10 Incompatible/important changes
11 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
13 - Adding disks with 'gnt-instance modify' now waits for the disks to sync per
14 default. Specify --no-wait-for-sync to override this behavior.
15 - The Ganeti python code now adheres to a private-module layout. In particular,
16 the module 'ganeti' is no longer in the python search path.
17 - On instance allocation, the iallocator now considers non-LVM storage
18 properly. In particular, actual file storage space information is used
19 when allocating space for a file/sharedfile instance.
23 - KVM hypervisors can now access RBD storage directly without having to
24 go through a block device.
25 - A new command 'gnt-cluster upgrade' was added that automates the upgrade
26 procedure between two Ganeti versions that are both 2.10 or higher.
32 *(Released Wed, 9 Oct 2013)*
34 Incompatible/important changes
35 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
37 - hroller now also plans for capacity to move non-redundant instances off
38 any node to be rebooted; the old behavior of completely ignoring any
39 non-redundant instances can be restored by adding the --ignore-non-redundant
41 - The cluster option '--no-lvm-storage' was removed in favor of the new option
42 '--enabled-disk-templates'.
43 - On instance creation, disk templates no longer need to be specified
44 with '-t'. The default disk template will be taken from the list of
45 enabled disk templates.
46 - The monitoring daemon is now running as root, in order to be able to collect
47 information only available to root (such as the state of Xen instances).
48 - The ConfD client is now IPv6 compatible.
49 - File and shared file storage is no longer dis/enabled at configure time,
50 but using the option '--enabled-disk-templates' at cluster initialization and
52 - The default directories for file and shared file storage are not anymore
53 specified at configure time, but taken from the cluster's configuration.
54 They can be set at cluster initialization and modification with
55 '--file-storage-dir' and '--shared-file-storage-dir'.
56 - Cluster verification now includes stricter checks regarding the
57 default file and shared file storage directories. It now checks that
58 the directories are explicitely allowed in the 'file-storage-paths' file and
59 that the directories exist on all nodes.
60 - The list of allowed disk templates in the instance policy and the list
61 of cluster-wide enabled disk templates is now checked for consistency
62 on cluster or group modification. On cluster initialization, the ipolicy
63 disk templates are ensured to be a subset of the cluster-wide enabled
69 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
70 the correct command syntax. It is possible to use different DRBD versions
71 on different nodes as long as they are compatible to each other. This
72 enables rolling upgrades of DRBD with no downtime. As permanent operation
73 of different DRBD versions within a node group is discouraged,
74 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
75 - New "inst-status-xen" data collector for the monitoring daemon, providing
76 information about the state of the xen instances on the nodes.
77 - New "lv" data collector for the monitoring daemon, collecting data about the
78 logical volumes on the nodes, and pairing them with the name of the instances
80 - New "diskstats" data collector, collecting the data from /proc/diskstats and
81 presenting them over the monitoring daemon interface.
82 - The ConfD client is now IPv6 compatible.
86 The following new dependencies have been added.
90 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
91 for the unit tests (and only used for testing).
95 - ``hslogger`` (http://software.complete.org/hslogger) is now always
96 required, even if confd is not enabled.
101 - Fix bug in cfgupgrade that led to failure when upgrading from 2.8 with
102 at least one DRBD instance.
103 - Fix bug in cfgupgrade that led to an invalid 2.8 configuration after
110 *(Released Tue, 1 Oct 2013)*
112 The first release candidate in the 2.9 series. Since 2.9.0 beta1:
115 - update of the documentation, in particular installation instructions
116 - merging of LD_* constants into DT_* constants
117 - python style changes to be compatible with newer versions of pylint
123 *(Released Thu, 29 Aug 2013)*
125 This was the first beta release of the 2.9 series. All important changes
126 are listed in the latest 2.9 entry.
132 *(Released Mon, 30 Sep 2013)*
134 Incompatible/important changes
135 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
137 - Instance policy can contain multiple instance specs, as described in
138 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
139 <design-partitioned>`. As a consequence, it's not possible to partially change
140 or override instance specs. Bounding specs (min and max) can be specified as a
141 whole using the new option ``--ipolicy-bounds-specs``, while standard
142 specs use the new option ``--ipolicy-std-specs``.
143 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
144 gnt-instance is a valid YAML object.
145 - hail now honors network restrictions when allocating nodes. This led to an
146 update of the IAllocator protocol. See the IAllocator documentation for
148 - confd now only answers static configuration request over the network. luxid
149 was extracted, listens on the local LUXI socket and responds to live queries.
150 This allows finer grained permissions if using separate users.
155 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
156 to always require authentication, ``--require-authentication``. It can
157 be specified in ``$sysconfdir/default/ganeti``.
158 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
159 be used to manage the disk templates to be used by instances in the cluster.
160 Initially, it will be set to a list that includes plain, drbd, if they were
161 enabled by specifying a volume group name, and file and sharedfile, if those
162 were enabled at configure time. Additionally, it will include all disk
163 templates that are currently used by instances. The order of disk templates
164 will be based on Ganeti's history of supporting them. In the future, the
165 first entry of the list will be used as a default disk template on instance
167 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
168 configuration back to the previous stable version.
169 - Disk templates in group ipolicy can be restored to the default value.
170 - Initial support for diskless instances and virtual clusters in QA.
171 - More QA and unit tests for instance policies.
172 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
173 describing why the opcode itself was executed.
174 - The monitoring daemon is now available. It allows users to query the cluster
175 for obtaining information about the status of the system. The daemon is only
176 responsible for providing the information over the network: the actual data
177 gathering is performed by data collectors (currently, only the DRBD status
178 collector is available).
179 - In order to help developers work on Ganeti, a new script
180 (``devel/build_chroot``) is provided, for building a chroot that contains all
181 the required development libraries and tools for compiling Ganeti on a Debian
183 - A new tool, ``harep``, for performing self-repair and recreation of instances
184 in Ganeti has been added.
185 - Split queries are enabled for tags, network, exports, cluster info, groups,
187 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
188 It prints the command line to set the current policies, to ease
190 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
191 devices for KVM virtio-net interfaces will get created with VNET_HDR
192 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
193 virtio-net interfaces, which prevents host kernel tainting and log
194 flooding, when dealing with broken or malicious virtio-net drivers.
195 It's set to true by default.
196 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
198 - Support 'viridian' parameter in Xen HVM
199 - Support DSA SSH keys in bootstrap
200 - To simplify the work of packaging frameworks that want to add the needed users
201 and groups in a split-user setup themselves, at build time three files in
202 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
203 the groups to be generated, the ``users`` file contains, one per line, the
204 users to be generated, optionally followed by their primary group, where
205 important. The ``groupmemberships`` file contains, one per line, additional
206 user-group membership relations that need to be established. The syntax of
207 these files will remain stable in all future versions.
212 The following new dependencies have been added:
215 - The ``curl`` library is not optional anymore for compiling the Haskell code.
216 - ``snap-server`` library (if monitoring is enabled).
219 - The minimum Python version needed to run Ganeti is now 2.6.
220 - ``yaml`` library (only for running the QA).
224 - Perform proper cleanup on termination of Haskell daemons
225 - Fix corner-case in handling of remaining retry time
231 *(Released Tue, 17 Sep 2013)*
233 - To simplify the work of packaging frameworks that want to add the needed users
234 and groups in a split-user setup themselves, at build time three files in
235 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
236 the groups to be generated, the ``users`` file contains, one per line, the
237 users to be generated, optionally followed by their primary group, where
238 important. The ``groupmemberships`` file contains, one per line, additional
239 user-group membership relations that need to be established. The syntax of
240 these files will remain stable in all future versions.
241 - Add a default to file-driver when unspecified over RAPI (Issue 571)
242 - Mark the DSA host pubkey as optional, and remove it during config downgrade
244 - Some documentation fixes
250 *(Released Tue, 27 Aug 2013)*
252 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
254 - Support 'viridian' parameter in Xen HVM (Issue 233)
255 - Include VCS version in ``gnt-cluster version``
256 - Support DSA SSH keys in bootstrap (Issue 338)
257 - Fix batch creation of instances
258 - Use FQDN to check master node status (Issue 551)
259 - Make the DRBD collector more failure-resilient
265 *(Released Fri, 2 Aug 2013)*
267 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
269 - Fix upgrading/downgrading from 2.7
270 - Increase maximum RAPI message size
271 - Documentation updates
272 - Split ``confd`` between ``luxid`` and ``confd``
273 - Merge 2.7 series up to the 2.7.1 release
274 - Allow the ``modify_etc_hosts`` option to be changed
275 - Add better debugging for ``luxid`` queries
276 - Expose bulk parameter for GetJobs in RAPI client
277 - Expose missing ``network`` fields in RAPI
278 - Add some ``cluster verify`` tests
279 - Some unittest fixes
280 - Fix a malfunction in ``hspace``'s tiered allocation
281 - Fix query compatibility between haskell and python implementations
282 - Add the ``vnet_hdr`` HV parameter for KVM
283 - Add ``--cleanup`` to instance failover
284 - Change the connected groups format in ``gnt-network info`` output; it
285 was previously displayed as a raw list by mistake. (Merged from 2.7)
291 *(Released Mon, 24 Jun 2013)*
293 This was the first beta release of the 2.8 series. All important changes
294 are listed in the latest 2.8 entry.
300 *(Released Thu, 26 Sep 2013)*
302 - Change the connected groups format in ``gnt-network info`` output; it
303 was previously displayed as a raw list by mistake
304 - Check disk template in right dict when copying
305 - Support multi-instance allocs without iallocator
306 - Fix some errors in the documentation
307 - Fix formatting of tuple in an error message
313 *(Released Thu, 25 Jul 2013)*
315 - Add logrotate functionality in daemon-util
316 - Add logrotate example file
317 - Add missing fields to network queries over rapi
318 - Fix network object timestamps
319 - Add support for querying network timestamps
320 - Fix a typo in the example crontab
321 - Fix a documentation typo
327 *(Released Thu, 04 Jul 2013)*
329 Incompatible/important changes
330 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
332 - Instance policies for disk size were documented to be on a per-disk
333 basis, but hail applied them to the sum of all disks. This has been
335 - ``hbal`` will now exit with status 0 if, during job execution over
336 LUXI, early exit has been requested and all jobs are successful;
337 before, exit status 1 was used, which cannot be differentiated from
339 - Compatibility with newer versions of rbd has been fixed
340 - ``gnt-instance batch-create`` has been changed to use the bulk create
341 opcode from Ganeti. This lead to incompatible changes in the format of
342 the JSON file. It's now not a custom dict anymore but a dict
343 compatible with the ``OpInstanceCreate`` opcode.
344 - Parent directories for file storage need to be listed in
345 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
346 write the file automatically based on old configuration values, but it
347 can not distribute it across all nodes and the file contents should be
348 verified. Use ``gnt-cluster copyfile
349 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
350 upgraded. The reason for requiring this list of paths now is that
351 before it would have been possible to inject new paths via RPC,
352 allowing files to be created in arbitrary locations. The RPC protocol
353 is protected using SSL/X.509 certificates, but as a design principle
354 Ganeti does not permit arbitrary paths to be passed.
355 - The parsing of the variants file for OSes (see
356 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
357 lines and comment lines (starting with ``#``) are ignored for better
359 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
360 longer available. ``gnt-node add`` now invokes a new tool on the
361 destination node, named ``prepare-node-join``, to configure the SSH
362 daemon. Paramiko is no longer necessary to configure nodes' SSH
363 daemons via ``gnt-node add``.
364 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
365 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
366 the flag is not reset after a master failover.
367 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
368 is recommended. 2.6 will be mandatory from the 2.8 series.
374 - New network management functionality to support automatic allocation
375 of IP addresses and managing of network parameters. See
376 :manpage:`gnt-network(8)` for more details.
