the job queue (mostly introduced during the
1861 2.2 development cycle, so not all are impacting 2.1)
1862 - Fixed instance migration with missing disk symlinks
1863 - Fixed handling of unknown jobs in ``gnt-job archive``
1864 - And many other small fixes/improvements
1868 - Enhanced both the unittest and the QA coverage
1869 - Switched the opcode validation to a generic model, and extended the
1870 validation to all opcode parameters
1871 - Changed more parts of the code that write shell scripts to use the
1873 - Switched the master daemon to use the asyncore library for the Luxi
1880 *(Released Thu, 17 Jun 2010)*
1882 - Added tool (``move-instance``) and infrastructure to move instances
1883 between separate clusters (see :doc:`separate documentation
1884 <move-instance>` and :doc:`design document <design-2.2>`)
1885 - Added per-request RPC timeout
1886 - RAPI now requires a Content-Type header for requests with a body (e.g.
1887 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1888 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1889 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1891 - Implemented initial support for running Ganeti daemons as separate
1892 users, see configure-time flags ``--with-user-prefix`` and
1893 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1895 - Instances can be removed after export (``gnt-backup export
1896 --remove-instance``)
1897 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1898 key (instead of 1024 bit)
1899 - Added new cluster configuration file for cluster domain secret
1900 - Import/export now use SSL instead of SSH
1901 - Added support for showing estimated time when exporting an instance,
1902 see the ``ganeti-os-interface(7)`` manpage and look for
1909 *(Released Tue, 16 Nov 2010)*
1911 Some more bugfixes. Unless critical bugs occur, this will be the last
1914 - Fix case of MAC special-values
1915 - Fix mac checker regex
1916 - backend: Fix typo causing "out of range" error
1917 - Add missing --units in gnt-instance list man page
1923 *(Released Tue, 24 Aug 2010)*
1926 - Don't ignore secondary node silently on non-mirrored disk templates
1928 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1929 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1930 - Properly document the usb_mouse parameter
1931 - Fix path in ganeti-rapi(8) (issue 116)
1932 - Adjust error message when the ganeti user's .ssh directory is
1934 - Add same-node-check when changing the disk template to drbd
1940 *(Released Fri, 16 Jul 2010)*
1943 - Add an option to only select some reboot types during qa/burnin.
1944 (on some hypervisors consequent reboots are not supported)
1945 - Fix infrequent race condition in master failover. Sometimes the old
1946 master ip address would be still detected as up for a short time
1947 after it was removed, causing failover to fail.
1948 - Decrease mlockall warnings when the ctypes module is missing. On
1949 Python 2.4 we support running even if no ctypes module is installed,
1950 but we were too verbose about this issue.
1951 - Fix building on old distributions, on which man doesn't have a
1953 - Fix RAPI not to ignore the MAC address on instance creation
1954 - Implement the old instance creation format in the RAPI client.
1960 *(Released Thu, 01 Jul 2010)*
1962 A small bugfix release:
1963 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1964 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1965 a non-existing option
1966 - Fix instance create: the --force-variant option was ignored
1967 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1968 - Fix modify node storage with non-FQDN arguments
1969 - Fix RAPI client to authenticate under Python 2.6 when used
1970 for more than 5 requests needing authentication
1971 - Fix gnt-instance modify -t (storage) giving a wrong error message
1972 when converting a non-shutdown drbd instance to plain
1978 *(Released Fri, 18 Jun 2010)*
1980 A small bugfix release:
1982 - Fix live migration of KVM instances started with older Ganeti
1983 versions which had fewer hypervisor parameters
1984 - Fix gnt-instance grow-disk on down instances
1985 - Fix an error-reporting bug during instance migration
1986 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1987 silently ignoring broken ones
1988 - Fix an RPC error reporting bug affecting, for example, RAPI client
1990 - Fix bug triggered by different API version os-es on different nodes
1991 - Fix a bug in instance startup with custom hvparams: OS level
1992 parameters would fail to be applied.
1993 - Fix the RAPI client under Python 2.6 (but more work is needed to
1994 make it work completely well with OpenSSL)
1995 - Fix handling of errors when resolving names from DNS
2001 *(Released Thu, 3 Jun 2010)*
2003 A medium sized development cycle. Some new features, and some
2004 fixes/small improvements/cleanups.
