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.
36 - The list of allowed disk templates in the instance policy and the list
37 of cluster-wide enabled disk templates is now checked for consistency
38 on cluster or group modification. On cluster initialization, the ipolicy
39 disk templates are ensured to be a subset of the cluster-wide enabled
45 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
46 the correct command syntax. It is possible to use different DRBD versions
47 on different nodes as long as they are compatible to each other. This
48 enables rolling upgrades of DRBD with no downtime. As permanent operation
49 of different DRBD versions within a node group is discouraged,
50 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
51 - New "inst-status-xen" data collector for the monitoring daemon, providing
52 information about the state of the xen instances on the nodes.
53 - New "lv" data collector for the monitoring daemon, collecting data about the
54 logical volumes on the nodes, and pairing them with the name of the instances
56 - New "diskstats" data collector, collecting the data from /proc/diskstats and
57 presenting them over the monitoring daemon interface.
58 - The ConfD client is now IPv6 compatible.
62 The following new dependencies have been added.
66 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
67 for the unit tests (and only used for testing).
73 *(Released Fri, 2 Aug 2013)*
75 Incompatible/important changes
76 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
78 - Instance policy can contain multiple instance specs, as described in
79 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
80 <design-partitioned>`. As a consequence, it's not possible to partially change
81 or override instance specs. Bounding specs (min and max) can be specified as a
82 whole using the new option ``--ipolicy-bounds-specs``, while standard
83 specs use the new option ``--ipolicy-std-specs``.
84 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
85 gnt-instance is a valid YAML object.
86 - hail now honors network restrictions when allocating nodes. This led to an
87 update of the IAllocator protocol. See the IAllocator documentation for
89 - confd now only answers static configuration request over the network. luxid
90 was extracted, listens on the local LUXI socket and responds to live queries.
91 This allows finer grained permissions if using separate users.
96 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
97 to always require authentication, ``--require-authentication``. It can
98 be specified in ``$sysconfdir/default/ganeti``.
99 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
100 be used to manage the disk templates to be used by instances in the cluster.
101 Initially, it will be set to a list that includes plain, drbd, if they were
102 enabled by specifying a volume group name, and file and sharedfile, if those
103 were enabled at configure time. Additionally, it will include all disk
104 templates that are currently used by instances. The order of disk templates
105 will be based on Ganeti's history of supporting them. In the future, the
106 first entry of the list will be used as a default disk template on instance
108 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
109 configuration back to the previous stable version.
110 - Disk templates in group ipolicy can be restored to the default value.
111 - Initial support for diskless instances and virtual clusters in QA.
112 - More QA and unit tests for instance policies.
113 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
114 describing why the opcode itself was executed.
115 - The monitoring daemon is now available. It allows users to query the cluster
116 for obtaining information about the status of the system. The daemon is only
117 responsible for providing the information over the network: the actual data
118 gathering is performed by data collectors (currently, only the DRBD status
119 collector is available).
120 - In order to help developers work on Ganeti, a new script
121 (``devel/build_chroot``) is provided, for building a chroot that contains all
122 the required development libraries and tools for compiling Ganeti on a Debian
124 - A new tool, ``harep``, for performing self-repair and recreation of instances
125 in Ganeti has been added.
126 - Split queries are enabled for tags, network, exports, cluster info, groups,
128 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
129 It prints the command line to set the current policies, to ease
131 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
132 devices for KVM virtio-net interfaces will get created with VNET_HDR
133 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
134 virtio-net interfaces, which prevents host kernel tainting and log
135 flooding, when dealing with broken or malicious virtio-net drivers.
136 It's set to true by default.
137 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
142 The following new dependencies have been added:
145 - The ``curl`` library is not optional anymore for compiling the Haskell code.
146 - ``snap-server`` library (if monitoring is enabled).
149 - The minimum Python version needed to run Ganeti is now 2.6.
150 - ``yaml`` library (only for running the QA).
155 - Fix upgrading/downgrading from 2.7
156 - Increase maximum RAPI message size
157 - Documentation updates
158 - Split ``confd`` between ``luxid`` and ``confd``
159 - Merge 2.7 series up to the 2.7.1 release
160 - Allow the ``modify_etc_hosts`` option to be changed
161 - Add better debugging for ``luxid`` queries
162 - Expose bulk parameter for GetJobs in RAPI client
163 - Expose missing ``network`` fields in RAPI
164 - Add some ``cluster verify`` tests
165 - Some unittest fixes
166 - Fix a malfunction in ``hspace``'s tiered allocation
167 - Fix query compatibility between haskell and python implementations
168 - Add the ``vnet_hdr`` HV parameter for KVM
169 - Add ``--cleanup`` to instance failover
170 - Change the connected groups format in ``gnt-network info`` output; it
171 was previously displayed as a raw list by mistake. (Merged from 2.7)
177 *(Released Mon, 24 Jun 2013)*
179 This was the first beta release of the 2.8 series. All important changes
180 are listed in the latest 2.8 entry.
186 *(Released Thu, 25 Jul 2013)*
188 - Add logrotate functionality in daemon-util
189 - Add logrotate example file
190 - Add missing fields to network queries over rapi
191 - Fix network object timestamps
192 - Add support for querying network timestamps
193 - Fix a typo in the example crontab
194 - Fix a documentation typo
200 *(Released Thu, 04 Jul 2013)*
202 Incompatible/important changes
203 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
205 - Instance policies for disk size were documented to be on a per-disk
206 basis, but hail applied them to the sum of all disks. This has been
208 - ``hbal`` will now exit with status 0 if, during job execution over
209 LUXI, early exit has been requested and all jobs are successful;
210 before, exit status 1 was used, which cannot be differentiated from
212 - Compatibility with newer versions of rbd has been fixed
213 - ``gnt-instance batch-create`` has been changed to use the bulk create
214 opcode from Ganeti. This lead to incompatible changes in the format of
215 the JSON file. It's now not a custom dict anymore but a dict
216 compatible with the ``OpInstanceCreate`` opcode.
217 - Parent directories for file storage need to be listed in
218 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
219 write the file automatically based on old configuration values, but it
220 can not distribute it across all nodes and the file contents should be
221 verified. Use ``gnt-cluster copyfile
222 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
223 upgraded. The reason for requiring this list of paths now is that
224 before it would have been possible to inject new paths via RPC,
225 allowing files to be created in arbitrary locations. The RPC protocol
226 is protected using SSL/X.509 certificates, but as a design principle
227 Ganeti does not permit arbitrary paths to be passed.
228 - The parsing of the variants file for OSes (see
229 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
230 lines and comment lines (starting with ``#``) are ignored for better
232 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
233 longer available. ``gnt-node add`` now invokes a new tool on the
234 destination node, named ``prepare-node-join``, to configure the SSH
235 daemon. Paramiko is no longer necessary to configure nodes' SSH
236 daemons via ``gnt-node add``.
237 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
238 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
239 the flag is not reset after a master failover.
240 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
241 is recommended. 2.6 will be mandatory from the 2.8 series.
247 - New network management functionality to support automatic allocation
248 of IP addresses and managing of network parameters. See
249 :manpage:`gnt-network(8)` for more details.
250 - New external storage backend, to allow managing arbitrary storage
251 systems external to the cluster. See
252 :manpage:`ganeti-extstorage-interface(7)`.
253 - New ``exclusive-storage`` node parameter added, restricted to
254 nodegroup level. When it's set to true, physical disks are assigned in
255 an exclusive fashion to instances, as documented in :doc:`Partitioned
256 Ganeti <design-partitioned>`. Currently, only instances using the
257 ``plain`` disk template are supported.
258 - The KVM hypervisor has been updated with many new hypervisor
259 parameters, including a generic one for passing arbitrary command line
260 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
261 compatible up to qemu 1.4.
262 - A new tool, called ``mon-collector``, is the stand-alone executor of
263 the data collectors for a monitoring system. As of this version, it
264 just includes the DRBD data collector, that can be executed by calling
265 ``mon-collector`` using the ``drbd`` parameter. See
266 :manpage:`mon-collector(7)`.
267 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
268 for RAPI users. It allows granting permissions to query for
269 information to a specific user without giving
270 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
271 - A new tool named ``node-cleanup`` has been added. It cleans remains of
272 a cluster from a machine by stopping all daemons, removing
273 certificates and ssconf files. Unless the ``--no-backup`` option is
274 given, copies of the certificates are made.
275 - Instance creations now support the use of opportunistic locking,
276 potentially speeding up the (parallel) creation of multiple instances.
277 This feature is currently only available via the :doc:`RAPI
278 <rapi>` interface and when an instance allocator is used. If the
279 ``opportunistic_locking`` parameter is set the opcode will try to
280 acquire as many locks as possible, but will not wait for any locks
281 held by other opcodes. If not enough resources can be found to
282 allocate the instance, the temporary error code
283 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
284 retried thereafter, with or without opportunistic locking.
285 - New experimental linux-ha resource scripts.
286 - Restricted-commands support: ganeti can now be asked (via command line
287 or rapi) to perform commands on a node. These are passed via ganeti
288 RPC rather than ssh. This functionality is restricted to commands
289 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
290 reasons. The file is not copied automatically.
296 - Diskless instances are now externally mirrored (Issue 237). This for
297 now has only been tested in conjunction with explicit target nodes for
299 - Queries not needing locks or RPC access to the node can now be
300 performed by the confd daemon, making them independent from jobs, and
301 thus faster to execute. This is selectable at configure time.
302 - The functionality for allocating multiple instances at once has been
303 overhauled and is now also available through :doc:`RAPI <rapi>`.
305 There are no significant changes from version 2.7.0~rc3.
311 *(Released Tue, 25 Jun 2013)*
313 - Fix permissions on the confd query socket (Issue 477)
314 - Fix permissions on the job archive dir (Issue 498)
315 - Fix handling of an internal exception in replace-disks (Issue 472)
316 - Fix gnt-node info handling of shortened names (Issue 497)
317 - Fix gnt-instance grow-disk when wiping is enabled
318 - Documentation improvements, and support for newer pandoc
319 - Fix hspace honoring ipolicy for disks (Issue 484)
320 - Improve handling of the ``kvm_extra`` HV parameter
326 *(Released Fri, 24 May 2013)*
328 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
330 - Disks added through ``gnt-instance modify`` or created through
331 ``gnt-instance recreate-disks`` are wiped, if the
332 ``prealloc_wipe_disks`` flag is set.
333 - If wiping newly created disks fails, the disks are removed. Also,
334 partial failures in creating disks through ``gnt-instance modify``
335 triggers a cleanup of the partially-created disks.