377 - New external storage backend, to allow managing arbitrary storage
378 systems external to the cluster. See
379 :manpage:`ganeti-extstorage-interface(7)`.
380 - New ``exclusive-storage`` node parameter added, restricted to
381 nodegroup level. When it's set to true, physical disks are assigned in
382 an exclusive fashion to instances, as documented in :doc:`Partitioned
383 Ganeti <design-partitioned>`. Currently, only instances using the
384 ``plain`` disk template are supported.
385 - The KVM hypervisor has been updated with many new hypervisor
386 parameters, including a generic one for passing arbitrary command line
387 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
388 compatible up to qemu 1.4.
389 - A new tool, called ``mon-collector``, is the stand-alone executor of
390 the data collectors for a monitoring system. As of this version, it
391 just includes the DRBD data collector, that can be executed by calling
392 ``mon-collector`` using the ``drbd`` parameter. See
393 :manpage:`mon-collector(7)`.
394 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
395 for RAPI users. It allows granting permissions to query for
396 information to a specific user without giving
397 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
398 - A new tool named ``node-cleanup`` has been added. It cleans remains of
399 a cluster from a machine by stopping all daemons, removing
400 certificates and ssconf files. Unless the ``--no-backup`` option is
401 given, copies of the certificates are made.
402 - Instance creations now support the use of opportunistic locking,
403 potentially speeding up the (parallel) creation of multiple instances.
404 This feature is currently only available via the :doc:`RAPI
405 <rapi>` interface and when an instance allocator is used. If the
406 ``opportunistic_locking`` parameter is set the opcode will try to
407 acquire as many locks as possible, but will not wait for any locks
408 held by other opcodes. If not enough resources can be found to
409 allocate the instance, the temporary error code
410 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
411 retried thereafter, with or without opportunistic locking.
412 - New experimental linux-ha resource scripts.
413 - Restricted-commands support: ganeti can now be asked (via command line
414 or rapi) to perform commands on a node. These are passed via ganeti
415 RPC rather than ssh. This functionality is restricted to commands
416 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
417 reasons. The file is not copied automatically.
423 - Diskless instances are now externally mirrored (Issue 237). This for
424 now has only been tested in conjunction with explicit target nodes for
426 - Queries not needing locks or RPC access to the node can now be
427 performed by the confd daemon, making them independent from jobs, and
428 thus faster to execute. This is selectable at configure time.
429 - The functionality for allocating multiple instances at once has been
430 overhauled and is now also available through :doc:`RAPI <rapi>`.
432 There are no significant changes from version 2.7.0~rc3.
438 *(Released Tue, 25 Jun 2013)*
440 - Fix permissions on the confd query socket (Issue 477)
441 - Fix permissions on the job archive dir (Issue 498)
442 - Fix handling of an internal exception in replace-disks (Issue 472)
443 - Fix gnt-node info handling of shortened names (Issue 497)
444 - Fix gnt-instance grow-disk when wiping is enabled
445 - Documentation improvements, and support for newer pandoc
446 - Fix hspace honoring ipolicy for disks (Issue 484)
447 - Improve handling of the ``kvm_extra`` HV parameter
453 *(Released Fri, 24 May 2013)*
455 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
457 - Disks added through ``gnt-instance modify`` or created through
458 ``gnt-instance recreate-disks`` are wiped, if the
459 ``prealloc_wipe_disks`` flag is set.
460 - If wiping newly created disks fails, the disks are removed. Also,
461 partial failures in creating disks through ``gnt-instance modify``
462 triggers a cleanup of the partially-created disks.
463 - Removing the master IP address doesn't fail if the address has been
465 - Fix ownership of the OS log dir
466 - Workaround missing SO_PEERCRED constant (Issue 191)
472 *(Released Fri, 3 May 2013)*
474 This was the first release candidate of the 2.7 series. Since beta3:
476 - Fix kvm compatibility with qemu 1.4 (Issue 389)
477 - Documentation updates (admin guide, upgrade notes, install
478 instructions) (Issue 372)
479 - Fix gnt-group list nodes and instances count (Issue 436)
480 - Fix compilation without non-mandatory libraries (Issue 441)
481 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
482 - Make confd logging more verbose at INFO level (Issue 435)
483 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
484 - Fix failure path for instance storage type conversion (Issue 229)
485 - Update htools text backend documentation
486 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
487 - Disable inter-cluster instance move for file-based instances, because
488 it is dependant on instance export, which is not supported for
489 file-based instances. (Issue 414)
490 - Fix gnt-job crashes on non-ascii characters (Issue 427)
491 - Fix volume group checks on non-vm-capable nodes (Issue 432)
497 *(Released Mon, 22 Apr 2013)*
499 This was the third beta release of the 2.7 series. Since beta2:
501 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
502 - Fix data loss on wrong usage of ``gnt-instance move``
503 - Properly export errors in confd-based job queries
504 - Add ``users-setup`` tool
505 - Fix iallocator protocol to report 0 as a disk size for diskless
506 instances. This avoids hail breaking when a diskless instance is
508 - Fix job queue directory permission problem that made confd job queries
509 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
510 for access to archived jobs (Issue 406).
511 - Limit the sizes of networks supported by ``gnt-network`` to something
512 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
513 - Fix bugs in instance disk template conversion
514 - Fix GHC 7 compatibility
515 - Fix ``burnin`` install path (Issue 426).
516 - Allow very small disk grows (Issue 347).
517 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
518 that noded doesn't import masterd code (Issue 419).
519 - Make sure the default metavg at cluster init is the same as the vg, if
520 unspecified (Issue 358).
521 - Fix cleanup of partially created disks (part of Issue 416)
527 *(Released Tue, 2 Apr 2013)*
529 This was the second beta release of the 2.7 series. Since beta1:
531 - Networks no longer have a "type" slot, since this information was
532 unused in Ganeti: instead of it tags should be used.
533 - The rapi client now has a ``target_node`` option to MigrateInstance.
534 - Fix early exit return code for hbal (Issue 386).
535 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
536 - Fix ``rbd showmapped`` output parsing (Issue 312).
537 - Networks are now referenced indexed by UUID, rather than name. This
538 will require running cfgupgrade, from 2.7.0beta1, if networks are in
540 - The OS environment now includes network information.
541 - Deleting of a network is now disallowed if any instance nic is using
542 it, to prevent dangling references.
543 - External storage is now documented in man pages.
544 - The exclusive_storage flag can now only be set at nodegroup level.
545 - Hbal can now submit an explicit priority with its jobs.
546 - Many network related locking fixes.
547 - Bump up the required pylint version to 0.25.1.
548 - Fix the ``no_remember`` option in RAPI client.
549 - Many ipolicy related tests, qa, and fixes.
550 - Many documentation improvements and fixes.
551 - Fix building with ``--disable-file-storage``.
552 - Fix ``-q`` option in htools, which was broken if passed more than
554 - Some haskell/python interaction improvements and fixes.
555 - Fix iallocator in case of missing LVM storage.
556 - Fix confd config load in case of ``--no-lvm-storage``.
557 - The confd/query functionality is now mentioned in the security
564 *(Released Wed, 6 Feb 2013)*
566 This was the first beta release of the 2.7 series. All important changes
567 are listed in the latest 2.7 entry.
573 *(Released Fri, 21 Dec 2012)*
575 Important behaviour change: hbal won't rebalance anymore instances which
576 have the ``auto_balance`` attribute set to false. This was the intention
577 all along, but until now it only skipped those from the N+1 memory
578 reservation (DRBD-specific).
580 A significant number of bug fixes in this release:
582 - Fixed disk adoption interaction with ipolicy checks.
583 - Fixed networking issues when instances are started, stopped or
584 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
585 - Fixed the warning in cluster verify for shared storage instances not
587 - Fixed removal of storage directory on shared file storage (issue 262).
588 - Fixed validation of LVM volume group name in OpClusterSetParams
589 (``gnt-cluster modify``) (issue 285).
590 - Fixed runtime memory increases (``gnt-instance modify -m``).
591 - Fixed live migration under Xen's ``xl`` mode.
592 - Fixed ``gnt-instance console`` with ``xl``.
593 - Fixed building with newer Haskell compiler/libraries.
594 - Fixed PID file writing in Haskell daemons (confd); this prevents
595 restart issues if confd was launched manually (outside of
596 ``daemon-util``) while another copy of it was running
597 - Fixed a type error when doing live migrations with KVM (issue 297) and
598 the error messages for failing migrations have been improved.
599 - Fixed opcode validation for the out-of-band commands (``gnt-node
601 - Fixed a type error when unsetting OS hypervisor parameters (issue
602 311); now it's possible to unset all OS-specific hypervisor
604 - Fixed the ``dry-run`` mode for many operations: verification of
605 results was over-zealous but didn't take into account the ``dry-run``
606 operation, resulting in "wrong" failures.
607 - Fixed bash completion in ``gnt-job list`` when the job queue has
608 hundreds of entries; especially with older ``bash`` versions, this
609 results in significant CPU usage.
611 And lastly, a few other improvements have been made:
613 - Added option to force master-failover without voting (issue 282).
614 - Clarified error message on lock conflict (issue 287).
615 - Logging of newly submitted jobs has been improved (issue 290).
616 - Hostname checks have been made uniform between instance rename and
618 - The ``--submit`` option is now supported by ``gnt-debug delay``.
619 - Shutting down the master daemon by sending SIGTERM now stops it from
620 processing jobs waiting for locks; instead, those jobs will be started
621 once again after the master daemon is started the next time (issue
623 - Support for Xen's ``xl`` program has been improved (besides the fixes
625 - Reduced logging noise in the Haskell confd daemon (only show one log
626 entry for each config reload, instead of two).
627 - Several man page updates and typo fixes.
633 *(Released Fri, 12 Oct 2012)*
635 A small bugfix release. Among the bugs fixed:
637 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
638 made the command unusable.
639 - Commands that issue many jobs don't fail anymore just because some jobs
640 take so long that other jobs are archived.
641 - Failures during ``gnt-instance reinstall`` are reflected by the exit
643 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
645 - When ``always_failover`` is set, ``--allow-failover`` is not required
646 in migrate commands anymore.
647 - ``bash_completion`` works even if extglob is disabled.
648 - Fixed bug with locks that made failover for RDB-based instances fail.
649 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
650 a random node instead of one based on the allocator metric.
651 - Support for newer versions of pylint and pep8.
652 - Hail doesn't fail anymore when trying to add an instance of type
653 ``file``, ``sharedfile`` or ``rbd``.
654 - Added new Makefile target to rebuild the whole distribution, so that
655 all files are included.
661 *(Released Fri, 27 Jul 2012)*
664 .. attention:: The ``LUXI`` protocol has been made more consistent
665 regarding its handling of command arguments. This, however, leads to
666 incompatibility issues with previous versions. Please ensure that you
667 restart Ganeti daemons soon after the upgrade, otherwise most
668 ``LUXI`` calls (job submission, setting/resetting the drain flag,
669 pausing/resuming the watcher, cancelling and archiving jobs, querying
670 the cluster configuration) will fail.
679 The current ``admin_up`` field, which used to denote whether an instance
680 should be running or not, has been removed. Instead, ``admin_state`` is
681 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
683 The rational behind this is that an instance being “down” can have
686 - it could be down during a reboot
687 - it could be temporarily be down for a reinstall
688 - or it could be down because it is deprecated and kept just for its
691 The previous Boolean state was making it difficult to do capacity
692 calculations: should Ganeti reserve memory for a down instance? Now, the
693 tri-state field makes it clear:
695 - in ``up`` and ``down`` state, all resources are reserved for the
696 instance, and it can be at any time brought up if it is down
697 - in ``offline`` state, only disk space is reserved for it, but not
700 The field can have an extra use: since the transition between ``up`` and
701 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
702 transition between ``offline`` and ``down`` is done via ``gnt-instance
703 modify``, it is possible to given different rights to users. For
704 example, owners of an instance could be allowed to start/stop it, but
705 not transition it out of the offline state.