2006 Significant features
2007 ~~~~~~~~~~~~~~~~~~~~
2009 The node deamon now tries to mlock itself into memory, unless the
2010 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2011 its logs, and falls back to console logging. This allows emergency
2012 features such as ``gnt-node powercycle`` to work even in the event of a
2013 broken node disk (tested offlining the disk hosting the node's
2014 filesystem and dropping its memory caches; don't try this at home)
2016 KVM: add vhost-net acceleration support. It can be tested with a new
2017 enough version of the kernel and of qemu-kvm.
2019 KVM: Add instance chrooting feature. If you use privilege dropping for
2020 your VMs you can also now force them to chroot to an empty directory,
2021 before starting the emulated guest.
2023 KVM: Add maximum migration bandwith and maximum downtime tweaking
2024 support (requires a new-enough version of qemu-kvm).
2026 Cluster verify will now warn if the master node doesn't have the master
2027 ip configured on it.
2029 Add a new (incompatible) instance creation request format to RAPI which
2030 supports all parameters (previously only a subset was supported, and it
2031 wasn't possible to extend the old format to accomodate all the new
2032 features. The old format is still supported, and a client can check for
2033 this feature, before using it, by checking for its presence in the
2034 ``features`` RAPI resource.
2036 Now with ancient latin support. Try it passing the ``--roman`` option to
2037 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2038 (requires the python-roman module to be installed, in order to work).
2043 As usual many internal code refactorings, documentation updates, and
2046 - Lots of improvements and cleanups to the experimental Remote API
2047 (RAPI) client library.
2048 - A new unit test suite for the core daemon libraries.
2049 - A fix to creating missing directories makes sure the umask is not
2050 applied anymore. This enforces the same directory permissions
2052 - Better handling terminating daemons with ctrl+c (used when running
2053 them in debugging mode).
2054 - Fix a race condition in live migrating a KVM instance, when stat()
2055 on the old proc status file returned EINVAL, which is an unexpected
2057 - Fixed manpage checking with newer man and utf-8 charachters. But now
2058 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2064 *(Released Fri, 7 May 2010)*
2066 Fix a bug which prevented untagged KVM instances from starting.
2072 *(Released Fri, 7 May 2010)*
2074 Another release with a long development cycle, during which many
2075 different features were added.
2077 Significant features
2078 ~~~~~~~~~~~~~~~~~~~~
2080 The KVM hypervisor now can run the individual instances as non-root, to
2081 reduce the impact of a VM being hijacked due to bugs in the
2082 hypervisor. It is possible to run all instances as a single (non-root)
2083 user, to manually specify a user for each instance, or to dynamically
2084 allocate a user out of a cluster-wide pool to each instance, with the
2085 guarantee that no two instances will run under the same user ID on any
2088 An experimental RAPI client library, that can be used standalone
2089 (without the other Ganeti libraries), is provided in the source tree as
2090 ``lib/rapi/client.py``. Note this client might change its interface in
2091 the future, as we iterate on its capabilities.
2093 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2094 replace the cluster's certificates and crypto keys. This might help in
2095 case they have been compromised, or have simply expired.
2097 A new disk option for instance creation has been added that allows one
2098 to "adopt" currently existing logical volumes, with data
2099 preservation. This should allow easier migration to Ganeti from
2100 unmanaged (or managed via other software) instances.
2102 Another disk improvement is the possibility to convert between redundant
2103 (DRBD) and plain (LVM) disk configuration for an instance. This should
2104 allow better scalability (starting with one node and growing the
2105 cluster, or shrinking a two-node cluster to one node).
2107 A new feature that could help with automated node failovers has been
2108 implemented: if a node sees itself as offline (by querying the master
2109 candidates), it will try to shutdown (hard) all instances and any active
2110 DRBD devices. This reduces the risk of duplicate instances if an
2111 external script automatically failovers the instances on such nodes. To
2112 enable this, the cluster parameter ``maintain_node_health`` should be
2113 enabled; in the future this option (per the name) will enable other
2114 automatic maintenance features.
2116 Instance export/import now will reuse the original instance
2117 specifications for all parameters; that means exporting an instance,
2118 deleting it and the importing it back should give an almost identical
2119 instance. Note that the default import behaviour has changed from
2120 before, where it created only one NIC; now it recreates the original
2123 Cluster verify has added a few new checks: SSL certificates validity,
2124 /etc/hosts consistency across the cluster, etc.