336 - Removing the master IP address doesn't fail if the address has been
338 - Fix ownership of the OS log dir
339 - Workaround missing SO_PEERCRED constant (Issue 191)
345 *(Released Fri, 3 May 2013)*
347 This was the first release candidate of the 2.7 series. Since beta3:
349 - Fix kvm compatibility with qemu 1.4 (Issue 389)
350 - Documentation updates (admin guide, upgrade notes, install
351 instructions) (Issue 372)
352 - Fix gnt-group list nodes and instances count (Issue 436)
353 - Fix compilation without non-mandatory libraries (Issue 441)
354 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
355 - Make confd logging more verbose at INFO level (Issue 435)
356 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
357 - Fix failure path for instance storage type conversion (Issue 229)
358 - Update htools text backend documentation
359 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
360 - Disable inter-cluster instance move for file-based instances, because
361 it is dependant on instance export, which is not supported for
362 file-based instances. (Issue 414)
363 - Fix gnt-job crashes on non-ascii characters (Issue 427)
364 - Fix volume group checks on non-vm-capable nodes (Issue 432)
370 *(Released Mon, 22 Apr 2013)*
372 This was the third beta release of the 2.7 series. Since beta2:
374 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
375 - Fix data loss on wrong usage of ``gnt-instance move``
376 - Properly export errors in confd-based job queries
377 - Add ``users-setup`` tool
378 - Fix iallocator protocol to report 0 as a disk size for diskless
379 instances. This avoids hail breaking when a diskless instance is
381 - Fix job queue directory permission problem that made confd job queries
382 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
383 for access to archived jobs (Issue 406).
384 - Limit the sizes of networks supported by ``gnt-network`` to something
385 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
386 - Fix bugs in instance disk template conversion
387 - Fix GHC 7 compatibility
388 - Fix ``burnin`` install path (Issue 426).
389 - Allow very small disk grows (Issue 347).
390 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
391 that noded doesn't import masterd code (Issue 419).
392 - Make sure the default metavg at cluster init is the same as the vg, if
393 unspecified (Issue 358).
394 - Fix cleanup of partially created disks (part of Issue 416)
400 *(Released Tue, 2 Apr 2013)*
402 This was the second beta release of the 2.7 series. Since beta1:
404 - Networks no longer have a "type" slot, since this information was
405 unused in Ganeti: instead of it tags should be used.
406 - The rapi client now has a ``target_node`` option to MigrateInstance.
407 - Fix early exit return code for hbal (Issue 386).
408 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
409 - Fix ``rbd showmapped`` output parsing (Issue 312).
410 - Networks are now referenced indexed by UUID, rather than name. This
411 will require running cfgupgrade, from 2.7.0beta1, if networks are in
413 - The OS environment now includes network information.
414 - Deleting of a network is now disallowed if any instance nic is using
415 it, to prevent dangling references.
416 - External storage is now documented in man pages.
417 - The exclusive_storage flag can now only be set at nodegroup level.
418 - Hbal can now submit an explicit priority with its jobs.
419 - Many network related locking fixes.
420 - Bump up the required pylint version to 0.25.1.
421 - Fix the ``no_remember`` option in RAPI client.
422 - Many ipolicy related tests, qa, and fixes.
423 - Many documentation improvements and fixes.
424 - Fix building with ``--disable-file-storage``.
425 - Fix ``-q`` option in htools, which was broken if passed more than
427 - Some haskell/python interaction improvements and fixes.
428 - Fix iallocator in case of missing LVM storage.
429 - Fix confd config load in case of ``--no-lvm-storage``.
430 - The confd/query functionality is now mentioned in the security
437 *(Released Wed, 6 Feb 2013)*
439 This was the first beta release of the 2.7 series. All important changes
440 are listed in the latest 2.7 entry.
446 *(Released Fri, 21 Dec 2012)*
448 Important behaviour change: hbal won't rebalance anymore instances which
449 have the ``auto_balance`` attribute set to false. This was the intention
450 all along, but until now it only skipped those from the N+1 memory
451 reservation (DRBD-specific).
453 A significant number of bug fixes in this release:
455 - Fixed disk adoption interaction with ipolicy checks.
456 - Fixed networking issues when instances are started, stopped or
457 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
458 - Fixed the warning in cluster verify for shared storage instances not
460 - Fixed removal of storage directory on shared file storage (issue 262).
461 - Fixed validation of LVM volume group name in OpClusterSetParams
462 (``gnt-cluster modify``) (issue 285).
463 - Fixed runtime memory increases (``gnt-instance modify -m``).
464 - Fixed live migration under Xen's ``xl`` mode.
465 - Fixed ``gnt-instance console`` with ``xl``.
466 - Fixed building with newer Haskell compiler/libraries.
467 - Fixed PID file writing in Haskell daemons (confd); this prevents
468 restart issues if confd was launched manually (outside of
469 ``daemon-util``) while another copy of it was running
470 - Fixed a type error when doing live migrations with KVM (issue 297) and
471 the error messages for failing migrations have been improved.
472 - Fixed opcode validation for the out-of-band commands (``gnt-node
474 - Fixed a type error when unsetting OS hypervisor parameters (issue
475 311); now it's possible to unset all OS-specific hypervisor
477 - Fixed the ``dry-run`` mode for many operations: verification of
478 results was over-zealous but didn't take into account the ``dry-run``
479 operation, resulting in "wrong" failures.
480 - Fixed bash completion in ``gnt-job list`` when the job queue has
481 hundreds of entries; especially with older ``bash`` versions, this
482 results in significant CPU usage.
484 And lastly, a few other improvements have been made:
486 - Added option to force master-failover without voting (issue 282).
487 - Clarified error message on lock conflict (issue 287).
488 - Logging of newly submitted jobs has been improved (issue 290).
489 - Hostname checks have been made uniform between instance rename and
491 - The ``--submit`` option is now supported by ``gnt-debug delay``.
492 - Shutting down the master daemon by sending SIGTERM now stops it from
493 processing jobs waiting for locks; instead, those jobs will be started
494 once again after the master daemon is started the next time (issue
496 - Support for Xen's ``xl`` program has been improved (besides the fixes
498 - Reduced logging noise in the Haskell confd daemon (only show one log
499 entry for each config reload, instead of two).
500 - Several man page updates and typo fixes.
506 *(Released Fri, 12 Oct 2012)*
508 A small bugfix release. Among the bugs fixed:
510 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
511 made the command unusable.
512 - Commands that issue many jobs don't fail anymore just because some jobs
513 take so long that other jobs are archived.
514 - Failures during ``gnt-instance reinstall`` are reflected by the exit
516 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
518 - When ``always_failover`` is set, ``--allow-failover`` is not required
519 in migrate commands anymore.
520 - ``bash_completion`` works even if extglob is disabled.
521 - Fixed bug with locks that made failover for RDB-based instances fail.
522 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
523 a random node instead of one based on the allocator metric.
524 - Support for newer versions of pylint and pep8.
525 - Hail doesn't fail anymore when trying to add an instance of type
526 ``file``, ``sharedfile`` or ``rbd``.
527 - Added new Makefile target to rebuild the whole distribution, so that
528 all files are included.
534 *(Released Fri, 27 Jul 2012)*
537 .. attention:: The ``LUXI`` protocol has been made more consistent
538 regarding its handling of command arguments. This, however, leads to
539 incompatibility issues with previous versions. Please ensure that you
540 restart Ganeti daemons soon after the upgrade, otherwise most
541 ``LUXI`` calls (job submission, setting/resetting the drain flag,
542 pausing/resuming the watcher, cancelling and archiving jobs, querying
543 the cluster configuration) will fail.
552 The current ``admin_up`` field, which used to denote whether an instance
553 should be running or not, has been removed. Instead, ``admin_state`` is
554 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
556 The rational behind this is that an instance being “down” can have
559 - it could be down during a reboot
560 - it could be temporarily be down for a reinstall
561 - or it could be down because it is deprecated and kept just for its
564 The previous Boolean state was making it difficult to do capacity
565 calculations: should Ganeti reserve memory for a down instance? Now, the
566 tri-state field makes it clear:
568 - in ``up`` and ``down`` state, all resources are reserved for the
569 instance, and it can be at any time brought up if it is down
570 - in ``offline`` state, only disk space is reserved for it, but not
573 The field can have an extra use: since the transition between ``up`` and
574 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
575 transition between ``offline`` and ``down`` is done via ``gnt-instance
576 modify``, it is possible to given different rights to users. For
577 example, owners of an instance could be allowed to start/stop it, but
578 not transition it out of the offline state.
580 Instance policies and specs
581 +++++++++++++++++++++++++++
583 In previous Ganeti versions, an instance creation request was not
584 limited on the minimum size and on the maximum size just by the cluster
585 resources. As such, any policy could be implemented only in third-party
586 clients (RAPI clients, or shell wrappers over ``gnt-*``
587 tools). Furthermore, calculating cluster capacity via ``hspace`` again
588 required external input with regards to instance sizes.
590 In order to improve these workflows and to allow for example better
591 per-node group differentiation, we introduced instance specs, which
594 - minimum instance disk size, disk count, memory size, cpu count
595 - maximum values for the above metrics
596 - and “standard” values (used in ``hspace`` to calculate the standard
599 The minimum/maximum values can be also customised at node-group level,
600 for example allowing more powerful hardware to support bigger instance
603 Beside the instance specs, there are a few other settings belonging to
604 the instance policy framework. It is possible now to customise, per
605 cluster and node-group:
607 - the list of allowed disk templates
608 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
609 - the maximum ratio of instance to spindles (see below for more
610 information) for local storage
612 All these together should allow all tools that talk to Ganeti to know
613 what are the ranges of allowed values for instances and the
614 over-subscription that is allowed.
616 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
617 instance configuration, and the physical CPU configuration from the
618 node. For the spindle ratios however, we didn't track before these
619 values, so new parameters have been added:
621 - a new node parameter ``spindle_count``, defaults to 1, customisable at
622 node group or node level
623 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
625 Note that spindles in this context doesn't need to mean actual
626 mechanical hard-drives; it's just a relative number for both the node
627 I/O capacity and instance I/O consumption.
629 Instance migration behaviour
630 ++++++++++++++++++++++++++++
632 While live-migration is in general desirable over failover, it is
633 possible that for some workloads it is actually worse, due to the
634 variable time of the “suspend” phase during live migration.
636 To allow the tools to work consistently over such instances (without
637 having to hard-code instance names), a new backend parameter
638 ``always_failover`` has been added to control the migration/failover
639 behaviour. When set to True, all migration requests for an instance will
640 instead fall-back to failover.
642 Instance memory ballooning
643 ++++++++++++++++++++++++++
645 Initial support for memory ballooning has been added. The memory for an
646 instance is no longer fixed (backend parameter ``memory``), but instead
647 can vary between minimum and maximum values (backend parameters
648 ``minmem`` and ``maxmem``). Currently we only change an instance's
651 - live migrating or failing over and instance and the target node
652 doesn't have enough memory
653 - user requests changing the memory via ``gnt-instance modify
659 In order to control the use of specific CPUs by instance, support for
660 controlling CPU pinning has been added for the Xen, HVM and LXC
661 hypervisors. This is controlled by a new hypervisor parameter
662 ``cpu_mask``; details about possible values for this are in the
663 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
664 VCPU-to-CPU mapping) form will work well only when all nodes in your
665 cluster have the same amount of CPUs.