707 Instance policies and specs
708 +++++++++++++++++++++++++++
710 In previous Ganeti versions, an instance creation request was not
711 limited on the minimum size and on the maximum size just by the cluster
712 resources. As such, any policy could be implemented only in third-party
713 clients (RAPI clients, or shell wrappers over ``gnt-*``
714 tools). Furthermore, calculating cluster capacity via ``hspace`` again
715 required external input with regards to instance sizes.
717 In order to improve these workflows and to allow for example better
718 per-node group differentiation, we introduced instance specs, which
721 - minimum instance disk size, disk count, memory size, cpu count
722 - maximum values for the above metrics
723 - and “standard” values (used in ``hspace`` to calculate the standard
726 The minimum/maximum values can be also customised at node-group level,
727 for example allowing more powerful hardware to support bigger instance
730 Beside the instance specs, there are a few other settings belonging to
731 the instance policy framework. It is possible now to customise, per
732 cluster and node-group:
734 - the list of allowed disk templates
735 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
736 - the maximum ratio of instance to spindles (see below for more
737 information) for local storage
739 All these together should allow all tools that talk to Ganeti to know
740 what are the ranges of allowed values for instances and the
741 over-subscription that is allowed.
743 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
744 instance configuration, and the physical CPU configuration from the
745 node. For the spindle ratios however, we didn't track before these
746 values, so new parameters have been added:
748 - a new node parameter ``spindle_count``, defaults to 1, customisable at
749 node group or node level
750 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
752 Note that spindles in this context doesn't need to mean actual
753 mechanical hard-drives; it's just a relative number for both the node
754 I/O capacity and instance I/O consumption.
756 Instance migration behaviour
757 ++++++++++++++++++++++++++++
759 While live-migration is in general desirable over failover, it is
760 possible that for some workloads it is actually worse, due to the
761 variable time of the “suspend” phase during live migration.
763 To allow the tools to work consistently over such instances (without
764 having to hard-code instance names), a new backend parameter
765 ``always_failover`` has been added to control the migration/failover
766 behaviour. When set to True, all migration requests for an instance will
767 instead fall-back to failover.
769 Instance memory ballooning
770 ++++++++++++++++++++++++++
772 Initial support for memory ballooning has been added. The memory for an
773 instance is no longer fixed (backend parameter ``memory``), but instead
774 can vary between minimum and maximum values (backend parameters
775 ``minmem`` and ``maxmem``). Currently we only change an instance's
778 - live migrating or failing over and instance and the target node
779 doesn't have enough memory
780 - user requests changing the memory via ``gnt-instance modify
786 In order to control the use of specific CPUs by instance, support for
787 controlling CPU pinning has been added for the Xen, HVM and LXC
788 hypervisors. This is controlled by a new hypervisor parameter
789 ``cpu_mask``; details about possible values for this are in the
790 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
791 VCPU-to-CPU mapping) form will work well only when all nodes in your
792 cluster have the same amount of CPUs.
797 Another area in which Ganeti was not customisable were the parameters
798 used for storage configuration, e.g. how many stripes to use for LVM,
799 DRBD resync configuration, etc.
801 To improve this area, we've added disks parameters, which are
802 customisable at cluster and node group level, and which allow to
803 specify various parameters for disks (DRBD has the most parameters
804 currently), for example:
806 - DRBD resync algorithm and parameters (e.g. speed)
807 - the default VG for meta-data volumes for DRBD
808 - number of stripes for LVM (plain disk template)
811 These parameters can be modified via ``gnt-cluster modify -D …`` and
812 ``gnt-group modify -D …``, and are used at either instance creation (in
813 case of LVM stripes, for example) or at disk “activation” time
816 Rados block device support
817 ++++++++++++++++++++++++++
819 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
820 denoted by the ``rbd`` disk template type. This is considered
821 experimental, feedback is welcome. For details on configuring it, see
822 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
827 The existing master IP functionality works well only in simple setups (a
828 single network shared by all nodes); however, if nodes belong to
829 different networks, then the ``/32`` setup and lack of routing
830 information is not enough.
832 To allow the master IP to function well in more complex cases, the
833 system was reworked as follows:
835 - a master IP netmask setting has been added
836 - the master IP activation/turn-down code was moved from the node daemon
838 - whether to run the Ganeti-supplied master IP script or a user-supplied
839 on is a ``gnt-cluster init`` setting
841 Details about the location of the standard and custom setup scripts are
842 in the man page :manpage:`gnt-cluster(8)`; for information about the
843 setup script protocol, look at the Ganeti-supplied script.
848 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
851 It is now possible to use TLS-protected connections, and when renewing
852 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
853 it is now possible to specify spice or spice CA certificates. Also, it
854 is possible to configure a password for SPICE sessions via the
855 hypervisor parameter ``spice_password_file``.
857 There are also new parameters to control the compression and streaming
858 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
859 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
860 for the spice parameters.
862 Lastly, it is now possible to see the SPICE connection information via
863 ``gnt-instance console``.
868 A new tool (``tools/ovfconverter``) has been added that supports
869 conversion between Ganeti and the `Open Virtualization Format
870 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
873 This relies on the ``qemu-img`` tool to convert the disk formats, so the
874 actual compatibility with other virtualization solutions depends on it.
879 The configuration query daemon (``ganeti-confd``) is now optional, and
880 has been rewritten in Haskell; whether to use the daemon at all, use the
881 Python (default) or the Haskell version is selectable at configure time
882 via the ``--enable-confd`` parameter, which can take one of the
883 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
884 daemon will result in a smaller footprint; for larger systems, we
885 welcome feedback on the Haskell version which might become the default
888 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
889 daemon running. The Python version doesn't implement the new call.
892 User interface changes
893 ~~~~~~~~~~~~~~~~~~~~~~
895 We have replaced the ``--disks`` option of ``gnt-instance
896 replace-disks`` with a more flexible ``--disk`` option, which allows
897 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
898 disk size and mode can be changed upon recreation (via ``gnt-instance
899 recreate-disks``, which accepts the same ``--disk`` option).
901 As many people are used to a ``show`` command, we have added that as an
902 alias to ``info`` on all ``gnt-*`` commands.
904 The ``gnt-instance grow-disk`` command has a new mode in which it can
905 accept the target size of the disk, instead of the delta; this can be
906 more safe since two runs in absolute mode will be idempotent, and
907 sometimes it's also easier to specify the desired size directly.
909 Also the handling of instances with regard to offline secondaries has
910 been improved. Instance operations should not fail because one of it's
911 secondary nodes is offline, even though it's safe to proceed.
913 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
914 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
915 minors to instance name.
920 RAPI coverage has improved, with (for example) new resources for
921 recreate-disks, node power-cycle, etc.
926 There is partial support for ``xl`` in the Xen hypervisor; feedback is
929 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
930 whether to still support Python 2.4 or move to Python 2.6 as minimum
933 Support for Fedora has been slightly improved; the provided example
934 init.d script should work better on it and the INSTALL file should
935 document the needed dependencies.
940 The deprecated ``QueryLocks`` LUXI request has been removed. Use
941 ``Query(what=QR_LOCK, ...)`` instead.
943 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
944 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
945 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
946 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
947 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
949 RAPI client: ``CertificateError`` now derives from
950 ``GanetiApiError``. This should make it more easy to handle Ganeti
953 Deprecation warnings due to PyCrypto/paramiko import in
954 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
955 make sure to run an up-to-date paramiko version, if you use this tool.
957 The QA scripts now depend on Python 2.5 or above (the main code base
958 still works with Python 2.4).
960 The configuration file (``config.data``) is now written without
961 indentation for performance reasons; if you want to edit it, it can be
962 re-formatted via ``tools/fmtjson``.
964 A number of bugs has been fixed in the cluster merge tool.
966 ``x509`` certification verification (used in import-export) has been
967 changed to allow the same clock skew as permitted by the cluster
968 verification. This will remove some rare but hard to diagnose errors in
975 *(Released Thu, 19 Jul 2012)*
977 Very few changes from rc4 to the final release, only bugfixes:
979 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
980 instance, fix CDROM booting for KVM instances)
981 - fixed node group modification of node parameters
982 - fixed issue in LUClusterVerifyGroup with multi-group clusters
983 - fixed generation of bash completion to ensure a stable ordering
990 *(Released Fri, 13 Jul 2012)*
992 Third release candidate for 2.6. The following changes were done from
995 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
996 - Fixed an inconsistency in the LUXI protocol with the provided
997 arguments (NOT backwards compatible)
998 - Fixed a bug with node groups ipolicy where ``min`` was greater than
999 the cluster ``std`` value
1000 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
1001 easier instance debugging on nodes (requires ``hconfd`` to work)
1007 *(Released Tue, 03 Jul 2012)*
1009 Second release candidate for 2.6. The following changes were done from
1012 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
1014 - Fixed a RAPI regression on missing beparams/memory
1015 - Fixed redistribution of files on offline nodes
1016 - Added possibility to run activate-disks even though secondaries are
1017 offline. With this change it relaxes also the strictness on some other
1018 commands which use activate disks internally:
1019 * ``gnt-instance start|reboot|rename|backup|export``
1020 - Made it possible to remove safely an instance if its secondaries are
1022 - Made it possible to reinstall even though secondaries are offline
1028 *(Released Mon, 25 Jun 2012)*
1030 First release candidate for 2.6. The following changes were done from
1033 - Fixed bugs with disk parameters and ``rbd`` templates as well as
1035 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
1036 behaviour. It supports now the modify operation
1037 - ``hcheck`` implemented to analyze cluster health and possibility of
1038 improving health by rebalance
1039 - ``hbal`` has been improved in dealing with split instances
1045 *(Released Mon, 11 Jun 2012)*
1047 Second beta release of 2.6. The following changes were done from beta2
1050 - Fixed ``daemon-util`` with non-root user models
1051 - Fixed creation of plain instances with ``--no-wait-for-sync``
1052 - Fix wrong iv_names when running ``cfgupgrade``
1053 - Export more information in RAPI group queries
1054 - Fixed bug when changing instance network interfaces
1055 - Extended burnin to do NIC changes
1056 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
1057 - Changed default for DRBD barriers
1058 - Fixed DRBD error reporting for syncer rate
1059 - Verify the options on disk parameters
1061 And of course various fixes to documentation and improved unittests and
1068 *(Released Wed, 23 May 2012)*
1070 First beta release of 2.6. The following changes were done from beta1 to
1073 - integrated patch for distributions without ``start-stop-daemon``
1074 - adapted example init.d script to work on Fedora
1075 - fixed log handling in Haskell daemons
1076 - adapted checks in the watcher for pycurl linked against libnss
1077 - add partial support for ``xl`` instead of ``xm`` for Xen
1078 - fixed a type issue in cluster verification
1079 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1082 Plus integrated fixes from the 2.5 branch:
1084 - fixed ``kvm-ifup`` to use ``/bin/bash``
1085 - fixed parallel build failures
1086 - KVM live migration when using a custom keymap
1092 *(Released Tue, 24 Jul 2012)*
1094 A small bugfix release, with no new features:
1096 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1097 different default shell (e.g. Debian, Ubuntu)
1098 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1099 243 and Debian bug #650664)
1100 - fixed compatibility with KVM versions that don't support multiple boot
1101 devices (fixes Issue 230 and Debian bug #624256)
1103 Additionally, a few fixes were done to the build system (fixed parallel
1104 build failures) and to the unittests (fixed race condition in test for
1105 FileID functions, and the default enable/disable mode for QA test is now
1112 *(Released Fri, 11 May 2012)*
1114 A small bugfix release.
1116 The main issues solved are on the topic of compatibility with newer LVM
1119 - fixed parsing of ``lv_attr`` field
1120 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1121 the ``--force`` flag is needed
1123 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1124 to accept kernel 3.x too (was hardcoded to 2.6.*).