2129 As usual, many internal changes were done, documentation fixes,
2132 - Fixed cluster initialization with disabled cluster storage (regression
2133 introduced in 2.1.1)
2134 - File-based storage supports growing the disks
2135 - Fixed behaviour of node role changes
2136 - Fixed cluster verify for some corner cases, plus a general rewrite of
2137 cluster verify to allow future extension with more checks
2138 - Fixed log spamming by watcher and node daemon (regression introduced
2140 - Fixed possible validation issues when changing the list of enabled
2142 - Fixed cleanup of /etc/hosts during node removal
2143 - Fixed RAPI response for invalid methods
2144 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2145 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2146 from ide disks, etc.)
2147 - Allow OS changes without re-installation (to record a changed OS
2148 outside of Ganeti, or to allow OS renames)
2149 - Allow instance creation without OS installation (useful for example if
2150 the OS will be installed manually, or restored from a backup not in
2152 - Implemented option to make cluster ``copyfile`` use the replication
2154 - Added list of enabled hypervisors to ssconf (possibly useful for
2156 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2158 - A partial form of node re-IP is possible via node readd, which now
2159 allows changed node primary IP
2160 - Command line utilities now show an informational message if the job is
2162 - The logs of the master daemon now show the PID/UID/GID of the
2169 *(Released Fri, 12 Mar 2010)*
2171 During the 2.1.0 long release candidate cycle, a lot of improvements and
2172 changes have accumulated with were released later as 2.1.1.
2177 The node evacuate command (``gnt-node evacuate``) was significantly
2178 rewritten, and as such the IAllocator protocol was changed - a new
2179 request type has been added. This unfortunate change during a stable
2180 series is designed to improve performance of node evacuations; on
2181 clusters with more than about five nodes and which are well-balanced,
2182 evacuation should proceed in parallel for all instances of the node
2183 being evacuated. As such, any existing IAllocator scripts need to be
2184 updated, otherwise the above command will fail due to the unknown
2185 request. The provided "dumb" allocator has not been updated; but the
2186 ganeti-htools package supports the new protocol since version 0.2.4.
2188 Another important change is increased validation of node and instance
2189 names. This might create problems in special cases, if invalid host
2190 names are being used.
2192 Also, a new layer of hypervisor parameters has been added, that sits at
2193 OS level between the cluster defaults and the instance ones. This allows
2194 customisation of virtualization parameters depending on the installed
2195 OS. For example instances with OS 'X' may have a different KVM kernel
2196 (or any other parameter) than the cluster defaults. This is intended to
2197 help managing a multiple OSes on the same cluster, without manual
2198 modification of each instance's parameters.
2200 A tool for merging clusters, ``cluster-merge``, has been added in the
2201 tools sub-directory.
2206 - Improved the int/float conversions that should make the code more
2207 robust in face of errors from the node daemons
2208 - Fixed the remove node code in case of internal configuration errors
2209 - Fixed the node daemon behaviour in face of inconsistent queue
2210 directory (e.g. read-only file-system where we can't open the files
2212 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2213 now it either aborts cleanly or, if given the new "auto_promote"
2214 parameter, will automatically promote other nodes as needed
2215 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2216 completely prevent Ganeti from working
2217 - Fixed bug for instance export when not all disks were successfully
2219 - Fixed behaviour of node add when the new node is slow in starting up
2221 - Fixed handling of signals in the LUXI client, which should improve
2222 behaviour of command-line scripts
2223 - Added checks for invalid node/instance names in the configuration (now
2224 flagged during cluster verify)
2225 - Fixed watcher behaviour for disk activation errors
2226 - Fixed two potentially endless loops in http library, which led to the
2227 RAPI daemon hanging and consuming 100% CPU in some cases
2228 - Fixed bug in RAPI daemon related to hashed passwords
2229 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2231 - Enhanced compatibility with non-Debian OSes, but not using absolute
2232 path in some commands and allowing customisation of the ssh
2233 configuration directory
2234 - Fixed possible future issue with new Python versions by abiding to the
2235 proper use of ``__slots__`` attribute on classes
2236 - Added checks that should prevent directory traversal attacks
2237 - Many documentation fixes based on feedback from users
2242 - Added an "early_release" more for instance replace disks and node