670 Another area in which Ganeti was not customisable were the parameters
671 used for storage configuration, e.g. how many stripes to use for LVM,
672 DRBD resync configuration, etc.
674 To improve this area, we've added disks parameters, which are
675 customisable at cluster and node group level, and which allow to
676 specify various parameters for disks (DRBD has the most parameters
677 currently), for example:
679 - DRBD resync algorithm and parameters (e.g. speed)
680 - the default VG for meta-data volumes for DRBD
681 - number of stripes for LVM (plain disk template)
684 These parameters can be modified via ``gnt-cluster modify -D …`` and
685 ``gnt-group modify -D …``, and are used at either instance creation (in
686 case of LVM stripes, for example) or at disk “activation” time
689 Rados block device support
690 ++++++++++++++++++++++++++
692 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
693 denoted by the ``rbd`` disk template type. This is considered
694 experimental, feedback is welcome. For details on configuring it, see
695 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
700 The existing master IP functionality works well only in simple setups (a
701 single network shared by all nodes); however, if nodes belong to
702 different networks, then the ``/32`` setup and lack of routing
703 information is not enough.
705 To allow the master IP to function well in more complex cases, the
706 system was reworked as follows:
708 - a master IP netmask setting has been added
709 - the master IP activation/turn-down code was moved from the node daemon
711 - whether to run the Ganeti-supplied master IP script or a user-supplied
712 on is a ``gnt-cluster init`` setting
714 Details about the location of the standard and custom setup scripts are
715 in the man page :manpage:`gnt-cluster(8)`; for information about the
716 setup script protocol, look at the Ganeti-supplied script.
721 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
724 It is now possible to use TLS-protected connections, and when renewing
725 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
726 it is now possible to specify spice or spice CA certificates. Also, it
727 is possible to configure a password for SPICE sessions via the
728 hypervisor parameter ``spice_password_file``.
730 There are also new parameters to control the compression and streaming
731 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
732 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
733 for the spice parameters.
735 Lastly, it is now possible to see the SPICE connection information via
736 ``gnt-instance console``.
741 A new tool (``tools/ovfconverter``) has been added that supports
742 conversion between Ganeti and the `Open Virtualization Format
743 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
746 This relies on the ``qemu-img`` tool to convert the disk formats, so the
747 actual compatibility with other virtualization solutions depends on it.
752 The configuration query daemon (``ganeti-confd``) is now optional, and
753 has been rewritten in Haskell; whether to use the daemon at all, use the
754 Python (default) or the Haskell version is selectable at configure time
755 via the ``--enable-confd`` parameter, which can take one of the
756 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
757 daemon will result in a smaller footprint; for larger systems, we
758 welcome feedback on the Haskell version which might become the default
761 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
762 daemon running. The Python version doesn't implement the new call.
765 User interface changes
766 ~~~~~~~~~~~~~~~~~~~~~~
768 We have replaced the ``--disks`` option of ``gnt-instance
769 replace-disks`` with a more flexible ``--disk`` option, which allows
770 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
771 disk size and mode can be changed upon recreation (via ``gnt-instance
772 recreate-disks``, which accepts the same ``--disk`` option).
774 As many people are used to a ``show`` command, we have added that as an
775 alias to ``info`` on all ``gnt-*`` commands.
777 The ``gnt-instance grow-disk`` command has a new mode in which it can
778 accept the target size of the disk, instead of the delta; this can be
779 more safe since two runs in absolute mode will be idempotent, and
780 sometimes it's also easier to specify the desired size directly.
782 Also the handling of instances with regard to offline secondaries has
783 been improved. Instance operations should not fail because one of it's
784 secondary nodes is offline, even though it's safe to proceed.
786 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
787 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
788 minors to instance name.
793 RAPI coverage has improved, with (for example) new resources for
794 recreate-disks, node power-cycle, etc.
799 There is partial support for ``xl`` in the Xen hypervisor; feedback is
802 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
803 whether to still support Python 2.4 or move to Python 2.6 as minimum
806 Support for Fedora has been slightly improved; the provided example
807 init.d script should work better on it and the INSTALL file should
808 document the needed dependencies.
813 The deprecated ``QueryLocks`` LUXI request has been removed. Use
814 ``Query(what=QR_LOCK, ...)`` instead.
816 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
817 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
818 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
819 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
820 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
822 RAPI client: ``CertificateError`` now derives from
823 ``GanetiApiError``. This should make it more easy to handle Ganeti
826 Deprecation warnings due to PyCrypto/paramiko import in
827 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
828 make sure to run an up-to-date paramiko version, if you use this tool.
830 The QA scripts now depend on Python 2.5 or above (the main code base
831 still works with Python 2.4).
833 The configuration file (``config.data``) is now written without
834 indentation for performance reasons; if you want to edit it, it can be
835 re-formatted via ``tools/fmtjson``.
837 A number of bugs has been fixed in the cluster merge tool.
839 ``x509`` certification verification (used in import-export) has been
840 changed to allow the same clock skew as permitted by the cluster
841 verification. This will remove some rare but hard to diagnose errors in
848 *(Released Thu, 19 Jul 2012)*
850 Very few changes from rc4 to the final release, only bugfixes:
852 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
853 instance, fix CDROM booting for KVM instances)
854 - fixed node group modification of node parameters
855 - fixed issue in LUClusterVerifyGroup with multi-group clusters
856 - fixed generation of bash completion to ensure a stable ordering
863 *(Released Fri, 13 Jul 2012)*
865 Third release candidate for 2.6. The following changes were done from
868 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
869 - Fixed an inconsistency in the LUXI protocol with the provided
870 arguments (NOT backwards compatible)
871 - Fixed a bug with node groups ipolicy where ``min`` was greater than
872 the cluster ``std`` value
873 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
874 easier instance debugging on nodes (requires ``hconfd`` to work)
880 *(Released Tue, 03 Jul 2012)*
882 Second release candidate for 2.6. The following changes were done from
885 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
887 - Fixed a RAPI regression on missing beparams/memory
888 - Fixed redistribution of files on offline nodes
889 - Added possibility to run activate-disks even though secondaries are
890 offline. With this change it relaxes also the strictness on some other
891 commands which use activate disks internally:
892 * ``gnt-instance start|reboot|rename|backup|export``
893 - Made it possible to remove safely an instance if its secondaries are
895 - Made it possible to reinstall even though secondaries are offline
901 *(Released Mon, 25 Jun 2012)*
903 First release candidate for 2.6. The following changes were done from
906 - Fixed bugs with disk parameters and ``rbd`` templates as well as
908 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
909 behaviour. It supports now the modify operation
910 - ``hcheck`` implemented to analyze cluster health and possibility of
911 improving health by rebalance
912 - ``hbal`` has been improved in dealing with split instances
918 *(Released Mon, 11 Jun 2012)*
920 Second beta release of 2.6. The following changes were done from beta2
923 - Fixed ``daemon-util`` with non-root user models
924 - Fixed creation of plain instances with ``--no-wait-for-sync``
925 - Fix wrong iv_names when running ``cfgupgrade``
926 - Export more information in RAPI group queries
927 - Fixed bug when changing instance network interfaces
928 - Extended burnin to do NIC changes
929 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
930 - Changed default for DRBD barriers
931 - Fixed DRBD error reporting for syncer rate
932 - Verify the options on disk parameters
934 And of course various fixes to documentation and improved unittests and
941 *(Released Wed, 23 May 2012)*
943 First beta release of 2.6. The following changes were done from beta1 to
946 - integrated patch for distributions without ``start-stop-daemon``
947 - adapted example init.d script to work on Fedora
948 - fixed log handling in Haskell daemons
949 - adapted checks in the watcher for pycurl linked against libnss
950 - add partial support for ``xl`` instead of ``xm`` for Xen
951 - fixed a type issue in cluster verification
952 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
955 Plus integrated fixes from the 2.5 branch:
957 - fixed ``kvm-ifup`` to use ``/bin/bash``
958 - fixed parallel build failures
959 - KVM live migration when using a custom keymap
965 *(Released Tue, 24 Jul 2012)*
967 A small bugfix release, with no new features:
969 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
970 different default shell (e.g. Debian, Ubuntu)
971 - fixed KVM startup and live migration with a custom keymap (fixes Issue
972 243 and Debian bug #650664)
973 - fixed compatibility with KVM versions that don't support multiple boot
974 devices (fixes Issue 230 and Debian bug #624256)
976 Additionally, a few fixes were done to the build system (fixed parallel
977 build failures) and to the unittests (fixed race condition in test for
978 FileID functions, and the default enable/disable mode for QA test is now
985 *(Released Fri, 11 May 2012)*
987 A small bugfix release.
989 The main issues solved are on the topic of compatibility with newer LVM
992 - fixed parsing of ``lv_attr`` field
993 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
994 the ``--force`` flag is needed
996 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
997 to accept kernel 3.x too (was hardcoded to 2.6.*).
999 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1000 of hook results and that also made display of other errors suboptimal
1001 was fixed; the code behaves now like 2.4 and earlier.
1003 Another change in 2.5, the cleanup of the OS scripts environment, is too
1004 aggressive: it removed even the ``PATH`` variable, which requires the OS
1005 scripts to *always* need to export it. Since this is a bit too strict,
1006 we now export a minimal PATH, the same that we export for hooks.
1008 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1009 integrated into this release.
1011 Finally, a few other miscellaneous changes were done (no new features,
1012 just small improvements):
1014 - Fix ``gnt-group --help`` display
1015 - Fix hardcoded Xen kernel path
1016 - Fix grow-disk handling of invalid units
1017 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1018 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1024 *(Released Thu, 12 Apr 2012)*
1026 Incompatible/important changes and bugfixes
1027 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1029 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1030 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1031 to match the underlying LUXI interface.
1032 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1033 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1034 not maintain backwards-compatibility as the underlying operation
1035 changed in an incompatible way. The RAPI client can talk to old
1036 servers, but it needs to be told so as the return value changed.
1037 - When creating file-based instances via RAPI, the ``file_driver``
1038 parameter no longer defaults to ``loop`` and must be specified.
1039 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1041 - Support for the undocumented and deprecated RAPI instance creation
1042 request format version 0 has been dropped. Use version 1, supported
1043 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1044 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1045 documentation <install>`.
1046 - The "cluster-verify" hooks are now executed per group by the
1047 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1048 if you just run ``gnt-cluster verify``, which generates one opcode per
1050 - The environment as passed to the OS scripts is cleared, and thus no
1051 environment variables defined in the node daemon's environment will be
1052 inherited by the scripts.
1053 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1055 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1056 support operations involving multiple node groups.
1057 - Offline nodes are ignored when failing over an instance.
1058 - Support for KVM version 1.0, which changed the version reporting format
1060 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1062 - ``Makefile`` is now compatible with Automake 1.11.2
1063 - Includes all bugfixes made in the 2.4 series
1068 - The ganeti-htools project has been merged into the ganeti-core source
1069 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1070 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1071 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1072 GPT-style partition tables (requires `parted
1073 <http://www.gnu.org/s/parted/>`_).
1074 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1075 - Allowed adding tags on instance creation.
1076 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1077 - Allow instances to be started in a paused state, enabling the user to
1078 see the complete console output on boot using the console.
1079 - Added new hypervisor flag to control default reboot behaviour
1080 (``reboot_behavior``).
1081 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1082 - Improved out-of-band management support:
1084 - Added ``gnt-node health`` command reporting the health status of
1086 - Added ``gnt-node power`` command to manage power status of nodes.
1087 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1089 - Instance migration can fall back to failover if instance is not
1091 - Filters can be used when listing nodes, instances, groups and locks;
1092 see :manpage:`ganeti(7)` manpage.