1126 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1127 of hook results and that also made display of other errors suboptimal
1128 was fixed; the code behaves now like 2.4 and earlier.
1130 Another change in 2.5, the cleanup of the OS scripts environment, is too
1131 aggressive: it removed even the ``PATH`` variable, which requires the OS
1132 scripts to *always* need to export it. Since this is a bit too strict,
1133 we now export a minimal PATH, the same that we export for hooks.
1135 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1136 integrated into this release.
1138 Finally, a few other miscellaneous changes were done (no new features,
1139 just small improvements):
1141 - Fix ``gnt-group --help`` display
1142 - Fix hardcoded Xen kernel path
1143 - Fix grow-disk handling of invalid units
1144 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1145 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1151 *(Released Thu, 12 Apr 2012)*
1153 Incompatible/important changes and bugfixes
1154 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1156 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1157 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1158 to match the underlying LUXI interface.
1159 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1160 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1161 not maintain backwards-compatibility as the underlying operation
1162 changed in an incompatible way. The RAPI client can talk to old
1163 servers, but it needs to be told so as the return value changed.
1164 - When creating file-based instances via RAPI, the ``file_driver``
1165 parameter no longer defaults to ``loop`` and must be specified.
1166 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1168 - Support for the undocumented and deprecated RAPI instance creation
1169 request format version 0 has been dropped. Use version 1, supported
1170 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1171 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1172 documentation <install>`.
1173 - The "cluster-verify" hooks are now executed per group by the
1174 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1175 if you just run ``gnt-cluster verify``, which generates one opcode per
1177 - The environment as passed to the OS scripts is cleared, and thus no
1178 environment variables defined in the node daemon's environment will be
1179 inherited by the scripts.
1180 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1182 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1183 support operations involving multiple node groups.
1184 - Offline nodes are ignored when failing over an instance.
1185 - Support for KVM version 1.0, which changed the version reporting format
1187 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1189 - ``Makefile`` is now compatible with Automake 1.11.2
1190 - Includes all bugfixes made in the 2.4 series
1195 - The ganeti-htools project has been merged into the ganeti-core source
1196 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1197 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1198 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1199 GPT-style partition tables (requires `parted
1200 <http://www.gnu.org/s/parted/>`_).
1201 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1202 - Allowed adding tags on instance creation.
1203 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1204 - Allow instances to be started in a paused state, enabling the user to
1205 see the complete console output on boot using the console.
1206 - Added new hypervisor flag to control default reboot behaviour
1207 (``reboot_behavior``).
1208 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1209 - Improved out-of-band management support:
1211 - Added ``gnt-node health`` command reporting the health status of
1213 - Added ``gnt-node power`` command to manage power status of nodes.
1214 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1216 - Instance migration can fall back to failover if instance is not
1218 - Filters can be used when listing nodes, instances, groups and locks;
1219 see :manpage:`ganeti(7)` manpage.
1220 - Added post-execution status as variables to :doc:`hooks <hooks>`
1222 - Instance tags are exported/imported together with the instance.
1223 - When given an explicit job ID, ``gnt-job info`` will work for archived
1225 - Jobs can define dependencies on other jobs (not yet supported via
1226 RAPI or command line, but used by internal commands and usable via
1229 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1232 - Instance failover is now available as a RAPI resource
1233 (``/2/instances/[instance_name]/failover``).
1234 - ``gnt-instance info`` defaults to static information if primary node
1236 - Opcodes have a new ``comment`` attribute.
1237 - Added basic SPICE support to KVM hypervisor.
1238 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1240 Node group improvements
1241 ~~~~~~~~~~~~~~~~~~~~~~~
1243 - ``gnt-cluster verify`` has been modified to check groups separately,
1244 thereby improving performance.
1245 - Node group support has been added to ``gnt-cluster verify-disks``,
1246 which now operates per node group.
1247 - Watcher has been changed to work better with node groups.
1249 - One process and state file per node group.
1250 - Slow watcher in one group doesn't block other group's watcher.
1252 - Added new command, ``gnt-group evacuate``, to move all instances in a
1253 node group to other groups.
1254 - Added ``gnt-instance change-group`` to move an instance to another
1256 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1257 per-group operations.
1258 - Node groups can be tagged.
1259 - Some operations switch from an exclusive to a shared lock as soon as
1261 - Instance's primary and secondary nodes' groups are now available as
1262 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1263 and ``snodes.group.uuid``).
1268 - Numerous updates to documentation and manpages.
1270 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1272 - Some opcode/job results are now also documented, see :doc:`RAPI
1275 - A lockset's internal lock is now also visible in lock monitor.
1276 - Log messages from job queue workers now contain information about the
1277 opcode they're processing.
1278 - ``gnt-instance console`` no longer requires the instance lock.
1279 - A short delay when waiting for job changes reduces the number of LUXI
1280 requests significantly.
1281 - DRBD metadata volumes are overwritten with zeros during disk creation.
1282 - Out-of-band commands no longer acquire the cluster lock in exclusive
1284 - ``devel/upload`` now uses correct permissions for directories.
1290 *(Released Fri, 23 Mar 2012)*
1292 This was the sixth release candidate of the 2.5 series.
1298 *(Released Mon, 9 Jan 2012)*
1300 This was the fifth release candidate of the 2.5 series.
1306 *(Released Thu, 27 Oct 2011)*
1308 This was the fourth release candidate of the 2.5 series.
1314 *(Released Wed, 26 Oct 2011)*
1316 This was the third release candidate of the 2.5 series.
1322 *(Released Tue, 18 Oct 2011)*
1324 This was the second release candidate of the 2.5 series.
1330 *(Released Tue, 4 Oct 2011)*
1332 This was the first release candidate of the 2.5 series.
1338 *(Released Wed, 31 Aug 2011)*
1340 This was the third beta release of the 2.5 series.
1346 *(Released Mon, 22 Aug 2011)*
1348 This was the second beta release of the 2.5 series.
1354 *(Released Fri, 12 Aug 2011)*
1356 This was the first beta release of the 2.5 series.
1362 *(Released Thu, 27 Oct 2011)*
1364 - Fixed bug when parsing command line parameter values ending in
1366 - Fixed assertion error after unclean master shutdown
1367 - Disable HTTP client pool for RPC, significantly reducing memory usage
1369 - Fixed queue archive creation with wrong permissions
1375 *(Released Tue, 23 Aug 2011)*
1379 - Fixed documentation for importing with ``--src-dir`` option
1380 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1381 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1387 *(Released Fri, 5 Aug 2011)*
1389 Many bug-fixes and a few small features:
1391 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1392 issues when you tried to add an instance with two MAC addresses in one
1394 - KVM: fixed per-instance stored UID value
1395 - KVM: configure bridged NICs at migration start
1396 - KVM: Fix a bug where instance will not start with never KVM versions
1398 - Added OS search path to ``gnt-cluster info``
1399 - Fixed an issue with ``file_storage_dir`` where you were forced to
1400 provide an absolute path, but the documentation states it is a
1401 relative path, the documentation was right
1402 - Added a new parameter to instance stop/start called ``--no-remember``
1403 that will make the state change to not be remembered
1404 - Implemented ``no_remember`` at RAPI level
1405 - Improved the documentation
1406 - Node evacuation: don't call IAllocator if node is already empty
1407 - Fixed bug in DRBD8 replace disks on current nodes
1408 - Fixed bug in recreate-disks for DRBD instances
1409 - Moved assertion checking locks in ``gnt-instance replace-disks``
1410 causing it to abort with not owning the right locks for some situation
1411 - Job queue: Fixed potential race condition when cancelling queued jobs
1412 - Fixed off-by-one bug in job serial generation
1413 - ``gnt-node volumes``: Fix instance names
1414 - Fixed aliases in bash completion
1415 - Fixed a bug in reopening log files after being sent a SIGHUP
1416 - Added a flag to burnin to allow specifying VCPU count
1417 - Bugfixes to non-root Ganeti configuration
1423 *(Released Thu, 12 May 2011)*
1425 Many bug-fixes and a few new small features:
1427 - Fixed a bug related to log opening failures
1428 - Fixed a bug in instance listing with orphan instances
1429 - Fixed a bug which prevented resetting the cluster-level node parameter
1430 ``oob_program`` to the default
1431 - Many fixes related to the ``cluster-merge`` tool
1432 - Fixed a race condition in the lock monitor, which caused failures
1433 during (at least) creation of many instances in parallel
1434 - Improved output for gnt-job info
1435 - Removed the quiet flag on some ssh calls which prevented debugging
1437 - Improved the N+1 failure messages in cluster verify by actually
1438 showing the memory values (needed and available)
1439 - Increased lock attempt timeouts so that when executing long operations
1440 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1441 too early and thus prevent the use of the 'fair' mechanism
1442 - Changed instance query data (``gnt-instance info``) to not acquire
1443 locks unless needed, thus allowing its use on locked instance if only
1444 static information is asked for
1445 - Improved behaviour with filesystems that do not support rename on an
1447 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1448 kept locks on all nodes during the wipe, which is unneeded
1449 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1450 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1451 10GiB) which caused the wipe to fail right at the end in some cases
1452 - Fixed master IP activation when doing master failover with no-voting
1453 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1454 the master node itself
1455 - Fixed potential data-loss in under disk full conditions, where Ganeti
1456 wouldn't check correctly the return code and would consider
1457 partially-written files 'correct'
1458 - Fixed bug related to multiple VGs and DRBD disk replacing
1459 - Added new disk parameter ``metavg`` that allows placement of the meta
1460 device for DRBD in a different volume group
1461 - Fixed error handling in the node daemon when the system libc doesn't
1462 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1463 - Fixed lock release during replace-disks, which kept cluster-wide locks
1464 when doing disk replaces with an iallocator script
1465 - Added check for missing bridges in cluster verify
1466 - Handle EPIPE errors while writing to the terminal better, so that
1467 piping the output to e.g. ``less`` doesn't cause a backtrace
1468 - Fixed rare case where a ^C during Luxi calls could have been
1469 interpreted as server errors, instead of simply terminating
1470 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1472 - Added a few more parameters to the KVM hypervisor, allowing a second
1473 CDROM, custom disk type for CDROMs and a floppy image
1474 - Removed redundant message in instance rename when the name is given
1476 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1477 disks on new nodes, allowing recreation when the original instance
1478 nodes are completely gone
1479 - Added option when converting disk templates to DRBD to skip waiting
1480 for the resync, in order to make the instance available sooner
1481 - Added two new variables to the OS scripts environment (containing the
1483 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1484 to allow use of ``pvgrub`` as bootloader
1485 - Changed the instance memory modifications to only check out-of-memory
1486 conditions on memory increases, and turned the secondary node warnings
1487 into errors (they can still be overridden via ``--force``)
1488 - Fixed the handling of a corner case when the Python installation gets
1489 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1490 to execute a command that doesn't exist
1491 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1492 node of the instance returned failures during instance shutdown; this
1493 adds the option ``--ignore-consistency`` to gnt-instance move
1495 And as usual, various improvements to the error messages, documentation
1502 *(Released Wed, 09 Mar 2011)*
1504 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1505 the RAPI users file if run twice (even with ``--dry-run``).