2243 evacuate, where we release locks earlier and thus allow higher
2244 parallelism within the cluster
2245 - Added watcher hooks, intended to allow the watcher to restart other
2246 daemons (e.g. from the ganeti-nbma project), but they can be used of
2247 course for any other purpose
2248 - Added a compile-time disable for DRBD barriers, to increase
2249 performance if the administrator trusts the power supply or the
2250 storage system to not lose writes
2251 - Added the option of using syslog for logging instead of, or in
2252 addition to, Ganeti's own log files
2253 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2254 can indeed boot from a paravirtual NIC
2255 - Added a generic debug level for many operations; while this is not
2256 used widely yet, it allows one to pass the debug value all the way to
2258 - Enhanced the hooks environment for instance moves (failovers,
2259 migrations) where the primary/secondary nodes changed during the
2260 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2261 - Enhanced data validations for many user-supplied values; one important
2262 item is the restrictions imposed on instance and node names, which
2263 might reject some (invalid) host names
2264 - Add a configure-time option to disable file-based storage, if it's not
2265 needed; this allows greater security separation between the master
2266 node and the other nodes from the point of view of the inter-node RPC
2268 - Added user notification in interactive tools if job is waiting in the
2269 job queue or trying to acquire locks
2270 - Added log messages when a job is waiting for locks
2271 - Added filtering by node tags in instance operations which admit
2272 multiple instances (start, stop, reboot, reinstall)
2273 - Added a new tool for cluster mergers, ``cluster-merge``
2274 - Parameters from command line which are of the form ``a=b,c=d`` can now
2275 use backslash escapes to pass in values which contain commas,
2276 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2278 - For KVM, the instance name is the first parameter passed to KVM, so
2279 that it's more visible in the process list
2285 *(Released Tue, 2 Mar 2010)*
2287 Ganeti 2.1 brings many improvements with it. Major changes:
2289 - Added infrastructure to ease automated disk repairs
2290 - Added new daemon to export configuration data in a cheaper way than
2291 using the remote API
2292 - Instance NICs can now be routed instead of being associated with a
2294 - Improved job locking logic to reduce impact of jobs acquiring multiple
2295 locks waiting for other long-running jobs
2297 In-depth implementation details can be found in the Ganeti 2.1 design
2303 - Added chroot hypervisor
2304 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2306 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2307 ``bootloader_path`` and ``bootloader_args``)
2308 - Added the ``use_localtime`` option for the xen-hvm and kvm
2309 hypervisors, and the default value for this has changed to false (in
2310 2.0 xen-hvm always enabled it)
2311 - Added luxi call to submit multiple jobs in one go
2312 - Added cluster initialization option to not modify ``/etc/hosts``
2314 - Added network interface parameters
2315 - Added dry run mode to some LUs
2316 - Added RAPI resources:
2318 - ``/2/instances/[instance_name]/info``
2319 - ``/2/instances/[instance_name]/replace-disks``
2320 - ``/2/nodes/[node_name]/evacuate``
2321 - ``/2/nodes/[node_name]/migrate``
2322 - ``/2/nodes/[node_name]/role``
2323 - ``/2/nodes/[node_name]/storage``
2324 - ``/2/nodes/[node_name]/storage/modify``
2325 - ``/2/nodes/[node_name]/storage/repair``
2327 - Added OpCodes to evacuate or migrate all instances on a node
2328 - Added new command to list storage elements on nodes (``gnt-node
2329 list-storage``) and modify them (``gnt-node modify-storage``)
2330 - Added new ssconf files with master candidate IP address
2331 (``ssconf_master_candidates_ips``), node primary IP address
2332 (``ssconf_node_primary_ips``) and node secondary IP address
2333 (``ssconf_node_secondary_ips``)
2334 - Added ``ganeti-confd`` and a client library to query the Ganeti
2335 configuration via UDP
2336 - Added ability to run hooks after cluster initialization and before
2338 - Added automatic mode for disk replace (``gnt-instance replace-disks
2340 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2341 after catastrophic data-loss
2342 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2344 - Added ``gnt-instance move`` command to move instances
2345 - Added ``gnt-cluster watcher`` command to control watcher
2346 - Added ``gnt-node powercycle`` command to powercycle nodes
2347 - Added new job status field ``lock_status``
2348 - Added parseable error codes to cluster verification (``gnt-cluster
2349 verify --error-codes``) and made output less verbose (use
2350 ``--verbose`` to restore previous behaviour)
2351 - Added UUIDs to the main config entities (cluster, nodes, instances)
2352 - Added support for OS variants