1093 - Added post-execution status as variables to :doc:`hooks <hooks>`
1095 - Instance tags are exported/imported together with the instance.
1096 - When given an explicit job ID, ``gnt-job info`` will work for archived
1098 - Jobs can define dependencies on other jobs (not yet supported via
1099 RAPI or command line, but used by internal commands and usable via
1102 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1105 - Instance failover is now available as a RAPI resource
1106 (``/2/instances/[instance_name]/failover``).
1107 - ``gnt-instance info`` defaults to static information if primary node
1109 - Opcodes have a new ``comment`` attribute.
1110 - Added basic SPICE support to KVM hypervisor.
1111 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1113 Node group improvements
1114 ~~~~~~~~~~~~~~~~~~~~~~~
1116 - ``gnt-cluster verify`` has been modified to check groups separately,
1117 thereby improving performance.
1118 - Node group support has been added to ``gnt-cluster verify-disks``,
1119 which now operates per node group.
1120 - Watcher has been changed to work better with node groups.
1122 - One process and state file per node group.
1123 - Slow watcher in one group doesn't block other group's watcher.
1125 - Added new command, ``gnt-group evacuate``, to move all instances in a
1126 node group to other groups.
1127 - Added ``gnt-instance change-group`` to move an instance to another
1129 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1130 per-group operations.
1131 - Node groups can be tagged.
1132 - Some operations switch from an exclusive to a shared lock as soon as
1134 - Instance's primary and secondary nodes' groups are now available as
1135 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1136 and ``snodes.group.uuid``).
1141 - Numerous updates to documentation and manpages.
1143 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1145 - Some opcode/job results are now also documented, see :doc:`RAPI
1148 - A lockset's internal lock is now also visible in lock monitor.
1149 - Log messages from job queue workers now contain information about the
1150 opcode they're processing.
1151 - ``gnt-instance console`` no longer requires the instance lock.
1152 - A short delay when waiting for job changes reduces the number of LUXI
1153 requests significantly.
1154 - DRBD metadata volumes are overwritten with zeros during disk creation.
1155 - Out-of-band commands no longer acquire the cluster lock in exclusive
1157 - ``devel/upload`` now uses correct permissions for directories.
1163 *(Released Fri, 23 Mar 2012)*
1165 This was the sixth release candidate of the 2.5 series.
1171 *(Released Mon, 9 Jan 2012)*
1173 This was the fifth release candidate of the 2.5 series.
1179 *(Released Thu, 27 Oct 2011)*
1181 This was the fourth release candidate of the 2.5 series.
1187 *(Released Wed, 26 Oct 2011)*
1189 This was the third release candidate of the 2.5 series.
1195 *(Released Tue, 18 Oct 2011)*
1197 This was the second release candidate of the 2.5 series.
1203 *(Released Tue, 4 Oct 2011)*
1205 This was the first release candidate of the 2.5 series.
1211 *(Released Wed, 31 Aug 2011)*
1213 This was the third beta release of the 2.5 series.
1219 *(Released Mon, 22 Aug 2011)*
1221 This was the second beta release of the 2.5 series.
1227 *(Released Fri, 12 Aug 2011)*
1229 This was the first beta release of the 2.5 series.
1235 *(Released Thu, 27 Oct 2011)*
1237 - Fixed bug when parsing command line parameter values ending in
1239 - Fixed assertion error after unclean master shutdown
1240 - Disable HTTP client pool for RPC, significantly reducing memory usage
1242 - Fixed queue archive creation with wrong permissions
1248 *(Released Tue, 23 Aug 2011)*
1252 - Fixed documentation for importing with ``--src-dir`` option
1253 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1254 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1260 *(Released Fri, 5 Aug 2011)*
1262 Many bug-fixes and a few small features:
1264 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1265 issues when you tried to add an instance with two MAC addresses in one
1267 - KVM: fixed per-instance stored UID value
1268 - KVM: configure bridged NICs at migration start
1269 - KVM: Fix a bug where instance will not start with never KVM versions
1271 - Added OS search path to ``gnt-cluster info``
1272 - Fixed an issue with ``file_storage_dir`` where you were forced to
1273 provide an absolute path, but the documentation states it is a
1274 relative path, the documentation was right
1275 - Added a new parameter to instance stop/start called ``--no-remember``
1276 that will make the state change to not be remembered
1277 - Implemented ``no_remember`` at RAPI level
1278 - Improved the documentation
1279 - Node evacuation: don't call IAllocator if node is already empty
1280 - Fixed bug in DRBD8 replace disks on current nodes
1281 - Fixed bug in recreate-disks for DRBD instances
1282 - Moved assertion checking locks in ``gnt-instance replace-disks``
1283 causing it to abort with not owning the right locks for some situation
1284 - Job queue: Fixed potential race condition when cancelling queued jobs
1285 - Fixed off-by-one bug in job serial generation
1286 - ``gnt-node volumes``: Fix instance names
1287 - Fixed aliases in bash completion
1288 - Fixed a bug in reopening log files after being sent a SIGHUP
1289 - Added a flag to burnin to allow specifying VCPU count
1290 - Bugfixes to non-root Ganeti configuration
1296 *(Released Thu, 12 May 2011)*
1298 Many bug-fixes and a few new small features:
1300 - Fixed a bug related to log opening failures
1301 - Fixed a bug in instance listing with orphan instances
1302 - Fixed a bug which prevented resetting the cluster-level node parameter
1303 ``oob_program`` to the default
1304 - Many fixes related to the ``cluster-merge`` tool
1305 - Fixed a race condition in the lock monitor, which caused failures
1306 during (at least) creation of many instances in parallel
1307 - Improved output for gnt-job info
1308 - Removed the quiet flag on some ssh calls which prevented debugging
1310 - Improved the N+1 failure messages in cluster verify by actually
1311 showing the memory values (needed and available)
1312 - Increased lock attempt timeouts so that when executing long operations
1313 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1314 too early and thus prevent the use of the 'fair' mechanism
1315 - Changed instance query data (``gnt-instance info``) to not acquire
1316 locks unless needed, thus allowing its use on locked instance if only
1317 static information is asked for
1318 - Improved behaviour with filesystems that do not support rename on an
1320 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1321 kept locks on all nodes during the wipe, which is unneeded
1322 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1323 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1324 10GiB) which caused the wipe to fail right at the end in some cases
1325 - Fixed master IP activation when doing master failover with no-voting
1326 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1327 the master node itself
1328 - Fixed potential data-loss in under disk full conditions, where Ganeti
1329 wouldn't check correctly the return code and would consider
1330 partially-written files 'correct'
1331 - Fixed bug related to multiple VGs and DRBD disk replacing
1332 - Added new disk parameter ``metavg`` that allows placement of the meta
1333 device for DRBD in a different volume group
1334 - Fixed error handling in the node daemon when the system libc doesn't
1335 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1336 - Fixed lock release during replace-disks, which kept cluster-wide locks
1337 when doing disk replaces with an iallocator script
1338 - Added check for missing bridges in cluster verify
1339 - Handle EPIPE errors while writing to the terminal better, so that
1340 piping the output to e.g. ``less`` doesn't cause a backtrace
1341 - Fixed rare case where a ^C during Luxi calls could have been
1342 interpreted as server errors, instead of simply terminating
1343 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1345 - Added a few more parameters to the KVM hypervisor, allowing a second
1346 CDROM, custom disk type for CDROMs and a floppy image
1347 - Removed redundant message in instance rename when the name is given
1349 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1350 disks on new nodes, allowing recreation when the original instance
1351 nodes are completely gone
1352 - Added option when converting disk templates to DRBD to skip waiting
1353 for the resync, in order to make the instance available sooner
1354 - Added two new variables to the OS scripts environment (containing the
1356 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1357 to allow use of ``pvgrub`` as bootloader
1358 - Changed the instance memory modifications to only check out-of-memory
1359 conditions on memory increases, and turned the secondary node warnings
1360 into errors (they can still be overridden via ``--force``)
1361 - Fixed the handling of a corner case when the Python installation gets
1362 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1363 to execute a command that doesn't exist
1364 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1365 node of the instance returned failures during instance shutdown; this
1366 adds the option ``--ignore-consistency`` to gnt-instance move
1368 And as usual, various improvements to the error messages, documentation
1375 *(Released Wed, 09 Mar 2011)*
1377 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1378 the RAPI users file if run twice (even with ``--dry-run``).
1380 The release fixes that bug (nothing else changed).
1386 *(Released Mon, 07 Mar 2011)*
1388 Final 2.4.0 release. Just a few small fixes:
1390 - Fixed RAPI node evacuate
1391 - Fixed the kvm-ifup script
1392 - Fixed internal error handling for special job cases
1393 - Updated man page to specify the escaping feature for options
1399 *(Released Mon, 28 Feb 2011)*
1401 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1402 that this feature wiped the disks of the wrong instance, leading to loss
1407 - Fixed title of query field containing instance name
1408 - Expanded the glossary in the documentation
1409 - Fixed one unittest (internal issue)
1415 *(Released Mon, 21 Feb 2011)*
1417 A number of bug fixes plus just a couple functionality changes.
1419 On the user-visible side, the ``gnt-* list`` command output has changed
1420 with respect to "special" field states. The current rc1 style of display
1421 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1422 the default output mode special fields states are displayed as follows:
1424 - Offline resource: ``*``
1425 - Unavailable/not applicable: ``-``
1426 - Data missing (RPC failure): ``?``
1427 - Unknown field: ``??``
1429 Another user-visible change is the addition of ``--force-join`` to
1434 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1435 - Fixed regression in RAPI/instance reinstall where all parameters were
1436 required (instead of optional)
1437 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1438 - Fixed iallocator usage (offline nodes were not considered offline)
1439 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1440 - Fixed hypervisor and OS parameter validation with respect to
1441 non-vm_capable nodes
1442 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1444 - Fixed ``tools/listrunner`` with respect to agent-based usage
1450 *(Released Fri, 4 Feb 2011)*
1452 Many changes and fixes since the beta1 release. While there were some
1453 internal changes, the code has been mostly stabilised for the RC
1456 Note: the dumb allocator was removed in this release, as it was not kept
1457 up-to-date with the IAllocator protocol changes. It is recommended to
1458 use the ``hail`` command from the ganeti-htools package.
1460 Note: the 2.4 and up versions of Ganeti are not compatible with the
1461 0.2.x branch of ganeti-htools. You need to upgrade to
1462 ganeti-htools-0.3.0 (or later).