1507 The release fixes that bug (nothing else changed).
1513 *(Released Mon, 07 Mar 2011)*
1515 Final 2.4.0 release. Just a few small fixes:
1517 - Fixed RAPI node evacuate
1518 - Fixed the kvm-ifup script
1519 - Fixed internal error handling for special job cases
1520 - Updated man page to specify the escaping feature for options
1526 *(Released Mon, 28 Feb 2011)*
1528 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1529 that this feature wiped the disks of the wrong instance, leading to loss
1534 - Fixed title of query field containing instance name
1535 - Expanded the glossary in the documentation
1536 - Fixed one unittest (internal issue)
1542 *(Released Mon, 21 Feb 2011)*
1544 A number of bug fixes plus just a couple functionality changes.
1546 On the user-visible side, the ``gnt-* list`` command output has changed
1547 with respect to "special" field states. The current rc1 style of display
1548 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1549 the default output mode special fields states are displayed as follows:
1551 - Offline resource: ``*``
1552 - Unavailable/not applicable: ``-``
1553 - Data missing (RPC failure): ``?``
1554 - Unknown field: ``??``
1556 Another user-visible change is the addition of ``--force-join`` to
1561 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1562 - Fixed regression in RAPI/instance reinstall where all parameters were
1563 required (instead of optional)
1564 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1565 - Fixed iallocator usage (offline nodes were not considered offline)
1566 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1567 - Fixed hypervisor and OS parameter validation with respect to
1568 non-vm_capable nodes
1569 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1571 - Fixed ``tools/listrunner`` with respect to agent-based usage
1577 *(Released Fri, 4 Feb 2011)*
1579 Many changes and fixes since the beta1 release. While there were some
1580 internal changes, the code has been mostly stabilised for the RC
1583 Note: the dumb allocator was removed in this release, as it was not kept
1584 up-to-date with the IAllocator protocol changes. It is recommended to
1585 use the ``hail`` command from the ganeti-htools package.
1587 Note: the 2.4 and up versions of Ganeti are not compatible with the
1588 0.2.x branch of ganeti-htools. You need to upgrade to
1589 ganeti-htools-0.3.0 (or later).
1591 Regressions fixed from 2.3
1592 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1594 - Fixed the ``gnt-cluster verify-disks`` command
1595 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1597 - Fixed disk adoption breakage
1598 - Fixed wrong headers in instance listing for field aliases
1603 - Fixed corner case in KVM handling of NICs
1604 - Fixed many cases of wrong handling of non-vm_capable nodes
1605 - Fixed a bug where a missing instance symlink was not possible to
1606 recreate with any ``gnt-*`` command (now ``gnt-instance
1607 activate-disks`` does it)
1608 - Fixed the volume group name as reported by ``gnt-cluster
1610 - Increased timeouts for the import-export code, hopefully leading to
1611 fewer aborts due network or instance timeouts
1612 - Fixed bug in ``gnt-node list-storage``
1613 - Fixed bug where not all daemons were started on cluster
1614 initialisation, but only at the first watcher run
1615 - Fixed many bugs in the OOB implementation
1616 - Fixed watcher behaviour in presence of instances with offline
1618 - Fixed instance list output for instances running on the wrong node
1619 - a few fixes to the cluster-merge tool, but it still cannot merge
1620 multi-node groups (currently it is not recommended to use this tool)
1626 - Improved network configuration for the KVM hypervisor
1627 - Added e1000 as a supported NIC for Xen-HVM
1628 - Improved the lvmstrap tool to also be able to use partitions, as
1629 opposed to full disks
1630 - Improved speed of disk wiping (the cluster parameter
1631 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1632 of instance creations
1633 - Added documentation for the OS parameters
1634 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1635 hypervisor is not responding
1636 - Added display of blacklisted and hidden OS information in
1637 ``gnt-cluster info``
1638 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1639 NIC and node parameters, which might create problems with currently
1640 invalid (but undetected) configuration files, but prevents validation
1641 failures when unrelated parameters are modified
1642 - Changed cluster initialisation to wait for the master daemon to become
1644 - Expanded the RAPI interface:
1646 - Added config redistribution resource
1647 - Added activation/deactivation of instance disks
1648 - Added export of console information
1650 - Implemented log file reopening on SIGHUP, which allows using
1651 logrotate(8) for the Ganeti log files
1652 - Added a basic OOB helper script as an example
1658 *(Released Fri, 14 Jan 2011)*
1663 - Fixed timezone issues when formatting timestamps
1664 - Added support for node groups, available via ``gnt-group`` and other
1666 - Added out-of-band framework and management, see :doc:`design
1667 document <design-oob>`
1668 - Removed support for roman numbers from ``gnt-node list`` and
1669 ``gnt-instance list``.
1670 - Allowed modification of master network interface via ``gnt-cluster
1671 modify --master-netdev``
1672 - Accept offline secondaries while shutting down instance disks
1673 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1674 - Added support for multiple LVM volume groups
1675 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1677 - Added commands to list available fields:
1679 - ``gnt-node list-fields``
1680 - ``gnt-group list-fields``
1681 - ``gnt-instance list-fields``
1683 - Updated documentation and man pages
1688 - Moved ``rapi_users`` file into separate directory, now named
1689 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1691 - Added new tool for running commands on many machines,
1692 ``tools/ganeti-listrunner``
1693 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1694 improving the ``gnt-instance console`` output
1695 - Allowed customisation of disk index separator at ``configure`` time
1696 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1697 - Added support for non-partitioned md disks in ``lvmstrap``
1698 - Added script to gracefully power off KVM instances
1699 - Split ``utils`` module into smaller parts
1700 - Changed query operations to return more detailed information, e.g.
1701 whether an information is unavailable due to an offline node. To use
1702 this new functionality, the LUXI call ``Query`` must be used. Field
1703 information is now stored by the master daemon and can be retrieved
1704 using ``QueryFields``. Instances, nodes and groups can also be queried
1705 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1706 exposed via RAPI). The following commands make use of this
1707 infrastructure change:
1709 - ``gnt-group list``
1710 - ``gnt-group list-fields``
1712 - ``gnt-node list-fields``
1713 - ``gnt-instance list``
1714 - ``gnt-instance list-fields``
1715 - ``gnt-debug locks``
1720 - New RAPI resources (see :doc:`rapi`):
1724 - ``/2/groups/[group_name]``
1725 - ``/2/groups/[group_name]/assign-nodes``
1726 - ``/2/groups/[group_name]/modify``
1727 - ``/2/groups/[group_name]/rename``
1728 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1732 - Implemented ``no_install`` for instance creation
1733 - Implemented OS parameters for instance reinstallation, allowing
1734 use of special settings on reinstallation (e.g. for preserving data)
1739 - Added IPv6 support in import/export
1740 - Pause DRBD synchronization while wiping disks on instance creation
1741 - Updated unittests and QA scripts
1742 - Improved network parameters passed to KVM
1743 - Converted man pages from docbook to reStructuredText
1749 *(Released Mon, 20 Dec 2010)*
1751 Released version 2.3.1~rc1 without any changes.
1757 *(Released Wed, 1 Dec 2010)*
1759 - impexpd: Disable OpenSSL compression in socat if possible (backport
1760 from master, commit e90739d625b, see :doc:`installation guide
1761 <install-quick>` for details)
1762 - Changed unittest coverage report to exclude test scripts
1763 - Added script to check version format
1769 *(Released Wed, 1 Dec 2010)*
1771 Released version 2.3.0~rc1 without any changes.
1777 *(Released Fri, 19 Nov 2010)*
1779 A number of bugfixes and documentation updates:
1781 - Update ganeti-os-interface documentation
1782 - Fixed a bug related to duplicate MACs or similar items which should be
1784 - Fix breakage in OS state modify
1785 - Reinstall instance: disallow offline secondaries (fixes bug related to
1786 OS changing but reinstall failing)
1787 - plus all the other fixes between 2.2.1 and 2.2.2
1793 *(Released Tue, 2 Nov 2010)*
1795 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1796 - Fixed master failover race with watcher
1797 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1799 - Accept previously stopped instance for export with instance removal
1800 - Simplify and extend the environment variables for instance OS scripts
1801 - Added new node flags, ``master_capable`` and ``vm_capable``
1802 - Added optional instance disk wiping prior during allocation. This is a
1803 cluster-wide option and can be set/modified using
1804 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1805 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1806 :doc:`install-quick`
1807 - Added a new watcher option (``--ignore-pause``)
1808 - Added option to ignore offline node on instance start/stop
1809 (``--ignore-offline``)
1810 - Allow overriding OS parameters with ``gnt-instance reinstall``
1811 - Added ability to change node's secondary IP address using ``gnt-node
1813 - Implemented privilege separation for all daemons except
1814 ``ganeti-noded``, see ``configure`` options
1815 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1817 - Implemented job priorities (see ``ganeti(7)`` manpage)
1818 - Ignore failures while shutting down instances during failover from
1820 - Exit daemon's bootstrap process only once daemon is ready
1821 - Export more information via ``LUInstanceQuery``/remote API
1822 - Improved documentation, QA and unittests
1823 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1824 restart if the file was created or changed
1825 - Added LUXI protocol version sent with each request and response,
1826 allowing detection of server/client mismatches
1827 - Moved the Python scripts among gnt-* and ganeti-* into modules
1828 - Moved all code related to setting up SSH to an external script,
1830 - Infrastructure changes for node group support in future versions
1836 *(Released Fri, 19 Nov 2010)*
1838 A few small bugs fixed, and some improvements to the build system:
1840 - Fix documentation regarding conversion to drbd
1841 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1843 - Fix error handling in node modify with multiple changes
1844 - Allow remote imports without checked names
1850 *(Released Tue, 19 Oct 2010)*
1852 - Disable SSL session ID cache in RPC client
1858 *(Released Thu, 14 Oct 2010)*
1860 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1861 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1863 - Fix problems with interaction between Curl and Python's HTTP server,
1864 resulting in increased speed in many RPC calls
1865 - Improve our release script to prevent breakage with older aclocal and
1872 *(Released Thu, 7 Oct 2010)*
1874 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1875 value retrieved from master
1876 - Added support for blacklisted or hidden OS definitions
1877 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1878 - Added support for -mem-path in KVM hypervisor abstraction layer
1879 - Allow overriding instance parameters in tool for inter-cluster
1880 instance moves (``tools/move-instance``)
1881 - Improved opcode summaries (e.g. in ``gnt-job list``)
1882 - Improve consistency of OS listing by sorting it
1883 - Documentation updates
1889 *(Released Fri, 8 Oct 2010)*
1891 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1897 *(Released Mon, 4 Oct 2010)*
1899 - Fixed regression in ``gnt-instance rename``
1905 *(Released Wed, 22 Sep 2010)*
1907 - Fixed OS_VARIANT variable for OS scripts
1908 - Fixed cluster tag operations via RAPI
1909 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1910 - Disabled RAPI CA checks in watcher
1916 *(Released Mon, 23 Aug 2010)*
1918 - Support DRBD versions of the format "a.b.c.d"
1920 - Re-introduce support for usage from multiple threads in RAPI client
1921 - Instance renames and modify via RAPI
1922 - Work around race condition between processing and archival in job
1924 - Mark opcodes following failed one as failed, too
1925 - Job field ``lock_status`` was removed due to difficulties making it
1926 work with the changed job queue in Ganeti 2.2; a better way to monitor
1927 locks is expected for a later 2.2.x release
1928 - Fixed dry-run behaviour with many commands
1929 - Support ``ssh-agent`` again when adding nodes
1930 - Many additional bugfixes
1936 *(Released Fri, 30 Jul 2010)*
1938 Important change: the internal RPC mechanism between Ganeti nodes has
1939 changed from using a home-grown http library (based on the Python base
1940 libraries) to use the PycURL library. This requires that PycURL is
1941 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1942 against GnuTLS by default. cURL's support for GnuTLS had known issues
1943 before cURL 7.21.0 and we recommend using the latest cURL release or
1944 linking against OpenSSL. Most other distributions already link PycURL
1945 and cURL against OpenSSL. The command::
1947 python -c 'import pycurl; print pycurl.version'
1949 can be used to determine the libraries PycURL and cURL are linked
1952 Other significant changes:
1954 - Rewrote much of the internals of the job queue, in order to achieve
1955 better parallelism; this decouples job query operations from the job
1956 processing, and it should allow much nicer behaviour of the master
1957 daemon under load, and it also has uncovered some long-standing bugs
1958 related to the job serialisation (now fixed)
1959 - Added a default iallocator setting to the cluster parameters,
1960 eliminating the need to always pass nodes or an iallocator for
1961 operations that require selection of new node(s)
1962 - Added experimental support for the LXC virtualization method
1963 - Added support for OS parameters, which allows the installation of
1964 instances to pass parameter to OS scripts in order to customise the
1966 - Added a hypervisor parameter controlling the migration type (live or
1967 non-live), since hypervisors have various levels of reliability; this
1968 has renamed the 'live' parameter to 'mode'
1969 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1970 logical volumes, meaning that cluster verify will ignore them and not
1971 flag their presence as errors
1972 - The watcher will now reset the error count for failed instances after
1973 8 hours, thus allowing self-healing if the problem that caused the
1974 instances to be down/fail to start has cleared in the meantime
1975 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1976 check for, and warn, if the drbd module parameter ``usermode_helper``
1977 is not consistent with the cluster-wide setting; this is needed to
1978 make diagnose easier of failed drbd creations
1979 - Started adding base IPv6 support, but this is not yet
1980 enabled/available for use
1981 - Rename operations (cluster, instance) will now return the new name,
1982 which is especially useful if a short name was passed in
1983 - Added support for instance migration in RAPI
1984 - Added a tool to pre-configure nodes for the SSH setup, before joining
1985 them to the cluster; this will allow in the future a simplified model
1986 for node joining (but not yet fully enabled in 2.2); this needs the
1987 paramiko python library
1988 - Fixed handling of name-resolving errors
1989 - Fixed consistency of job results on the error path
1990 - Fixed master-failover race condition when executed multiple times in
1992 - Fixed many bugs related to the job queue (mostly introduced during the
1993 2.2 development cycle, so not all are impacting 2.1)
1994 - Fixed instance migration with missing disk symlinks
1995 - Fixed handling of unknown jobs in ``gnt-job archive``
1996 - And many other small fixes/improvements
2000 - Enhanced both the unittest and the QA coverage
2001 - Switched the opcode validation to a generic model, and extended the
2002 validation to all opcode parameters
2003 - Changed more parts of the code that write shell scripts to use the
2005 - Switched the master daemon to use the asyncore library for the Luxi
2012 *(Released Thu, 17 Jun 2010)*
2014 - Added tool (``move-instance``) and infrastructure to move instances
2015 between separate clusters (see :doc:`separate documentation
2016 <move-instance>` and :doc:`design document <design-2.2>`)
2017 - Added per-request RPC timeout
2018 - RAPI now requires a Content-Type header for requests with a body (e.g.