2353 - Added support for hashed passwords in the Ganeti remote API users file
2355 - Added option to specify maximum timeout on instance shutdown
2356 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2357 - Added new helper script to start and stop Ganeti daemons
2358 (``daemon-util``), with the intent to reduce the work necessary to
2359 adjust Ganeti for non-Debian distributions and to start/stop daemons
2361 - Added more unittests
2362 - Fixed critical bug in ganeti-masterd startup
2363 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2364 now customisable at the cluster level for both the KVM and Xen
2365 hypervisors using the new ``migration_port`` parameter
2366 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2367 reinstalling an instance
2368 - Allowed ``@`` in tag names
2369 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2370 - Many documentation updates
2371 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2372 - ``gnt-instance reinstall`` can now reinstall multiple instances
2373 - Updated many command line parameters
2374 - Introduced new OS API version 15
2375 - No longer support a default hypervisor
2376 - Treat virtual LVs as inexistent
2377 - Improved job locking logic to reduce lock contention
2378 - Match instance and node names case insensitively
2379 - Reimplemented bash completion script to be more complete
2386 *(Released Thu, 4 Feb 2010)*
2388 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2389 - Fix a string formatting bug
2390 - Improve safety of the code in some error paths
2391 - Improve data validation in the master of values returned from nodes
2397 *(Released Thu, 17 Dec 2009)*
2399 - Fix security issue due to missing validation of iallocator names; this
2400 allows local and remote execution of arbitrary executables
2401 - Fix failure of gnt-node list during instance removal
2402 - Ship the RAPI documentation in the archive
2408 *(Released Wed, 30 Sep 2009)*
2410 - Fixed many wrong messages
2411 - Fixed a few bugs related to the locking library
2412 - Fixed MAC checking at instance creation time
2413 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2414 - Fixed a few issues related to signal handling in both daemons and
2416 - Fixed the example startup script provided
2417 - Fixed insserv dependencies in the example startup script (patch from
2419 - Fixed handling of drained nodes in the iallocator framework
2420 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2422 - Fixed error related to invalid job IDs in job polling
2423 - Fixed job/opcode persistence on unclean master shutdown
2424 - Fixed handling of partial job processing after unclean master
2426 - Fixed error reporting from LUs, previously all errors were converted
2427 into execution errors
2428 - Fixed error reporting from burnin
2429 - Decreased significantly the memory usage of the job queue
2430 - Optimised slightly multi-job submission
2431 - Optimised slightly opcode loading
2432 - Backported the multi-job submit framework from the development
2433 branch; multi-instance start and stop should be faster
2434 - Added script to clean archived jobs after 21 days; this will reduce
2435 the size of the queue directory
2436 - Added some extra checks in disk size tracking
2437 - Added an example ethers hook script
2438 - Added a cluster parameter that prevents Ganeti from modifying of
2440 - Added more node information to RAPI responses
2441 - Added a ``gnt-job watch`` command that allows following the ouput of a
2443 - Added a bind-address option to ganeti-rapi
2444 - Added more checks to the configuration verify
2445 - Enhanced the burnin script such that some operations can be retried
2447 - Converted instance reinstall to multi-instance model
2453 *(Released Fri, 7 Aug 2009)*
2455 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2456 to allow using the pre-2.0.2 behaviour in activation, if any existing
2457 instances have mismatched disk sizes in the configuration
2458 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2459 any configuration mismatches for disk sizes
2460 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2461 failover to work on two-node clusters
2462 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2464 - Fixed detection of OS script errors in ``gnt-backup export``
2465 - Fixed exit code of ``gnt-backup export``
2471 *(Released Fri, 17 Jul 2009)*
2473 - Added experimental support for stripped logical volumes; this should
2474 enhance performance but comes with a higher complexity in the block
2475 device handling; stripping is only enabled when passing
2476 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2477 affected even in the non-stripped mode
2478 - Improved resiliency against transient failures at the end of DRBD
2479 resyncs, and in general of DRBD resync checks
2480 - Fixed a couple of issues with exports and snapshot errors
2481 - Fixed a couple of issues in instance listing
2482 - Added display of the disk size in ``gnt-instance info``