1464 Regressions fixed from 2.3
1465 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1467 - Fixed the ``gnt-cluster verify-disks`` command
1468 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1470 - Fixed disk adoption breakage
1471 - Fixed wrong headers in instance listing for field aliases
1476 - Fixed corner case in KVM handling of NICs
1477 - Fixed many cases of wrong handling of non-vm_capable nodes
1478 - Fixed a bug where a missing instance symlink was not possible to
1479 recreate with any ``gnt-*`` command (now ``gnt-instance
1480 activate-disks`` does it)
1481 - Fixed the volume group name as reported by ``gnt-cluster
1483 - Increased timeouts for the import-export code, hopefully leading to
1484 fewer aborts due network or instance timeouts
1485 - Fixed bug in ``gnt-node list-storage``
1486 - Fixed bug where not all daemons were started on cluster
1487 initialisation, but only at the first watcher run
1488 - Fixed many bugs in the OOB implementation
1489 - Fixed watcher behaviour in presence of instances with offline
1491 - Fixed instance list output for instances running on the wrong node
1492 - a few fixes to the cluster-merge tool, but it still cannot merge
1493 multi-node groups (currently it is not recommended to use this tool)
1499 - Improved network configuration for the KVM hypervisor
1500 - Added e1000 as a supported NIC for Xen-HVM
1501 - Improved the lvmstrap tool to also be able to use partitions, as
1502 opposed to full disks
1503 - Improved speed of disk wiping (the cluster parameter
1504 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1505 of instance creations
1506 - Added documentation for the OS parameters
1507 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1508 hypervisor is not responding
1509 - Added display of blacklisted and hidden OS information in
1510 ``gnt-cluster info``
1511 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1512 NIC and node parameters, which might create problems with currently
1513 invalid (but undetected) configuration files, but prevents validation
1514 failures when unrelated parameters are modified
1515 - Changed cluster initialisation to wait for the master daemon to become
1517 - Expanded the RAPI interface:
1519 - Added config redistribution resource
1520 - Added activation/deactivation of instance disks
1521 - Added export of console information
1523 - Implemented log file reopening on SIGHUP, which allows using
1524 logrotate(8) for the Ganeti log files
1525 - Added a basic OOB helper script as an example
1531 *(Released Fri, 14 Jan 2011)*
1536 - Fixed timezone issues when formatting timestamps
1537 - Added support for node groups, available via ``gnt-group`` and other
1539 - Added out-of-band framework and management, see :doc:`design
1540 document <design-oob>`
1541 - Removed support for roman numbers from ``gnt-node list`` and
1542 ``gnt-instance list``.
1543 - Allowed modification of master network interface via ``gnt-cluster
1544 modify --master-netdev``
1545 - Accept offline secondaries while shutting down instance disks
1546 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1547 - Added support for multiple LVM volume groups
1548 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1550 - Added commands to list available fields:
1552 - ``gnt-node list-fields``
1553 - ``gnt-group list-fields``
1554 - ``gnt-instance list-fields``
1556 - Updated documentation and man pages
1561 - Moved ``rapi_users`` file into separate directory, now named
1562 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1564 - Added new tool for running commands on many machines,
1565 ``tools/ganeti-listrunner``
1566 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1567 improving the ``gnt-instance console`` output
1568 - Allowed customisation of disk index separator at ``configure`` time
1569 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1570 - Added support for non-partitioned md disks in ``lvmstrap``
1571 - Added script to gracefully power off KVM instances
1572 - Split ``utils`` module into smaller parts
1573 - Changed query operations to return more detailed information, e.g.
1574 whether an information is unavailable due to an offline node. To use
1575 this new functionality, the LUXI call ``Query`` must be used. Field
1576 information is now stored by the master daemon and can be retrieved
1577 using ``QueryFields``. Instances, nodes and groups can also be queried
1578 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1579 exposed via RAPI). The following commands make use of this
1580 infrastructure change:
1582 - ``gnt-group list``
1583 - ``gnt-group list-fields``
1585 - ``gnt-node list-fields``
1586 - ``gnt-instance list``
1587 - ``gnt-instance list-fields``
1588 - ``gnt-debug locks``
1593 - New RAPI resources (see :doc:`rapi`):
1597 - ``/2/groups/[group_name]``
1598 - ``/2/groups/[group_name]/assign-nodes``
1599 - ``/2/groups/[group_name]/modify``
1600 - ``/2/groups/[group_name]/rename``
1601 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1605 - Implemented ``no_install`` for instance creation
1606 - Implemented OS parameters for instance reinstallation, allowing
1607 use of special settings on reinstallation (e.g. for preserving data)
1612 - Added IPv6 support in import/export
1613 - Pause DRBD synchronization while wiping disks on instance creation
1614 - Updated unittests and QA scripts
1615 - Improved network parameters passed to KVM
1616 - Converted man pages from docbook to reStructuredText
1622 *(Released Mon, 20 Dec 2010)*
1624 Released version 2.3.1~rc1 without any changes.
1630 *(Released Wed, 1 Dec 2010)*
1632 - impexpd: Disable OpenSSL compression in socat if possible (backport
1633 from master, commit e90739d625b, see :doc:`installation guide
1634 <install-quick>` for details)
1635 - Changed unittest coverage report to exclude test scripts
1636 - Added script to check version format
1642 *(Released Wed, 1 Dec 2010)*
1644 Released version 2.3.0~rc1 without any changes.
1650 *(Released Fri, 19 Nov 2010)*
1652 A number of bugfixes and documentation updates:
1654 - Update ganeti-os-interface documentation
1655 - Fixed a bug related to duplicate MACs or similar items which should be
1657 - Fix breakage in OS state modify
1658 - Reinstall instance: disallow offline secondaries (fixes bug related to
1659 OS changing but reinstall failing)
1660 - plus all the other fixes between 2.2.1 and 2.2.2
1666 *(Released Tue, 2 Nov 2010)*
1668 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1669 - Fixed master failover race with watcher
1670 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1672 - Accept previously stopped instance for export with instance removal
1673 - Simplify and extend the environment variables for instance OS scripts
1674 - Added new node flags, ``master_capable`` and ``vm_capable``
1675 - Added optional instance disk wiping prior during allocation. This is a
1676 cluster-wide option and can be set/modified using
1677 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1678 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1679 :doc:`install-quick`
1680 - Added a new watcher option (``--ignore-pause``)
1681 - Added option to ignore offline node on instance start/stop
1682 (``--ignore-offline``)
1683 - Allow overriding OS parameters with ``gnt-instance reinstall``
1684 - Added ability to change node's secondary IP address using ``gnt-node
1686 - Implemented privilege separation for all daemons except
1687 ``ganeti-noded``, see ``configure`` options
1688 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1690 - Implemented job priorities (see ``ganeti(7)`` manpage)
1691 - Ignore failures while shutting down instances during failover from
1693 - Exit daemon's bootstrap process only once daemon is ready
1694 - Export more information via ``LUInstanceQuery``/remote API
1695 - Improved documentation, QA and unittests
1696 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1697 restart if the file was created or changed
1698 - Added LUXI protocol version sent with each request and response,
1699 allowing detection of server/client mismatches
1700 - Moved the Python scripts among gnt-* and ganeti-* into modules
1701 - Moved all code related to setting up SSH to an external script,
1703 - Infrastructure changes for node group support in future versions
1709 *(Released Fri, 19 Nov 2010)*
1711 A few small bugs fixed, and some improvements to the build system:
1713 - Fix documentation regarding conversion to drbd
1714 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1716 - Fix error handling in node modify with multiple changes
1717 - Allow remote imports without checked names
1723 *(Released Tue, 19 Oct 2010)*
1725 - Disable SSL session ID cache in RPC client
1731 *(Released Thu, 14 Oct 2010)*
1733 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1734 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1736 - Fix problems with interaction between Curl and Python's HTTP server,
1737 resulting in increased speed in many RPC calls
1738 - Improve our release script to prevent breakage with older aclocal and
1745 *(Released Thu, 7 Oct 2010)*
1747 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1748 value retrieved from master
1749 - Added support for blacklisted or hidden OS definitions
1750 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1751 - Added support for -mem-path in KVM hypervisor abstraction layer
1752 - Allow overriding instance parameters in tool for inter-cluster
1753 instance moves (``tools/move-instance``)
1754 - Improved opcode summaries (e.g. in ``gnt-job list``)
1755 - Improve consistency of OS listing by sorting it
1756 - Documentation updates
1762 *(Released Fri, 8 Oct 2010)*
1764 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1770 *(Released Mon, 4 Oct 2010)*
1772 - Fixed regression in ``gnt-instance rename``
1778 *(Released Wed, 22 Sep 2010)*
1780 - Fixed OS_VARIANT variable for OS scripts
1781 - Fixed cluster tag operations via RAPI
1782 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1783 - Disabled RAPI CA checks in watcher
1789 *(Released Mon, 23 Aug 2010)*
1791 - Support DRBD versions of the format "a.b.c.d"
1793 - Re-introduce support for usage from multiple threads in RAPI client
1794 - Instance renames and modify via RAPI
1795 - Work around race condition between processing and archival in job
1797 - Mark opcodes following failed one as failed, too
1798 - Job field ``lock_status`` was removed due to difficulties making it
1799 work with the changed job queue in Ganeti 2.2; a better way to monitor
1800 locks is expected for a later 2.2.x release
1801 - Fixed dry-run behaviour with many commands
1802 - Support ``ssh-agent`` again when adding nodes
1803 - Many additional bugfixes
1809 *(Released Fri, 30 Jul 2010)*
1811 Important change: the internal RPC mechanism between Ganeti nodes has
1812 changed from using a home-grown http library (based on the Python base
1813 libraries) to use the PycURL library. This requires that PycURL is
1814 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1815 against GnuTLS by default. cURL's support for GnuTLS had known issues
1816 before cURL 7.21.0 and we recommend using the latest cURL release or
1817 linking against OpenSSL. Most other distributions already link PycURL
1818 and cURL against OpenSSL. The command::
1820 python -c 'import pycurl; print pycurl.version'
1822 can be used to determine the libraries PycURL and cURL are linked
1825 Other significant changes:
1827 - Rewrote much of the internals of the job queue, in order to achieve
1828 better parallelism; this decouples job query operations from the job
1829 processing, and it should allow much nicer behaviour of the master
1830 daemon under load, and it also has uncovered some long-standing bugs
1831 related to the job serialisation (now fixed)
1832 - Added a default iallocator setting to the cluster parameters,
1833 eliminating the need to always pass nodes or an iallocator for
1834 operations that require selection of new node(s)
1835 - Added experimental support for the LXC virtualization method
1836 - Added support for OS parameters, which allows the installation of
1837 instances to pass parameter to OS scripts in order to customise the
1839 - Added a hypervisor parameter controlling the migration type (live or
1840 non-live), since hypervisors have various levels of reliability; this
1841 has renamed the 'live' parameter to 'mode'
1842 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1843 logical volumes, meaning that cluster verify will ignore them and not
1844 flag their presence as errors
1845 - The watcher will now reset the error count for failed instances after
1846 8 hours, thus allowing self-healing if the problem that caused the
1847 instances to be down/fail to start has cleared in the meantime
1848 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1849 check for, and warn, if the drbd module parameter ``usermode_helper``
1850 is not consistent with the cluster-wide setting; this is needed to
1851 make diagnose easier of failed drbd creations
1852 - Started adding base IPv6 support, but this is not yet
1853 enabled/available for use
1854 - Rename operations (cluster, instance) will now return the new name,
1855 which is especially useful if a short name was passed in
1856 - Added support for instance migration in RAPI
1857 - Added a tool to pre-configure nodes for the SSH setup, before joining
1858 them to the cluster; this will allow in the future a simplified model
1859 for node joining (but not yet fully enabled in 2.2); this needs the
1860 paramiko python library
1861 - Fixed handling of name-resolving errors
1862 - Fixed consistency of job results on the error path
1863 - Fixed master-failover race condition when executed multiple times in
1865 - Fixed many bugs related to the job queue (mostly introduced during the
1866 2.2 development cycle, so not all are impacting 2.1)
1867 - Fixed instance migration with missing disk symlinks
1868 - Fixed handling of unknown jobs in ``gnt-job archive``
1869 - And many other small fixes/improvements
1873 - Enhanced both the unittest and the QA coverage
1874 - Switched the opcode validation to a generic model, and extended the
1875 validation to all opcode parameters
1876 - Changed more parts of the code that write shell scripts to use the
1878 - Switched the master daemon to use the asyncore library for the Luxi
1885 *(Released Thu, 17 Jun 2010)*
1887 - Added tool (``move-instance``) and infrastructure to move instances
1888 between separate clusters (see :doc:`separate documentation
1889 <move-instance>` and :doc:`design document <design-2.2>`)
1890 - Added per-request RPC timeout
1891 - RAPI now requires a Content-Type header for requests with a body (e.g.