2019 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
2020 :rfc:`2616` (HTTP/1.1), section 7.2.1)
2021 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
2023 - Implemented initial support for running Ganeti daemons as separate
2024 users, see configure-time flags ``--with-user-prefix`` and
2025 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
2027 - Instances can be removed after export (``gnt-backup export
2028 --remove-instance``)
2029 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
2030 key (instead of 1024 bit)
2031 - Added new cluster configuration file for cluster domain secret
2032 - Import/export now use SSL instead of SSH
2033 - Added support for showing estimated time when exporting an instance,
2034 see the ``ganeti-os-interface(7)`` manpage and look for
2041 *(Released Tue, 16 Nov 2010)*
2043 Some more bugfixes. Unless critical bugs occur, this will be the last
2046 - Fix case of MAC special-values
2047 - Fix mac checker regex
2048 - backend: Fix typo causing "out of range" error
2049 - Add missing --units in gnt-instance list man page
2055 *(Released Tue, 24 Aug 2010)*
2058 - Don't ignore secondary node silently on non-mirrored disk templates
2060 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
2061 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
2062 - Properly document the usb_mouse parameter
2063 - Fix path in ganeti-rapi(8) (issue 116)
2064 - Adjust error message when the ganeti user's .ssh directory is
2066 - Add same-node-check when changing the disk template to drbd
2072 *(Released Fri, 16 Jul 2010)*
2075 - Add an option to only select some reboot types during qa/burnin.
2076 (on some hypervisors consequent reboots are not supported)
2077 - Fix infrequent race condition in master failover. Sometimes the old
2078 master ip address would be still detected as up for a short time
2079 after it was removed, causing failover to fail.
2080 - Decrease mlockall warnings when the ctypes module is missing. On
2081 Python 2.4 we support running even if no ctypes module is installed,
2082 but we were too verbose about this issue.
2083 - Fix building on old distributions, on which man doesn't have a
2085 - Fix RAPI not to ignore the MAC address on instance creation
2086 - Implement the old instance creation format in the RAPI client.
2092 *(Released Thu, 01 Jul 2010)*
2094 A small bugfix release:
2095 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2096 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2097 a non-existing option
2098 - Fix instance create: the --force-variant option was ignored
2099 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2100 - Fix modify node storage with non-FQDN arguments
2101 - Fix RAPI client to authenticate under Python 2.6 when used
2102 for more than 5 requests needing authentication
2103 - Fix gnt-instance modify -t (storage) giving a wrong error message
2104 when converting a non-shutdown drbd instance to plain
2110 *(Released Fri, 18 Jun 2010)*
2112 A small bugfix release:
2114 - Fix live migration of KVM instances started with older Ganeti
2115 versions which had fewer hypervisor parameters
2116 - Fix gnt-instance grow-disk on down instances
2117 - Fix an error-reporting bug during instance migration
2118 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2119 silently ignoring broken ones
2120 - Fix an RPC error reporting bug affecting, for example, RAPI client
2122 - Fix bug triggered by different API version os-es on different nodes
2123 - Fix a bug in instance startup with custom hvparams: OS level
2124 parameters would fail to be applied.
2125 - Fix the RAPI client under Python 2.6 (but more work is needed to
2126 make it work completely well with OpenSSL)
2127 - Fix handling of errors when resolving names from DNS
2133 *(Released Thu, 3 Jun 2010)*
2135 A medium sized development cycle. Some new features, and some
2136 fixes/small improvements/cleanups.
2138 Significant features
2139 ~~~~~~~~~~~~~~~~~~~~
2141 The node deamon now tries to mlock itself into memory, unless the
2142 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2143 its logs, and falls back to console logging. This allows emergency
2144 features such as ``gnt-node powercycle`` to work even in the event of a
2145 broken node disk (tested offlining the disk hosting the node's
2146 filesystem and dropping its memory caches; don't try this at home)
2148 KVM: add vhost-net acceleration support. It can be tested with a new
2149 enough version of the kernel and of qemu-kvm.
2151 KVM: Add instance chrooting feature. If you use privilege dropping for
2152 your VMs you can also now force them to chroot to an empty directory,
2153 before starting the emulated guest.
2155 KVM: Add maximum migration bandwith and maximum downtime tweaking
2156 support (requires a new-enough version of qemu-kvm).
2158 Cluster verify will now warn if the master node doesn't have the master
2159 ip configured on it.
2161 Add a new (incompatible) instance creation request format to RAPI which
2162 supports all parameters (previously only a subset was supported, and it
2163 wasn't possible to extend the old format to accomodate all the new
2164 features. The old format is still supported, and a client can check for
2165 this feature, before using it, by checking for its presence in the
2166 ``features`` RAPI resource.
2168 Now with ancient latin support. Try it passing the ``--roman`` option to
2169 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2170 (requires the python-roman module to be installed, in order to work).
2175 As usual many internal code refactorings, documentation updates, and
2178 - Lots of improvements and cleanups to the experimental Remote API
2179 (RAPI) client library.
2180 - A new unit test suite for the core daemon libraries.
2181 - A fix to creating missing directories makes sure the umask is not
2182 applied anymore. This enforces the same directory permissions
2184 - Better handling terminating daemons with ctrl+c (used when running
2185 them in debugging mode).
2186 - Fix a race condition in live migrating a KVM instance, when stat()
2187 on the old proc status file returned EINVAL, which is an unexpected
2189 - Fixed manpage checking with newer man and utf-8 charachters. But now
2190 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2196 *(Released Fri, 7 May 2010)*
2198 Fix a bug which prevented untagged KVM instances from starting.
2204 *(Released Fri, 7 May 2010)*
2206 Another release with a long development cycle, during which many
2207 different features were added.
2209 Significant features
2210 ~~~~~~~~~~~~~~~~~~~~
2212 The KVM hypervisor now can run the individual instances as non-root, to
2213 reduce the impact of a VM being hijacked due to bugs in the
2214 hypervisor. It is possible to run all instances as a single (non-root)
2215 user, to manually specify a user for each instance, or to dynamically
2216 allocate a user out of a cluster-wide pool to each instance, with the
2217 guarantee that no two instances will run under the same user ID on any
2220 An experimental RAPI client library, that can be used standalone
2221 (without the other Ganeti libraries), is provided in the source tree as
2222 ``lib/rapi/client.py``. Note this client might change its interface in
2223 the future, as we iterate on its capabilities.
2225 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2226 replace the cluster's certificates and crypto keys. This might help in
2227 case they have been compromised, or have simply expired.
2229 A new disk option for instance creation has been added that allows one
2230 to "adopt" currently existing logical volumes, with data
2231 preservation. This should allow easier migration to Ganeti from
2232 unmanaged (or managed via other software) instances.
2234 Another disk improvement is the possibility to convert between redundant
2235 (DRBD) and plain (LVM) disk configuration for an instance. This should
2236 allow better scalability (starting with one node and growing the
2237 cluster, or shrinking a two-node cluster to one node).
2239 A new feature that could help with automated node failovers has been
2240 implemented: if a node sees itself as offline (by querying the master
2241 candidates), it will try to shutdown (hard) all instances and any active
2242 DRBD devices. This reduces the risk of duplicate instances if an
2243 external script automatically failovers the instances on such nodes. To
2244 enable this, the cluster parameter ``maintain_node_health`` should be
2245 enabled; in the future this option (per the name) will enable other
2246 automatic maintenance features.
2248 Instance export/import now will reuse the original instance
2249 specifications for all parameters; that means exporting an instance,
2250 deleting it and the importing it back should give an almost identical
2251 instance. Note that the default import behaviour has changed from
2252 before, where it created only one NIC; now it recreates the original
2255 Cluster verify has added a few new checks: SSL certificates validity,
2256 /etc/hosts consistency across the cluster, etc.
2261 As usual, many internal changes were done, documentation fixes,
2264 - Fixed cluster initialization with disabled cluster storage (regression
2265 introduced in 2.1.1)
2266 - File-based storage supports growing the disks
2267 - Fixed behaviour of node role changes
2268 - Fixed cluster verify for some corner cases, plus a general rewrite of
2269 cluster verify to allow future extension with more checks
2270 - Fixed log spamming by watcher and node daemon (regression introduced
2272 - Fixed possible validation issues when changing the list of enabled
2274 - Fixed cleanup of /etc/hosts during node removal
2275 - Fixed RAPI response for invalid methods
2276 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2277 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2278 from ide disks, etc.)
2279 - Allow OS changes without re-installation (to record a changed OS
2280 outside of Ganeti, or to allow OS renames)
2281 - Allow instance creation without OS installation (useful for example if
2282 the OS will be installed manually, or restored from a backup not in
2284 - Implemented option to make cluster ``copyfile`` use the replication
2286 - Added list of enabled hypervisors to ssconf (possibly useful for
2288 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2290 - A partial form of node re-IP is possible via node readd, which now
2291 allows changed node primary IP
2292 - Command line utilities now show an informational message if the job is
2294 - The logs of the master daemon now show the PID/UID/GID of the
2301 *(Released Fri, 12 Mar 2010)*
2303 During the 2.1.0 long release candidate cycle, a lot of improvements and
2304 changes have accumulated with were released later as 2.1.1.