2483 - Fixed checking for valid OSes in instance creation
2484 - Fixed handling of the "vcpus" parameter in instance listing and in
2485 general of invalid parameters
2486 - Fixed http server library, and thus RAPI, to handle invalid
2487 username/password combinations correctly; this means that now they
2488 report unauthorized for queries too, not only for modifications,
2489 allowing earlier detect of configuration problems
2490 - Added a new "role" node list field, equivalent to the master/master
2491 candidate/drained/offline flags combinations
2492 - Fixed cluster modify and changes of candidate pool size
2493 - Fixed cluster verify error messages for wrong files on regular nodes
2494 - Fixed a couple of issues with node demotion from master candidate role
2495 - Fixed node readd issues
2496 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2497 - Added a new ``--no-voting`` option for masterfailover to fix failover
2498 on two-nodes clusters when the former master node is unreachable
2499 - Added instance reinstall over RAPI
2505 *(Released Tue, 16 Jun 2009)*
2507 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2508 allow re-adding the start in single-user option (regression from 1.2)
2509 - the watcher writes the instance status to a file, to allow monitoring
2510 to report the instance status (from the master) based on cached
2511 results of the watcher's queries; while this can get stale if the
2512 watcher is being locked due to other work on the cluster, this is
2513 still an improvement
2514 - the watcher now also restarts the node daemon and the rapi daemon if
2516 - fixed the watcher to handle full and drained queue cases
2517 - hooks export more instance data in the environment, which helps if
2518 hook scripts need to take action based on the instance's properties
2519 (no longer need to query back into ganeti)
2520 - instance failovers when the instance is stopped do not check for free
2521 RAM, so that failing over a stopped instance is possible in low memory
2523 - rapi uses queries for tags instead of jobs (for less job traffic), and
2524 for cluster tags it won't talk to masterd at all but read them from
2526 - a couple of error handling fixes in RAPI
2527 - drbd handling: improved the error handling of inconsistent disks after
2528 resync to reduce the frequency of "there are some degraded disks for
2529 this instance" messages
2530 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2531 error handling path called a wrong function name)
2537 *(Released Wed, 27 May 2009)*
2539 - no changes from rc5
2545 *(Released Wed, 20 May 2009)*
2547 - fix a couple of bugs (validation, argument checks)
2548 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2549 - some small improvements to RAPI and IAllocator
2550 - make watcher automatically start the master daemon if down
2556 *(Released Mon, 27 Apr 2009)*
2558 - change the OS list to not require locks; this helps with big clusters
2559 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2560 volume group is broken
2561 - ``gnt-instance info``, without any arguments, doesn't run for all
2562 instances anymore; either pass ``--all`` or pass the desired
2563 instances; this helps against mistakes on big clusters where listing
2564 the information for all instances takes a long time
2565 - miscellaneous doc and man pages fixes
2571 *(Released Wed, 8 Apr 2009)*
2573 - Change the internal locking model of some ``gnt-node`` commands, in
2574 order to reduce contention (and blocking of master daemon) when
2575 batching many creation/reinstall jobs
2576 - Fixes to Xen soft reboot
2577 - No longer build documentation at build time, instead distribute it in
2578 the archive, in order to reduce the need for the whole docbook/rst
2585 *(Released Fri, 27 Mar 2009)*
2587 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2588 - Fix watcher startup sequence, improves the behaviour of busy clusters
2589 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2590 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2591 bind address and other places
2592 - Some documentation fixes and updates
2598 *(Released Mon, 2 Mar 2009)*
2600 - More documentation updates, now all docs should be more-or-less
2602 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2604 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2606 - Fix an issue related to $libdir/run/ganeti and cluster creation
2612 *(Released Thu, 19 Feb 2009)*
2614 - Xen PVM and KVM have switched the default value for the instance root
2615 disk to the first partition on the first drive, instead of the whole
2616 drive; this means that the OS installation scripts must be changed
2618 - Man pages have been updated
2619 - RAPI has been switched by default to HTTPS, and the exported functions
2620 should all work correctly
2621 - RAPI v1 has been removed
2622 - Many improvements to the KVM hypervisor
2623 - Block device errors are now better reported
2624 - Many other bugfixes and small improvements
2630 *(Released Mon, 26 Jan 2009)*