1892 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1893 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1894 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1896 - Implemented initial support for running Ganeti daemons as separate
1897 users, see configure-time flags ``--with-user-prefix`` and
1898 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1900 - Instances can be removed after export (``gnt-backup export
1901 --remove-instance``)
1902 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1903 key (instead of 1024 bit)
1904 - Added new cluster configuration file for cluster domain secret
1905 - Import/export now use SSL instead of SSH
1906 - Added support for showing estimated time when exporting an instance,
1907 see the ``ganeti-os-interface(7)`` manpage and look for
1914 *(Released Tue, 16 Nov 2010)*
1916 Some more bugfixes. Unless critical bugs occur, this will be the last
1919 - Fix case of MAC special-values
1920 - Fix mac checker regex
1921 - backend: Fix typo causing "out of range" error
1922 - Add missing --units in gnt-instance list man page
1928 *(Released Tue, 24 Aug 2010)*
1931 - Don't ignore secondary node silently on non-mirrored disk templates
1933 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1934 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1935 - Properly document the usb_mouse parameter
1936 - Fix path in ganeti-rapi(8) (issue 116)
1937 - Adjust error message when the ganeti user's .ssh directory is
1939 - Add same-node-check when changing the disk template to drbd
1945 *(Released Fri, 16 Jul 2010)*
1948 - Add an option to only select some reboot types during qa/burnin.
1949 (on some hypervisors consequent reboots are not supported)
1950 - Fix infrequent race condition in master failover. Sometimes the old
1951 master ip address would be still detected as up for a short time
1952 after it was removed, causing failover to fail.
1953 - Decrease mlockall warnings when the ctypes module is missing. On
1954 Python 2.4 we support running even if no ctypes module is installed,
1955 but we were too verbose about this issue.
1956 - Fix building on old distributions, on which man doesn't have a
1958 - Fix RAPI not to ignore the MAC address on instance creation
1959 - Implement the old instance creation format in the RAPI client.
1965 *(Released Thu, 01 Jul 2010)*
1967 A small bugfix release:
1968 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1969 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1970 a non-existing option
1971 - Fix instance create: the --force-variant option was ignored
1972 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1973 - Fix modify node storage with non-FQDN arguments
1974 - Fix RAPI client to authenticate under Python 2.6 when used
1975 for more than 5 requests needing authentication
1976 - Fix gnt-instance modify -t (storage) giving a wrong error message
1977 when converting a non-shutdown drbd instance to plain
1983 *(Released Fri, 18 Jun 2010)*
1985 A small bugfix release:
1987 - Fix live migration of KVM instances started with older Ganeti
1988 versions which had fewer hypervisor parameters
1989 - Fix gnt-instance grow-disk on down instances
1990 - Fix an error-reporting bug during instance migration
1991 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1992 silently ignoring broken ones
1993 - Fix an RPC error reporting bug affecting, for example, RAPI client
1995 - Fix bug triggered by different API version os-es on different nodes
1996 - Fix a bug in instance startup with custom hvparams: OS level
1997 parameters would fail to be applied.
1998 - Fix the RAPI client under Python 2.6 (but more work is needed to
1999 make it work completely well with OpenSSL)
2000 - Fix handling of errors when resolving names from DNS
2006 *(Released Thu, 3 Jun 2010)*
2008 A medium sized development cycle. Some new features, and some
2009 fixes/small improvements/cleanups.
2011 Significant features
2012 ~~~~~~~~~~~~~~~~~~~~
2014 The node deamon now tries to mlock itself into memory, unless the
2015 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2016 its logs, and falls back to console logging. This allows emergency
2017 features such as ``gnt-node powercycle`` to work even in the event of a
2018 broken node disk (tested offlining the disk hosting the node's
2019 filesystem and dropping its memory caches; don't try this at home)
2021 KVM: add vhost-net acceleration support. It can be tested with a new
2022 enough version of the kernel and of qemu-kvm.
2024 KVM: Add instance chrooting feature. If you use privilege dropping for
2025 your VMs you can also now force them to chroot to an empty directory,
2026 before starting the emulated guest.
2028 KVM: Add maximum migration bandwith and maximum downtime tweaking
2029 support (requires a new-enough version of qemu-kvm).
2031 Cluster verify will now warn if the master node doesn't have the master
2032 ip configured on it.
2034 Add a new (incompatible) instance creation request format to RAPI which
2035 supports all parameters (previously only a subset was supported, and it
2036 wasn't possible to extend the old format to accomodate all the new
2037 features. The old format is still supported, and a client can check for
2038 this feature, before using it, by checking for its presence in the
2039 ``features`` RAPI resource.
2041 Now with ancient latin support. Try it passing the ``--roman`` option to
2042 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2043 (requires the python-roman module to be installed, in order to work).
2048 As usual many internal code refactorings, documentation updates, and
2051 - Lots of improvements and cleanups to the experimental Remote API
2052 (RAPI) client library.
2053 - A new unit test suite for the core daemon libraries.
2054 - A fix to creating missing directories makes sure the umask is not
2055 applied anymore. This enforces the same directory permissions
2057 - Better handling terminating daemons with ctrl+c (used when running
2058 them in debugging mode).
2059 - Fix a race condition in live migrating a KVM instance, when stat()
2060 on the old proc status file returned EINVAL, which is an unexpected
2062 - Fixed manpage checking with newer man and utf-8 charachters. But now
2063 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2069 *(Released Fri, 7 May 2010)*
2071 Fix a bug which prevented untagged KVM instances from starting.
2077 *(Released Fri, 7 May 2010)*
2079 Another release with a long development cycle, during which many
2080 different features were added.
2082 Significant features
2083 ~~~~~~~~~~~~~~~~~~~~
2085 The KVM hypervisor now can run the individual instances as non-root, to
2086 reduce the impact of a VM being hijacked due to bugs in the
2087 hypervisor. It is possible to run all instances as a single (non-root)
2088 user, to manually specify a user for each instance, or to dynamically
2089 allocate a user out of a cluster-wide pool to each instance, with the
2090 guarantee that no two instances will run under the same user ID on any
2093 An experimental RAPI client library, that can be used standalone
2094 (without the other Ganeti libraries), is provided in the source tree as
2095 ``lib/rapi/client.py``. Note this client might change its interface in
2096 the future, as we iterate on its capabilities.
2098 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2099 replace the cluster's certificates and crypto keys. This might help in
2100 case they have been compromised, or have simply expired.
2102 A new disk option for instance creation has been added that allows one
2103 to "adopt" currently existing logical volumes, with data
2104 preservation. This should allow easier migration to Ganeti from
2105 unmanaged (or managed via other software) instances.
2107 Another disk improvement is the possibility to convert between redundant
2108 (DRBD) and plain (LVM) disk configuration for an instance. This should
2109 allow better scalability (starting with one node and growing the
2110 cluster, or shrinking a two-node cluster to one node).
2112 A new feature that could help with automated node failovers has been
2113 implemented: if a node sees itself as offline (by querying the master
2114 candidates), it will try to shutdown (hard) all instances and any active
2115 DRBD devices. This reduces the risk of duplicate instances if an
2116 external script automatically failovers the instances on such nodes. To
2117 enable this, the cluster parameter ``maintain_node_health`` should be
2118 enabled; in the future this option (per the name) will enable other
2119 automatic maintenance features.
2121 Instance export/import now will reuse the original instance
2122 specifications for all parameters; that means exporting an instance,
2123 deleting it and the importing it back should give an almost identical
2124 instance. Note that the default import behaviour has changed from
2125 before, where it created only one NIC; now it recreates the original
2128 Cluster verify has added a few new checks: SSL certificates validity,
2129 /etc/hosts consistency across the cluster, etc.
2134 As usual, many internal changes were done, documentation fixes,
2137 - Fixed cluster initialization with disabled cluster storage (regression
2138 introduced in 2.1.1)
2139 - File-based storage supports growing the disks
2140 - Fixed behaviour of node role changes
2141 - Fixed cluster verify for some corner cases, plus a general rewrite of
2142 cluster verify to allow future extension with more checks
2143 - Fixed log spamming by watcher and node daemon (regression introduced
2145 - Fixed possible validation issues when changing the list of enabled
2147 - Fixed cleanup of /etc/hosts during node removal
2148 - Fixed RAPI response for invalid methods
2149 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2150 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2151 from ide disks, etc.)
2152 - Allow OS changes without re-installation (to record a changed OS
2153 outside of Ganeti, or to allow OS renames)
2154 - Allow instance creation without OS installation (useful for example if
2155 the OS will be installed manually, or restored from a backup not in
2157 - Implemented option to make cluster ``copyfile`` use the replication
2159 - Added list of enabled hypervisors to ssconf (possibly useful for
2161 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2163 - A partial form of node re-IP is possible via node readd, which now
2164 allows changed node primary IP
2165 - Command line utilities now show an informational message if the job is
2167 - The logs of the master daemon now show the PID/UID/GID of the
2174 *(Released Fri, 12 Mar 2010)*
2176 During the 2.1.0 long release candidate cycle, a lot of improvements and
2177 changes have accumulated with were released later as 2.1.1.
2182 The node evacuate command (``gnt-node evacuate``) was significantly
2183 rewritten, and as such the IAllocator protocol was changed - a new
2184 request type has been added. This unfortunate change during a stable
2185 series is designed to improve performance of node evacuations; on
2186 clusters with more than about five nodes and which are well-balanced,
2187 evacuation should proceed in parallel for all instances of the node
2188 being evacuated. As such, any existing IAllocator scripts need to be
2189 updated, otherwise the above command will fail due to the unknown
2190 request. The provided "dumb" allocator has not been updated; but the
2191 ganeti-htools package supports the new protocol since version 0.2.4.
2193 Another important change is increased validation of node and instance
2194 names. This might create problems in special cases, if invalid host
2195 names are being used.
2197 Also, a new layer of hypervisor parameters has been added, that sits at
2198 OS level between the cluster defaults and the instance ones. This allows
2199 customisation of virtualization parameters depending on the installed
2200 OS. For example instances with OS 'X' may have a different KVM kernel
2201 (or any other parameter) than the cluster defaults. This is intended to
2202 help managing a multiple OSes on the same cluster, without manual
2203 modification of each instance's parameters.