2309 The node evacuate command (``gnt-node evacuate``) was significantly
2310 rewritten, and as such the IAllocator protocol was changed - a new
2311 request type has been added. This unfortunate change during a stable
2312 series is designed to improve performance of node evacuations; on
2313 clusters with more than about five nodes and which are well-balanced,
2314 evacuation should proceed in parallel for all instances of the node
2315 being evacuated. As such, any existing IAllocator scripts need to be
2316 updated, otherwise the above command will fail due to the unknown
2317 request. The provided "dumb" allocator has not been updated; but the
2318 ganeti-htools package supports the new protocol since version 0.2.4.
2320 Another important change is increased validation of node and instance
2321 names. This might create problems in special cases, if invalid host
2322 names are being used.
2324 Also, a new layer of hypervisor parameters has been added, that sits at
2325 OS level between the cluster defaults and the instance ones. This allows
2326 customisation of virtualization parameters depending on the installed
2327 OS. For example instances with OS 'X' may have a different KVM kernel
2328 (or any other parameter) than the cluster defaults. This is intended to
2329 help managing a multiple OSes on the same cluster, without manual
2330 modification of each instance's parameters.
2332 A tool for merging clusters, ``cluster-merge``, has been added in the
2333 tools sub-directory.
2338 - Improved the int/float conversions that should make the code more
2339 robust in face of errors from the node daemons
2340 - Fixed the remove node code in case of internal configuration errors
2341 - Fixed the node daemon behaviour in face of inconsistent queue
2342 directory (e.g. read-only file-system where we can't open the files
2344 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2345 now it either aborts cleanly or, if given the new "auto_promote"
2346 parameter, will automatically promote other nodes as needed
2347 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2348 completely prevent Ganeti from working
2349 - Fixed bug for instance export when not all disks were successfully
2351 - Fixed behaviour of node add when the new node is slow in starting up
2353 - Fixed handling of signals in the LUXI client, which should improve
2354 behaviour of command-line scripts
2355 - Added checks for invalid node/instance names in the configuration (now
2356 flagged during cluster verify)
2357 - Fixed watcher behaviour for disk activation errors
2358 - Fixed two potentially endless loops in http library, which led to the
2359 RAPI daemon hanging and consuming 100% CPU in some cases
2360 - Fixed bug in RAPI daemon related to hashed passwords
2361 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2363 - Enhanced compatibility with non-Debian OSes, but not using absolute
2364 path in some commands and allowing customisation of the ssh
2365 configuration directory
2366 - Fixed possible future issue with new Python versions by abiding to the
2367 proper use of ``__slots__`` attribute on classes
2368 - Added checks that should prevent directory traversal attacks
2369 - Many documentation fixes based on feedback from users
2374 - Added an "early_release" more for instance replace disks and node
2375 evacuate, where we release locks earlier and thus allow higher
2376 parallelism within the cluster
2377 - Added watcher hooks, intended to allow the watcher to restart other
2378 daemons (e.g. from the ganeti-nbma project), but they can be used of
2379 course for any other purpose
2380 - Added a compile-time disable for DRBD barriers, to increase
2381 performance if the administrator trusts the power supply or the
2382 storage system to not lose writes
2383 - Added the option of using syslog for logging instead of, or in
2384 addition to, Ganeti's own log files
2385 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2386 can indeed boot from a paravirtual NIC
2387 - Added a generic debug level for many operations; while this is not
2388 used widely yet, it allows one to pass the debug value all the way to
2390 - Enhanced the hooks environment for instance moves (failovers,
2391 migrations) where the primary/secondary nodes changed during the
2392 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2393 - Enhanced data validations for many user-supplied values; one important
2394 item is the restrictions imposed on instance and node names, which
2395 might reject some (invalid) host names
2396 - Add a configure-time option to disable file-based storage, if it's not
2397 needed; this allows greater security separation between the master
2398 node and the other nodes from the point of view of the inter-node RPC
2400 - Added user notification in interactive tools if job is waiting in the
2401 job queue or trying to acquire locks
2402 - Added log messages when a job is waiting for locks
2403 - Added filtering by node tags in instance operations which admit
2404 multiple instances (start, stop, reboot, reinstall)
2405 - Added a new tool for cluster mergers, ``cluster-merge``
2406 - Parameters from command line which are of the form ``a=b,c=d`` can now
2407 use backslash escapes to pass in values which contain commas,
2408 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2410 - For KVM, the instance name is the first parameter passed to KVM, so
2411 that it's more visible in the process list
2417 *(Released Tue, 2 Mar 2010)*
2419 Ganeti 2.1 brings many improvements with it. Major changes:
2421 - Added infrastructure to ease automated disk repairs
2422 - Added new daemon to export configuration data in a cheaper way than
2423 using the remote API
2424 - Instance NICs can now be routed instead of being associated with a
2426 - Improved job locking logic to reduce impact of jobs acquiring multiple
2427 locks waiting for other long-running jobs
2429 In-depth implementation details can be found in the Ganeti 2.1 design
2435 - Added chroot hypervisor
2436 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2438 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2439 ``bootloader_path`` and ``bootloader_args``)
2440 - Added the ``use_localtime`` option for the xen-hvm and kvm
2441 hypervisors, and the default value for this has changed to false (in
2442 2.0 xen-hvm always enabled it)
2443 - Added luxi call to submit multiple jobs in one go
2444 - Added cluster initialization option to not modify ``/etc/hosts``
2446 - Added network interface parameters
2447 - Added dry run mode to some LUs
2448 - Added RAPI resources:
2450 - ``/2/instances/[instance_name]/info``
2451 - ``/2/instances/[instance_name]/replace-disks``
2452 - ``/2/nodes/[node_name]/evacuate``
2453 - ``/2/nodes/[node_name]/migrate``
2454 - ``/2/nodes/[node_name]/role``
2455 - ``/2/nodes/[node_name]/storage``
2456 - ``/2/nodes/[node_name]/storage/modify``
2457 - ``/2/nodes/[node_name]/storage/repair``
2459 - Added OpCodes to evacuate or migrate all instances on a node
2460 - Added new command to list storage elements on nodes (``gnt-node
2461 list-storage``) and modify them (``gnt-node modify-storage``)
2462 - Added new ssconf files with master candidate IP address
2463 (``ssconf_master_candidates_ips``), node primary IP address
2464 (``ssconf_node_primary_ips``) and node secondary IP address
2465 (``ssconf_node_secondary_ips``)
2466 - Added ``ganeti-confd`` and a client library to query the Ganeti
2467 configuration via UDP
2468 - Added ability to run hooks after cluster initialization and before
2470 - Added automatic mode for disk replace (``gnt-instance replace-disks
2472 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2473 after catastrophic data-loss
2474 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2476 - Added ``gnt-instance move`` command to move instances
2477 - Added ``gnt-cluster watcher`` command to control watcher
2478 - Added ``gnt-node powercycle`` command to powercycle nodes
2479 - Added new job status field ``lock_status``
2480 - Added parseable error codes to cluster verification (``gnt-cluster
2481 verify --error-codes``) and made output less verbose (use
2482 ``--verbose`` to restore previous behaviour)
2483 - Added UUIDs to the main config entities (cluster, nodes, instances)
2484 - Added support for OS variants
2485 - Added support for hashed passwords in the Ganeti remote API users file
2487 - Added option to specify maximum timeout on instance shutdown
2488 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2489 - Added new helper script to start and stop Ganeti daemons
2490 (``daemon-util``), with the intent to reduce the work necessary to
2491 adjust Ganeti for non-Debian distributions and to start/stop daemons
2493 - Added more unittests
2494 - Fixed critical bug in ganeti-masterd startup
2495 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2496 now customisable at the cluster level for both the KVM and Xen
2497 hypervisors using the new ``migration_port`` parameter
2498 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2499 reinstalling an instance
2500 - Allowed ``@`` in tag names
2501 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2502 - Many documentation updates
2503 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2504 - ``gnt-instance reinstall`` can now reinstall multiple instances
2505 - Updated many command line parameters
2506 - Introduced new OS API version 15
2507 - No longer support a default hypervisor
2508 - Treat virtual LVs as inexistent
2509 - Improved job locking logic to reduce lock contention
2510 - Match instance and node names case insensitively
2511 - Reimplemented bash completion script to be more complete
2518 *(Released Thu, 4 Feb 2010)*
2520 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2521 - Fix a string formatting bug
2522 - Improve safety of the code in some error paths
2523 - Improve data validation in the master of values returned from nodes
2529 *(Released Thu, 17 Dec 2009)*
2531 - Fix security issue due to missing validation of iallocator names; this
2532 allows local and remote execution of arbitrary executables
2533 - Fix failure of gnt-node list during instance removal
2534 - Ship the RAPI documentation in the archive
2540 *(Released Wed, 30 Sep 2009)*
2542 - Fixed many wrong messages
2543 - Fixed a few bugs related to the locking library
2544 - Fixed MAC checking at instance creation time
2545 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2546 - Fixed a few issues related to signal handling in both daemons and
2548 - Fixed the example startup script provided
2549 - Fixed insserv dependencies in the example startup script (patch from
2551 - Fixed handling of drained nodes in the iallocator framework
2552 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2554 - Fixed error related to invalid job IDs in job polling
2555 - Fixed job/opcode persistence on unclean master shutdown
2556 - Fixed handling of partial job processing after unclean master
2558 - Fixed error reporting from LUs, previously all errors were converted
2559 into execution errors
2560 - Fixed error reporting from burnin
2561 - Decreased significantly the memory usage of the job queue
2562 - Optimised slightly multi-job submission
2563 - Optimised slightly opcode loading
2564 - Backported the multi-job submit framework from the development
2565 branch; multi-instance start and stop should be faster
2566 - Added script to clean archived jobs after 21 days; this will reduce
2567 the size of the queue directory
2568 - Added some extra checks in disk size tracking
2569 - Added an example ethers hook script
2570 - Added a cluster parameter that prevents Ganeti from modifying of
2572 - Added more node information to RAPI responses
2573 - Added a ``gnt-job watch`` command that allows following the ouput of a
2575 - Added a bind-address option to ganeti-rapi
2576 - Added more checks to the configuration verify
2577 - Enhanced the burnin script such that some operations can be retried
2579 - Converted instance reinstall to multi-instance model
2585 *(Released Fri, 7 Aug 2009)*
2587 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2588 to allow using the pre-2.0.2 behaviour in activation, if any existing
2589 instances have mismatched disk sizes in the configuration
2590 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2591 any configuration mismatches for disk sizes
2592 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2593 failover to work on two-node clusters
2594 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2596 - Fixed detection of OS script errors in ``gnt-backup export``
2597 - Fixed exit code of ``gnt-backup export``
2603 *(Released Fri, 17 Jul 2009)*
2605 - Added experimental support for stripped logical volumes; this should
2606 enhance performance but comes with a higher complexity in the block
2607 device handling; stripping is only enabled when passing
2608 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2609 affected even in the non-stripped mode
2610 - Improved resiliency against transient failures at the end of DRBD
2611 resyncs, and in general of DRBD resync checks
2612 - Fixed a couple of issues with exports and snapshot errors
2613 - Fixed a couple of issues in instance listing
2614 - Added display of the disk size in ``gnt-instance info``
2615 - Fixed checking for valid OSes in instance creation
2616 - Fixed handling of the "vcpus" parameter in instance listing and in
2617 general of invalid parameters
2618 - Fixed http server library, and thus RAPI, to handle invalid
2619 username/password combinations correctly; this means that now they
2620 report unauthorized for queries too, not only for modifications,
2621 allowing earlier detect of configuration problems
2622 - Added a new "role" node list field, equivalent to the master/master
2623 candidate/drained/offline flags combinations
2624 - Fixed cluster modify and changes of candidate pool size
2625 - Fixed cluster verify error messages for wrong files on regular nodes
2626 - Fixed a couple of issues with node demotion from master candidate role
2627 - Fixed node readd issues