2632 - Version 2 is a general rewrite of the code and therefore the
2633 differences are too many to list, see the design document for 2.0 in
2634 the ``doc/`` subdirectory for more details
2635 - In this beta version there is not yet a migration path from 1.2 (there
2636 will be one in the final 2.0 release)
2637 - A few significant changes are:
2639 - all commands are executed by a daemon (``ganeti-masterd``) and the
2640 various ``gnt-*`` commands are just front-ends to it
2641 - all the commands are entered into, and executed from a job queue,
2642 see the ``gnt-job(8)`` manpage
2643 - the RAPI daemon supports read-write operations, secured by basic
2644 HTTP authentication on top of HTTPS
2645 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2646 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2647 to migrate to DRBD 8 first while still running Ganeti 1.2)
2648 - DRBD devices are using statically allocated minor numbers, which
2649 will be assigned to existing instances during the migration process
2650 - there is support for both Xen PVM and Xen HVM instances running on
2652 - KVM virtualization is supported too
2653 - file-based storage has been implemented, which means that it is
2654 possible to run the cluster without LVM and DRBD storage, for
2655 example using a shared filesystem exported from shared storage (and
2656 still have live migration)
2662 *(Released Tue, 13 Jan 2009)*
2664 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2665 - Reuse the old instance mac address by default on instance import, if
2666 the instance name is the same.
2667 - Handle situations in which the node info rpc returns incomplete
2669 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2670 - Improved version of batcher:
2672 - state file support
2673 - instance mac address support
2674 - support for HVM clusters/instances
2676 - Add an option to show the number of cpu sockets and nodes in
2678 - Support OSes that handle more than one version of the OS api (but do
2679 not change the current API in any other way)
2680 - Fix ``gnt-node migrate``
2681 - ``gnt-debug`` man page
2682 - Fixes various more typos and small issues
2683 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2689 *(Released Wed, 24 Sep 2008)*
2691 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2692 type of disk exported to fully virtualized instances.
2693 - provide access to the serial console of HVM instances
2694 - instance auto_balance flag, set by default. If turned off it will
2695 avoid warnings on cluster verify if there is not enough memory to fail
2696 over an instance. in the future it will prevent automatically failing
2697 it over when we will support that.
2698 - batcher tool for instance creation, see ``tools/README.batcher``
2699 - ``gnt-instance reinstall --select-os`` to interactively select a new
2700 operating system when reinstalling an instance.
2701 - when changing the memory amount on instance modify a check has been
2702 added that the instance will be able to start. also warnings are
2703 emitted if the instance will not be able to fail over, if auto_balance
2705 - documentation fixes
2706 - sync fields between ``gnt-instance list/modify/add/import``
2707 - fix a race condition in drbd when the sync speed was set after giving
2708 the device a remote peer.
2714 *(Released Tue, 22 Jul 2008)*
2716 - note: the allowed size and number of tags per object were reduced
2717 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2718 - fixed twisted 8.x compatibility
2719 - fixed ``gnt-instance replace-disks`` with iallocator
2720 - add TCP keepalives on twisted connections to detect restarted nodes
2721 - disk increase support, see ``gnt-instance grow-disk``
2722 - implement bulk node/instance query for RAPI
2723 - add tags in node/instance listing (optional)
2724 - experimental migration (and live migration) support, read the man page
2725 for ``gnt-instance migrate``
2726 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2727 has some small improvements in handling its state file
2733 *(Released Fri, 13 Jun 2008)*
2735 - Experimental readonly, REST-based remote API implementation;
2736 automatically started on master node, TCP port 5080, if enabled by
2737 ``--enable-rapi`` parameter to configure script.
2738 - Instance allocator support. Add and import instance accept a
2739 ``--iallocator`` parameter, and call that instance allocator to decide
2740 which node to use for the instance. The iallocator document describes
2741 what's expected from an allocator script.
2742 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2743 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2744 node is lost there is still enough memory to fail over the instances
2746 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2747 ``gnt-cluster verify``, to check for site-specific compliance. All the
2748 hooks will run, and their output, if any, will be displayed. Any
2749 failing hook will make the verification return an error value.