2205 A tool for merging clusters, ``cluster-merge``, has been added in the
2206 tools sub-directory.
2211 - Improved the int/float conversions that should make the code more
2212 robust in face of errors from the node daemons
2213 - Fixed the remove node code in case of internal configuration errors
2214 - Fixed the node daemon behaviour in face of inconsistent queue
2215 directory (e.g. read-only file-system where we can't open the files
2217 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2218 now it either aborts cleanly or, if given the new "auto_promote"
2219 parameter, will automatically promote other nodes as needed
2220 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2221 completely prevent Ganeti from working
2222 - Fixed bug for instance export when not all disks were successfully
2224 - Fixed behaviour of node add when the new node is slow in starting up
2226 - Fixed handling of signals in the LUXI client, which should improve
2227 behaviour of command-line scripts
2228 - Added checks for invalid node/instance names in the configuration (now
2229 flagged during cluster verify)
2230 - Fixed watcher behaviour for disk activation errors
2231 - Fixed two potentially endless loops in http library, which led to the
2232 RAPI daemon hanging and consuming 100% CPU in some cases
2233 - Fixed bug in RAPI daemon related to hashed passwords
2234 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2236 - Enhanced compatibility with non-Debian OSes, but not using absolute
2237 path in some commands and allowing customisation of the ssh
2238 configuration directory
2239 - Fixed possible future issue with new Python versions by abiding to the
2240 proper use of ``__slots__`` attribute on classes
2241 - Added checks that should prevent directory traversal attacks
2242 - Many documentation fixes based on feedback from users
2247 - Added an "early_release" more for instance replace disks and node
2248 evacuate, where we release locks earlier and thus allow higher
2249 parallelism within the cluster
2250 - Added watcher hooks, intended to allow the watcher to restart other
2251 daemons (e.g. from the ganeti-nbma project), but they can be used of
2252 course for any other purpose
2253 - Added a compile-time disable for DRBD barriers, to increase
2254 performance if the administrator trusts the power supply or the
2255 storage system to not lose writes
2256 - Added the option of using syslog for logging instead of, or in
2257 addition to, Ganeti's own log files
2258 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2259 can indeed boot from a paravirtual NIC
2260 - Added a generic debug level for many operations; while this is not
2261 used widely yet, it allows one to pass the debug value all the way to
2263 - Enhanced the hooks environment for instance moves (failovers,
2264 migrations) where the primary/secondary nodes changed during the
2265 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2266 - Enhanced data validations for many user-supplied values; one important
2267 item is the restrictions imposed on instance and node names, which
2268 might reject some (invalid) host names
2269 - Add a configure-time option to disable file-based storage, if it's not
2270 needed; this allows greater security separation between the master
2271 node and the other nodes from the point of view of the inter-node RPC
2273 - Added user notification in interactive tools if job is waiting in the
2274 job queue or trying to acquire locks
2275 - Added log messages when a job is waiting for locks
2276 - Added filtering by node tags in instance operations which admit
2277 multiple instances (start, stop, reboot, reinstall)
2278 - Added a new tool for cluster mergers, ``cluster-merge``
2279 - Parameters from command line which are of the form ``a=b,c=d`` can now
2280 use backslash escapes to pass in values which contain commas,
2281 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2283 - For KVM, the instance name is the first parameter passed to KVM, so
2284 that it's more visible in the process list
2290 *(Released Tue, 2 Mar 2010)*
2292 Ganeti 2.1 brings many improvements with it. Major changes:
2294 - Added infrastructure to ease automated disk repairs
2295 - Added new daemon to export configuration data in a cheaper way than
2296 using the remote API
2297 - Instance NICs can now be routed instead of being associated with a
2299 - Improved job locking logic to reduce impact of jobs acquiring multiple
2300 locks waiting for other long-running jobs
2302 In-depth implementation details can be found in the Ganeti 2.1 design
2308 - Added chroot hypervisor
2309 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2311 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2312 ``bootloader_path`` and ``bootloader_args``)
2313 - Added the ``use_localtime`` option for the xen-hvm and kvm
2314 hypervisors, and the default value for this has changed to false (in
2315 2.0 xen-hvm always enabled it)
2316 - Added luxi call to submit multiple jobs in one go
2317 - Added cluster initialization option to not modify ``/etc/hosts``
2319 - Added network interface parameters
2320 - Added dry run mode to some LUs
2321 - Added RAPI resources:
2323 - ``/2/instances/[instance_name]/info``
2324 - ``/2/instances/[instance_name]/replace-disks``
2325 - ``/2/nodes/[node_name]/evacuate``
2326 - ``/2/nodes/[node_name]/migrate``
2327 - ``/2/nodes/[node_name]/role``
2328 - ``/2/nodes/[node_name]/storage``
2329 - ``/2/nodes/[node_name]/storage/modify``
2330 - ``/2/nodes/[node_name]/storage/repair``
2332 - Added OpCodes to evacuate or migrate all instances on a node
2333 - Added new command to list storage elements on nodes (``gnt-node
2334 list-storage``) and modify them (``gnt-node modify-storage``)
2335 - Added new ssconf files with master candidate IP address
2336 (``ssconf_master_candidates_ips``), node primary IP address
2337 (``ssconf_node_primary_ips``) and node secondary IP address
2338 (``ssconf_node_secondary_ips``)
2339 - Added ``ganeti-confd`` and a client library to query the Ganeti
2340 configuration via UDP
2341 - Added ability to run hooks after cluster initialization and before
2343 - Added automatic mode for disk replace (``gnt-instance replace-disks
2345 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2346 after catastrophic data-loss
2347 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2349 - Added ``gnt-instance move`` command to move instances
2350 - Added ``gnt-cluster watcher`` command to control watcher
2351 - Added ``gnt-node powercycle`` command to powercycle nodes
2352 - Added new job status field ``lock_status``
2353 - Added parseable error codes to cluster verification (``gnt-cluster
2354 verify --error-codes``) and made output less verbose (use
2355 ``--verbose`` to restore previous behaviour)
2356 - Added UUIDs to the main config entities (cluster, nodes, instances)
2357 - Added support for OS variants
2358 - Added support for hashed passwords in the Ganeti remote API users file
2360 - Added option to specify maximum timeout on instance shutdown
2361 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2362 - Added new helper script to start and stop Ganeti daemons
2363 (``daemon-util``), with the intent to reduce the work necessary to
2364 adjust Ganeti for non-Debian distributions and to start/stop daemons
2366 - Added more unittests
2367 - Fixed critical bug in ganeti-masterd startup
2368 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2369 now customisable at the cluster level for both the KVM and Xen
2370 hypervisors using the new ``migration_port`` parameter
2371 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2372 reinstalling an instance
2373 - Allowed ``@`` in tag names
2374 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2375 - Many documentation updates
2376 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2377 - ``gnt-instance reinstall`` can now reinstall multiple instances
2378 - Updated many command line parameters
2379 - Introduced new OS API version 15
2380 - No longer support a default hypervisor
2381 - Treat virtual LVs as inexistent
2382 - Improved job locking logic to reduce lock contention
2383 - Match instance and node names case insensitively
2384 - Reimplemented bash completion script to be more complete
2391 *(Released Thu, 4 Feb 2010)*
2393 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2394 - Fix a string formatting bug
2395 - Improve safety of the code in some error paths
2396 - Improve data validation in the master of values returned from nodes
2402 *(Released Thu, 17 Dec 2009)*
2404 - Fix security issue due to missing validation of iallocator names; this
2405 allows local and remote execution of arbitrary executables
2406 - Fix failure of gnt-node list during instance removal
2407 - Ship the RAPI documentation in the archive
2413 *(Released Wed, 30 Sep 2009)*
2415 - Fixed many wrong messages
2416 - Fixed a few bugs related to the locking library
2417 - Fixed MAC checking at instance creation time
2418 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2419 - Fixed a few issues related to signal handling in both daemons and
2421 - Fixed the example startup script provided
2422 - Fixed insserv dependencies in the example startup script (patch from
2424 - Fixed handling of drained nodes in the iallocator framework
2425 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2427 - Fixed error related to invalid job IDs in job polling
2428 - Fixed job/opcode persistence on unclean master shutdown
2429 - Fixed handling of partial job processing after unclean master
2431 - Fixed error reporting from LUs, previously all errors were converted
2432 into execution errors
2433 - Fixed error reporting from burnin
2434 - Decreased significantly the memory usage of the job queue
2435 - Optimised slightly multi-job submission
2436 - Optimised slightly opcode loading
2437 - Backported the multi-job submit framework from the development
2438 branch; multi-instance start and stop should be faster
2439 - Added script to clean archived jobs after 21 days; this will reduce
2440 the size of the queue directory
2441 - Added some extra checks in disk size tracking
2442 - Added an example ethers hook script
2443 - Added a cluster parameter that prevents Ganeti from modifying of
2445 - Added more node information to RAPI responses
2446 - Added a ``gnt-job watch`` command that allows following the ouput of a
2448 - Added a bind-address option to ganeti-rapi
2449 - Added more checks to the configuration verify
2450 - Enhanced the burnin script such that some operations can be retried
2452 - Converted instance reinstall to multi-instance model
2458 *(Released Fri, 7 Aug 2009)*
2460 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2461 to allow using the pre-2.0.2 behaviour in activation, if any existing
2462 instances have mismatched disk sizes in the configuration
2463 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2464 any configuration mismatches for disk sizes
2465 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2466 failover to work on two-node clusters
2467 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2469 - Fixed detection of OS script errors in ``gnt-backup export``
2470 - Fixed exit code of ``gnt-backup export``
2476 *(Released Fri, 17 Jul 2009)*
2478 - Added experimental support for stripped logical volumes; this should
2479 enhance performance but comes with a higher complexity in the block
2480 device handling; stripping is only enabled when passing
2481 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2482 affected even in the non-stripped mode
2483 - Improved resiliency against transient failures at the end of DRBD
2484 resyncs, and in general of DRBD resync checks
2485 - Fixed a couple of issues with exports and snapshot errors
2486 - Fixed a couple of issues in instance listing
2487 - Added display of the disk size in ``gnt-instance info``
2488 - Fixed checking for valid OSes in instance creation
2489 - Fixed handling of the "vcpus" parameter in instance listing and in
2490 general of invalid parameters
2491 - Fixed http server library, and thus RAPI, to handle invalid
2492 username/password combinations correctly; this means that now they
2493 report unauthorized for queries too, not only for modifications,
2494 allowing earlier detect of configuration problems
2495 - Added a new "role" node list field, equivalent to the master/master
2496 candidate/drained/offline flags combinations
2497 - Fixed cluster modify and changes of candidate pool size
2498 - Fixed cluster verify error messages for wrong files on regular nodes
2499 - Fixed a couple of issues with node demotion from master candidate role
2500 - Fixed node readd issues
2501 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2502 - Added a new ``--no-voting`` option for masterfailover to fix failover
2503 on two-nodes clusters when the former master node is unreachable
2504 - Added instance reinstall over RAPI
2510 *(Released Tue, 16 Jun 2009)*
2512 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2513 allow re-adding the start in single-user option (regression from 1.2)
2514 - the watcher writes the instance status to a file, to allow monitoring
2515 to report the instance status (from the master) based on cached
2516 results of the watcher's queries; while this can get stale if the
2517 watcher is being locked due to other work on the cluster, this is
2518 still an improvement