2628 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2629 - Added a new ``--no-voting`` option for masterfailover to fix failover
2630 on two-nodes clusters when the former master node is unreachable
2631 - Added instance reinstall over RAPI
2637 *(Released Tue, 16 Jun 2009)*
2639 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2640 allow re-adding the start in single-user option (regression from 1.2)
2641 - the watcher writes the instance status to a file, to allow monitoring
2642 to report the instance status (from the master) based on cached
2643 results of the watcher's queries; while this can get stale if the
2644 watcher is being locked due to other work on the cluster, this is
2645 still an improvement
2646 - the watcher now also restarts the node daemon and the rapi daemon if
2648 - fixed the watcher to handle full and drained queue cases
2649 - hooks export more instance data in the environment, which helps if
2650 hook scripts need to take action based on the instance's properties
2651 (no longer need to query back into ganeti)
2652 - instance failovers when the instance is stopped do not check for free
2653 RAM, so that failing over a stopped instance is possible in low memory
2655 - rapi uses queries for tags instead of jobs (for less job traffic), and
2656 for cluster tags it won't talk to masterd at all but read them from
2658 - a couple of error handling fixes in RAPI
2659 - drbd handling: improved the error handling of inconsistent disks after
2660 resync to reduce the frequency of "there are some degraded disks for
2661 this instance" messages
2662 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2663 error handling path called a wrong function name)
2669 *(Released Wed, 27 May 2009)*
2671 - no changes from rc5
2677 *(Released Wed, 20 May 2009)*
2679 - fix a couple of bugs (validation, argument checks)
2680 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2681 - some small improvements to RAPI and IAllocator
2682 - make watcher automatically start the master daemon if down
2688 *(Released Mon, 27 Apr 2009)*
2690 - change the OS list to not require locks; this helps with big clusters
2691 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2692 volume group is broken
2693 - ``gnt-instance info``, without any arguments, doesn't run for all
2694 instances anymore; either pass ``--all`` or pass the desired
2695 instances; this helps against mistakes on big clusters where listing
2696 the information for all instances takes a long time
2697 - miscellaneous doc and man pages fixes
2703 *(Released Wed, 8 Apr 2009)*
2705 - Change the internal locking model of some ``gnt-node`` commands, in
2706 order to reduce contention (and blocking of master daemon) when
2707 batching many creation/reinstall jobs
2708 - Fixes to Xen soft reboot
2709 - No longer build documentation at build time, instead distribute it in
2710 the archive, in order to reduce the need for the whole docbook/rst
2717 *(Released Fri, 27 Mar 2009)*
2719 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2720 - Fix watcher startup sequence, improves the behaviour of busy clusters
2721 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2722 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2723 bind address and other places
2724 - Some documentation fixes and updates
2730 *(Released Mon, 2 Mar 2009)*
2732 - More documentation updates, now all docs should be more-or-less
2734 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2736 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2738 - Fix an issue related to $libdir/run/ganeti and cluster creation
2744 *(Released Thu, 19 Feb 2009)*
2746 - Xen PVM and KVM have switched the default value for the instance root
2747 disk to the first partition on the first drive, instead of the whole
2748 drive; this means that the OS installation scripts must be changed
2750 - Man pages have been updated
2751 - RAPI has been switched by default to HTTPS, and the exported functions
2752 should all work correctly
2753 - RAPI v1 has been removed
2754 - Many improvements to the KVM hypervisor
2755 - Block device errors are now better reported
2756 - Many other bugfixes and small improvements
2762 *(Released Mon, 26 Jan 2009)*
2764 - Version 2 is a general rewrite of the code and therefore the
2765 differences are too many to list, see the design document for 2.0 in
2766 the ``doc/`` subdirectory for more details
2767 - In this beta version there is not yet a migration path from 1.2 (there
2768 will be one in the final 2.0 release)
2769 - A few significant changes are:
2771 - all commands are executed by a daemon (``ganeti-masterd``) and the
2772 various ``gnt-*`` commands are just front-ends to it
2773 - all the commands are entered into, and executed from a job queue,
2774 see the ``gnt-job(8)`` manpage
2775 - the RAPI daemon supports read-write operations, secured by basic
2776 HTTP authentication on top of HTTPS
2777 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2778 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2779 to migrate to DRBD 8 first while still running Ganeti 1.2)
2780 - DRBD devices are using statically allocated minor numbers, which
2781 will be assigned to existing instances during the migration process
2782 - there is support for both Xen PVM and Xen HVM instances running on
2784 - KVM virtualization is supported too
2785 - file-based storage has been implemented, which means that it is
2786 possible to run the cluster without LVM and DRBD storage, for
2787 example using a shared filesystem exported from shared storage (and
2788 still have live migration)
2794 *(Released Tue, 13 Jan 2009)*
2796 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2797 - Reuse the old instance mac address by default on instance import, if
2798 the instance name is the same.
2799 - Handle situations in which the node info rpc returns incomplete
2801 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2802 - Improved version of batcher:
2804 - state file support
2805 - instance mac address support
2806 - support for HVM clusters/instances
2808 - Add an option to show the number of cpu sockets and nodes in
2810 - Support OSes that handle more than one version of the OS api (but do
2811 not change the current API in any other way)
2812 - Fix ``gnt-node migrate``
2813 - ``gnt-debug`` man page
2814 - Fixes various more typos and small issues
2815 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2821 *(Released Wed, 24 Sep 2008)*
2823 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2824 type of disk exported to fully virtualized instances.
2825 - provide access to the serial console of HVM instances
2826 - instance auto_balance flag, set by default. If turned off it will
2827 avoid warnings on cluster verify if there is not enough memory to fail
2828 over an instance. in the future it will prevent automatically failing
2829 it over when we will support that.
2830 - batcher tool for instance creation, see ``tools/README.batcher``
2831 - ``gnt-instance reinstall --select-os`` to interactively select a new
2832 operating system when reinstalling an instance.
2833 - when changing the memory amount on instance modify a check has been
2834 added that the instance will be able to start. also warnings are
2835 emitted if the instance will not be able to fail over, if auto_balance
2837 - documentation fixes
2838 - sync fields between ``gnt-instance list/modify/add/import``
2839 - fix a race condition in drbd when the sync speed was set after giving
2840 the device a remote peer.
2846 *(Released Tue, 22 Jul 2008)*
2848 - note: the allowed size and number of tags per object were reduced
2849 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2850 - fixed twisted 8.x compatibility
2851 - fixed ``gnt-instance replace-disks`` with iallocator
2852 - add TCP keepalives on twisted connections to detect restarted nodes
2853 - disk increase support, see ``gnt-instance grow-disk``
2854 - implement bulk node/instance query for RAPI
2855 - add tags in node/instance listing (optional)
2856 - experimental migration (and live migration) support, read the man page
2857 for ``gnt-instance migrate``
2858 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2859 has some small improvements in handling its state file
2865 *(Released Fri, 13 Jun 2008)*
2867 - Experimental readonly, REST-based remote API implementation;
2868 automatically started on master node, TCP port 5080, if enabled by
2869 ``--enable-rapi`` parameter to configure script.
2870 - Instance allocator support. Add and import instance accept a
2871 ``--iallocator`` parameter, and call that instance allocator to decide
2872 which node to use for the instance. The iallocator document describes
2873 what's expected from an allocator script.
2874 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2875 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2876 node is lost there is still enough memory to fail over the instances
2878 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2879 ``gnt-cluster verify``, to check for site-specific compliance. All the
2880 hooks will run, and their output, if any, will be displayed. Any
2881 failing hook will make the verification return an error value.
2882 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2883 primary and secondary interfaces
2884 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2885 that is still declared as part of the cluster and has failed.
2886 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2887 specifying output fields, that just adds the chosen fields to the
2889 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2890 export from the filesystem.
2891 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2892 have been added. Using them you can enable/disable acpi and pae
2893 support, and specify a path for a cd image to be exported to the
2894 instance. These parameters as the name suggest only work on HVM
2896 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2897 PAE support will be set to the previously hardcoded values, but the
2898 (previously hardcoded) path to the CDROM ISO image will be unset and
2899 if required, needs to be set manually with ``gnt-instance modify``
2901 - The address to which an instance's VNC console is bound is now
2902 selectable per-instance, rather than being cluster wide. Of course
2903 this only applies to instances controlled via VNC, so currently just
2904 applies to HVM clusters.
2910 *(Released Mon, 18 Feb 2008)*
2912 - more tweaks to the disk activation code (especially helpful for DRBD)
2913 - change the default ``gnt-instance list`` output format, now there is
2914 one combined status field (see the manpage for the exact values this
2916 - some more fixes for the mac export to hooks change
2917 - make Ganeti not break with DRBD 8.2.x (which changed the version
2918 format in ``/proc/drbd``) (issue 24)
2919 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2920 template, allowing migration from DRBD0.7+MD to DRBD8
2926 *(Released Wed, 30 Jan 2008)*
2928 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2930 - add command aliases infrastructure and a few aliases
2931 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2932 man pages and the ``--help`` option of ``gnt-node
2933 list``/``gnt-instance list``
2934 - fix ``gnt-backup list`` with down nodes (issue 21)
2935 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2936 - fix the dist archive and add a check for including svn/git files in
2938 - some developer-related changes: improve the burnin and the QA suite,
2939 add an upload script for testing during development
2945 *(Released Wed, 16 Jan 2008)*
2947 - experimental HVM support, read the install document, section
2948 "Initializing the cluster"
2949 - allow for the PVM hypervisor per-instance kernel and initrd paths
2950 - add a new command ``gnt-cluster verify-disks`` which uses a new
2951 algorithm to improve the reconnection of the DRBD pairs if the device
2952 on the secondary node has gone away
2953 - make logical volume code auto-activate LVs at disk activation time
2954 - slightly improve the speed of activating disks
2955 - allow specification of the MAC address at instance creation time, and
2956 changing it later via ``gnt-instance modify``
2957 - fix handling of external commands that generate lots of output on
2959 - update documentation with regard to minimum version of DRBD8 supported
2965 *(Released Tue, 4 Dec 2007)*
2967 - Log the ``xm create`` output to the node daemon log on failure (to
2968 help diagnosing the error)
2969 - In debug mode, log all external commands output if failed to the logs
2970 - Change parsing of lvm commands to ignore stderr
2976 *(Released Wed, 28 Nov 2007)*
2978 - Another round of updates to the DRBD 8 code to deal with more failures
2979 in the replace secondary node operation
2980 - Some more logging of failures in disk operations (lvm, drbd)
2981 - A few documentation updates
2988 *(Released Tue, 13 Nov 2007)*
2990 - Change configuration file format from Python's Pickle to JSON.
2991 Upgrading is possible using the cfgupgrade utility.
2992 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2993 faster replace disks and is more stable (DRBD 8 has many improvements
2994 compared to DRBD 0.7)
2995 - Added command line tags support (see man pages for ``gnt-instance``,
2996 ``gnt-node``, ``gnt-cluster``)
2997 - Added instance rename support
2998 - Added multi-instance startup/shutdown
2999 - Added cluster rename support
3000 - Added ``gnt-node evacuate`` to simplify some node operations
3001 - Added instance reboot operation that can speedup reboot as compared to
3003 - Soften the requirement that hostnames are in FQDN format
3004 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
3006 - Removed dependency on debian's patched fping that uses the
3007 non-standard ``-S`` option
3008 - Now the OS definitions are searched for in multiple, configurable
3009 paths (easier for distros to package)
3010 - Some changes to the hooks infrastructure (especially the new
3011 post-configuration update hook)
3012 - Other small bugfixes
3014 .. vim: set textwidth=72 syntax=rst :