2750 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2751 primary and secondary interfaces
2752 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2753 that is still declared as part of the cluster and has failed.
2754 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2755 specifying output fields, that just adds the chosen fields to the
2757 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2758 export from the filesystem.
2759 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2760 have been added. Using them you can enable/disable acpi and pae
2761 support, and specify a path for a cd image to be exported to the
2762 instance. These parameters as the name suggest only work on HVM
2764 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2765 PAE support will be set to the previously hardcoded values, but the
2766 (previously hardcoded) path to the CDROM ISO image will be unset and
2767 if required, needs to be set manually with ``gnt-instance modify``
2769 - The address to which an instance's VNC console is bound is now
2770 selectable per-instance, rather than being cluster wide. Of course
2771 this only applies to instances controlled via VNC, so currently just
2772 applies to HVM clusters.
2778 *(Released Mon, 18 Feb 2008)*
2780 - more tweaks to the disk activation code (especially helpful for DRBD)
2781 - change the default ``gnt-instance list`` output format, now there is
2782 one combined status field (see the manpage for the exact values this
2784 - some more fixes for the mac export to hooks change
2785 - make Ganeti not break with DRBD 8.2.x (which changed the version
2786 format in ``/proc/drbd``) (issue 24)
2787 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2788 template, allowing migration from DRBD0.7+MD to DRBD8
2794 *(Released Wed, 30 Jan 2008)*
2796 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2798 - add command aliases infrastructure and a few aliases
2799 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2800 man pages and the ``--help`` option of ``gnt-node
2801 list``/``gnt-instance list``
2802 - fix ``gnt-backup list`` with down nodes (issue 21)
2803 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2804 - fix the dist archive and add a check for including svn/git files in
2806 - some developer-related changes: improve the burnin and the QA suite,
2807 add an upload script for testing during development
2813 *(Released Wed, 16 Jan 2008)*
2815 - experimental HVM support, read the install document, section
2816 "Initializing the cluster"
2817 - allow for the PVM hypervisor per-instance kernel and initrd paths
2818 - add a new command ``gnt-cluster verify-disks`` which uses a new
2819 algorithm to improve the reconnection of the DRBD pairs if the device
2820 on the secondary node has gone away
2821 - make logical volume code auto-activate LVs at disk activation time
2822 - slightly improve the speed of activating disks
2823 - allow specification of the MAC address at instance creation time, and
2824 changing it later via ``gnt-instance modify``
2825 - fix handling of external commands that generate lots of output on
2827 - update documentation with regard to minimum version of DRBD8 supported
2833 *(Released Tue, 4 Dec 2007)*
2835 - Log the ``xm create`` output to the node daemon log on failure (to
2836 help diagnosing the error)
2837 - In debug mode, log all external commands output if failed to the logs
2838 - Change parsing of lvm commands to ignore stderr
2844 *(Released Wed, 28 Nov 2007)*
2846 - Another round of updates to the DRBD 8 code to deal with more failures
2847 in the replace secondary node operation
2848 - Some more logging of failures in disk operations (lvm, drbd)
2849 - A few documentation updates
2856 *(Released Tue, 13 Nov 2007)*
2858 - Change configuration file format from Python's Pickle to JSON.
2859 Upgrading is possible using the cfgupgrade utility.
2860 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2861 faster replace disks and is more stable (DRBD 8 has many improvements
2862 compared to DRBD 0.7)
2863 - Added command line tags support (see man pages for ``gnt-instance``,
2864 ``gnt-node``, ``gnt-cluster``)
2865 - Added instance rename support
2866 - Added multi-instance startup/shutdown
2867 - Added cluster rename support
2868 - Added ``gnt-node evacuate`` to simplify some node operations
2869 - Added instance reboot operation that can speedup reboot as compared to
2871 - Soften the requirement that hostnames are in FQDN format
2872 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2874 - Removed dependency on debian's patched fping that uses the
2875 non-standard ``-S`` option
2876 - Now the OS definitions are searched for in multiple, configurable
2877 paths (easier for distros to package)
2878 - Some changes to the hooks infrastructure (especially the new
2879 post-configuration update hook)
2880 - Other small bugfixes
2882 .. vim: set textwidth=72 syntax=rst :