2519 - the watcher now also restarts the node daemon and the rapi daemon if
2521 - fixed the watcher to handle full and drained queue cases
2522 - hooks export more instance data in the environment, which helps if
2523 hook scripts need to take action based on the instance's properties
2524 (no longer need to query back into ganeti)
2525 - instance failovers when the instance is stopped do not check for free
2526 RAM, so that failing over a stopped instance is possible in low memory
2528 - rapi uses queries for tags instead of jobs (for less job traffic), and
2529 for cluster tags it won't talk to masterd at all but read them from
2531 - a couple of error handling fixes in RAPI
2532 - drbd handling: improved the error handling of inconsistent disks after
2533 resync to reduce the frequency of "there are some degraded disks for
2534 this instance" messages
2535 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2536 error handling path called a wrong function name)
2542 *(Released Wed, 27 May 2009)*
2544 - no changes from rc5
2550 *(Released Wed, 20 May 2009)*
2552 - fix a couple of bugs (validation, argument checks)
2553 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2554 - some small improvements to RAPI and IAllocator
2555 - make watcher automatically start the master daemon if down
2561 *(Released Mon, 27 Apr 2009)*
2563 - change the OS list to not require locks; this helps with big clusters
2564 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2565 volume group is broken
2566 - ``gnt-instance info``, without any arguments, doesn't run for all
2567 instances anymore; either pass ``--all`` or pass the desired
2568 instances; this helps against mistakes on big clusters where listing
2569 the information for all instances takes a long time
2570 - miscellaneous doc and man pages fixes
2576 *(Released Wed, 8 Apr 2009)*
2578 - Change the internal locking model of some ``gnt-node`` commands, in
2579 order to reduce contention (and blocking of master daemon) when
2580 batching many creation/reinstall jobs
2581 - Fixes to Xen soft reboot
2582 - No longer build documentation at build time, instead distribute it in
2583 the archive, in order to reduce the need for the whole docbook/rst
2590 *(Released Fri, 27 Mar 2009)*
2592 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2593 - Fix watcher startup sequence, improves the behaviour of busy clusters
2594 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2595 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2596 bind address and other places
2597 - Some documentation fixes and updates
2603 *(Released Mon, 2 Mar 2009)*
2605 - More documentation updates, now all docs should be more-or-less
2607 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2609 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2611 - Fix an issue related to $libdir/run/ganeti and cluster creation
2617 *(Released Thu, 19 Feb 2009)*
2619 - Xen PVM and KVM have switched the default value for the instance root
2620 disk to the first partition on the first drive, instead of the whole
2621 drive; this means that the OS installation scripts must be changed
2623 - Man pages have been updated
2624 - RAPI has been switched by default to HTTPS, and the exported functions
2625 should all work correctly
2626 - RAPI v1 has been removed
2627 - Many improvements to the KVM hypervisor
2628 - Block device errors are now better reported
2629 - Many other bugfixes and small improvements
2635 *(Released Mon, 26 Jan 2009)*
2637 - Version 2 is a general rewrite of the code and therefore the
2638 differences are too many to list, see the design document for 2.0 in
2639 the ``doc/`` subdirectory for more details
2640 - In this beta version there is not yet a migration path from 1.2 (there
2641 will be one in the final 2.0 release)
2642 - A few significant changes are:
2644 - all commands are executed by a daemon (``ganeti-masterd``) and the
2645 various ``gnt-*`` commands are just front-ends to it
2646 - all the commands are entered into, and executed from a job queue,
2647 see the ``gnt-job(8)`` manpage
2648 - the RAPI daemon supports read-write operations, secured by basic
2649 HTTP authentication on top of HTTPS
2650 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2651 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2652 to migrate to DRBD 8 first while still running Ganeti 1.2)
2653 - DRBD devices are using statically allocated minor numbers, which
2654 will be assigned to existing instances during the migration process
2655 - there is support for both Xen PVM and Xen HVM instances running on
2657 - KVM virtualization is supported too
2658 - file-based storage has been implemented, which means that it is
2659 possible to run the cluster without LVM and DRBD storage, for
2660 example using a shared filesystem exported from shared storage (and
2661 still have live migration)
2667 *(Released Tue, 13 Jan 2009)*
2669 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2670 - Reuse the old instance mac address by default on instance import, if
2671 the instance name is the same.
2672 - Handle situations in which the node info rpc returns incomplete
2674 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2675 - Improved version of batcher:
2677 - state file support
2678 - instance mac address support
2679 - support for HVM clusters/instances
2681 - Add an option to show the number of cpu sockets and nodes in
2683 - Support OSes that handle more than one version of the OS api (but do
2684 not change the current API in any other way)
2685 - Fix ``gnt-node migrate``
2686 - ``gnt-debug`` man page
2687 - Fixes various more typos and small issues
2688 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2694 *(Released Wed, 24 Sep 2008)*
2696 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2697 type of disk exported to fully virtualized instances.
2698 - provide access to the serial console of HVM instances
2699 - instance auto_balance flag, set by default. If turned off it will
2700 avoid warnings on cluster verify if there is not enough memory to fail
2701 over an instance. in the future it will prevent automatically failing
2702 it over when we will support that.
2703 - batcher tool for instance creation, see ``tools/README.batcher``
2704 - ``gnt-instance reinstall --select-os`` to interactively select a new
2705 operating system when reinstalling an instance.
2706 - when changing the memory amount on instance modify a check has been
2707 added that the instance will be able to start. also warnings are
2708 emitted if the instance will not be able to fail over, if auto_balance
2710 - documentation fixes
2711 - sync fields between ``gnt-instance list/modify/add/import``
2712 - fix a race condition in drbd when the sync speed was set after giving
2713 the device a remote peer.
2719 *(Released Tue, 22 Jul 2008)*
2721 - note: the allowed size and number of tags per object were reduced
2722 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2723 - fixed twisted 8.x compatibility
2724 - fixed ``gnt-instance replace-disks`` with iallocator
2725 - add TCP keepalives on twisted connections to detect restarted nodes
2726 - disk increase support, see ``gnt-instance grow-disk``
2727 - implement bulk node/instance query for RAPI
2728 - add tags in node/instance listing (optional)
2729 - experimental migration (and live migration) support, read the man page
2730 for ``gnt-instance migrate``
2731 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2732 has some small improvements in handling its state file
2738 *(Released Fri, 13 Jun 2008)*
2740 - Experimental readonly, REST-based remote API implementation;
2741 automatically started on master node, TCP port 5080, if enabled by
2742 ``--enable-rapi`` parameter to configure script.
2743 - Instance allocator support. Add and import instance accept a
2744 ``--iallocator`` parameter, and call that instance allocator to decide
2745 which node to use for the instance. The iallocator document describes
2746 what's expected from an allocator script.
2747 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2748 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2749 node is lost there is still enough memory to fail over the instances
2751 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2752 ``gnt-cluster verify``, to check for site-specific compliance. All the
2753 hooks will run, and their output, if any, will be displayed. Any
2754 failing hook will make the verification return an error value.
2755 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2756 primary and secondary interfaces
2757 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2758 that is still declared as part of the cluster and has failed.
2759 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2760 specifying output fields, that just adds the chosen fields to the
2762 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2763 export from the filesystem.
2764 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2765 have been added. Using them you can enable/disable acpi and pae
2766 support, and specify a path for a cd image to be exported to the
2767 instance. These parameters as the name suggest only work on HVM
2769 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2770 PAE support will be set to the previously hardcoded values, but the
2771 (previously hardcoded) path to the CDROM ISO image will be unset and
2772 if required, needs to be set manually with ``gnt-instance modify``
2774 - The address to which an instance's VNC console is bound is now
2775 selectable per-instance, rather than being cluster wide. Of course
2776 this only applies to instances controlled via VNC, so currently just
2777 applies to HVM clusters.
2783 *(Released Mon, 18 Feb 2008)*
2785 - more tweaks to the disk activation code (especially helpful for DRBD)
2786 - change the default ``gnt-instance list`` output format, now there is
2787 one combined status field (see the manpage for the exact values this
2789 - some more fixes for the mac export to hooks change
2790 - make Ganeti not break with DRBD 8.2.x (which changed the version
2791 format in ``/proc/drbd``) (issue 24)
2792 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2793 template, allowing migration from DRBD0.7+MD to DRBD8
2799 *(Released Wed, 30 Jan 2008)*
2801 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2803 - add command aliases infrastructure and a few aliases
2804 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2805 man pages and the ``--help`` option of ``gnt-node
2806 list``/``gnt-instance list``
2807 - fix ``gnt-backup list`` with down nodes (issue 21)
2808 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2809 - fix the dist archive and add a check for including svn/git files in
2811 - some developer-related changes: improve the burnin and the QA suite,
2812 add an upload script for testing during development
2818 *(Released Wed, 16 Jan 2008)*
2820 - experimental HVM support, read the install document, section
2821 "Initializing the cluster"
2822 - allow for the PVM hypervisor per-instance kernel and initrd paths
2823 - add a new command ``gnt-cluster verify-disks`` which uses a new
2824 algorithm to improve the reconnection of the DRBD pairs if the device
2825 on the secondary node has gone away
2826 - make logical volume code auto-activate LVs at disk activation time
2827 - slightly improve the speed of activating disks
2828 - allow specification of the MAC address at instance creation time, and
2829 changing it later via ``gnt-instance modify``
2830 - fix handling of external commands that generate lots of output on
2832 - update documentation with regard to minimum version of DRBD8 supported
2838 *(Released Tue, 4 Dec 2007)*
2840 - Log the ``xm create`` output to the node daemon log on failure (to
2841 help diagnosing the error)
2842 - In debug mode, log all external commands output if failed to the logs
2843 - Change parsing of lvm commands to ignore stderr
2849 *(Released Wed, 28 Nov 2007)*
2851 - Another round of updates to the DRBD 8 code to deal with more failures
2852 in the replace secondary node operation
2853 - Some more logging of failures in disk operations (lvm, drbd)
2854 - A few documentation updates
2861 *(Released Tue, 13 Nov 2007)*
2863 - Change configuration file format from Python's Pickle to JSON.
2864 Upgrading is possible using the cfgupgrade utility.
2865 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2866 faster replace disks and is more stable (DRBD 8 has many improvements
2867 compared to DRBD 0.7)
2868 - Added command line tags support (see man pages for ``gnt-instance``,
2869 ``gnt-node``, ``gnt-cluster``)
2870 - Added instance rename support
2871 - Added multi-instance startup/shutdown
2872 - Added cluster rename support
2873 - Added ``gnt-node evacuate`` to simplify some node operations
2874 - Added instance reboot operation that can speedup reboot as compared to
2876 - Soften the requirement that hostnames are in FQDN format
2877 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2879 - Removed dependency on debian's patched fping that uses the
2880 non-standard ``-S`` option
2881 - Now the OS definitions are searched for in multiple, configurable
2882 paths (easier for distros to package)
2883 - Some changes to the hooks infrastructure (especially the new
2884 post-configuration update hook)
2885 - Other small bugfixes
2887 .. vim: set textwidth=72 syntax=rst :