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.
20 *(Released Thu, 29 Aug 2013)*
22 Incompatible/important changes
23 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
25 - hroller now also plans for capacity to move non-redundant instances off
26 any node to be rebooted; the old behavior of completely ignoring any
27 non-redundant instances can be restored by adding the --ignore-non-redundant
29 - The cluster option '--no-lvm-storage' was removed in favor of the new option
30 '--enabled-disk-templates'.
31 - On instance creation, disk templates no longer need to be specified
32 with '-t'. The default disk template will be taken from the list of
33 enabled disk templates.
34 - The monitoring daemon is now running as root, in order to be able to collect
35 information only available to root (such as the state of Xen instances).
36 - The ConfD client is now IPv6 compatible.
37 - File and shared file storage is no longer dis/enabled at configure time,
38 but using the option '--enabled-disk-templates' at cluster initialization and
40 - The default directories for file and shared file storage are not anymore
41 specified at configure time, but taken from the cluster's configuration.
42 They can be set at cluster initialization and modification with
43 '--file-storage-dir' and '--shared-file-storage-dir'.
44 - Cluster verification now includes stricter checks regarding the
45 default file and shared file storage directories. It now checks that
46 the directories are explicitely allowed in the 'file-storage-paths' file and
47 that the directories exist on all nodes.
48 - The list of allowed disk templates in the instance policy and the list
49 of cluster-wide enabled disk templates is now checked for consistency
50 on cluster or group modification. On cluster initialization, the ipolicy
51 disk templates are ensured to be a subset of the cluster-wide enabled
57 - DRBD 8.4 support. Depending on the installed DRBD version, Ganeti now uses
58 the correct command syntax. It is possible to use different DRBD versions
59 on different nodes as long as they are compatible to each other. This
60 enables rolling upgrades of DRBD with no downtime. As permanent operation
61 of different DRBD versions within a node group is discouraged,
62 ``gnt-cluster verify`` will emit a warning if it detects such a situation.
63 - New "inst-status-xen" data collector for the monitoring daemon, providing
64 information about the state of the xen instances on the nodes.
65 - New "lv" data collector for the monitoring daemon, collecting data about the
66 logical volumes on the nodes, and pairing them with the name of the instances
68 - New "diskstats" data collector, collecting the data from /proc/diskstats and
69 presenting them over the monitoring daemon interface.
70 - The ConfD client is now IPv6 compatible.
74 The following new dependencies have been added.
78 - ``python-mock`` (http://www.voidspace.org.uk/python/mock/) is now a required
79 for the unit tests (and only used for testing).
85 *(Released Tue, 17 Sep 2013)*
87 Incompatible/important changes
88 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
90 - Instance policy can contain multiple instance specs, as described in
91 the “Constrained instance sizes” section of :doc:`Partitioned Ganeti
92 <design-partitioned>`. As a consequence, it's not possible to partially change
93 or override instance specs. Bounding specs (min and max) can be specified as a
94 whole using the new option ``--ipolicy-bounds-specs``, while standard
95 specs use the new option ``--ipolicy-std-specs``.
96 - The output of the info command of gnt-cluster, gnt-group, gnt-node,
97 gnt-instance is a valid YAML object.
98 - hail now honors network restrictions when allocating nodes. This led to an
99 update of the IAllocator protocol. See the IAllocator documentation for
101 - confd now only answers static configuration request over the network. luxid
102 was extracted, listens on the local LUXI socket and responds to live queries.
103 This allows finer grained permissions if using separate users.
108 - The :doc:`Remote API <rapi>` daemon now supports a command line flag
109 to always require authentication, ``--require-authentication``. It can
110 be specified in ``$sysconfdir/default/ganeti``.
111 - A new cluster attribute 'enabled_disk_templates' is introduced. It will
112 be used to manage the disk templates to be used by instances in the cluster.
113 Initially, it will be set to a list that includes plain, drbd, if they were
114 enabled by specifying a volume group name, and file and sharedfile, if those
115 were enabled at configure time. Additionally, it will include all disk
116 templates that are currently used by instances. The order of disk templates
117 will be based on Ganeti's history of supporting them. In the future, the
118 first entry of the list will be used as a default disk template on instance
120 - ``cfgupgrade`` now supports a ``--downgrade`` option to bring the
121 configuration back to the previous stable version.
122 - Disk templates in group ipolicy can be restored to the default value.
123 - Initial support for diskless instances and virtual clusters in QA.
124 - More QA and unit tests for instance policies.
125 - Every opcode now contains a reason trail (visible through ``gnt-job info``)
126 describing why the opcode itself was executed.
127 - The monitoring daemon is now available. It allows users to query the cluster
128 for obtaining information about the status of the system. The daemon is only
129 responsible for providing the information over the network: the actual data
130 gathering is performed by data collectors (currently, only the DRBD status
131 collector is available).
132 - In order to help developers work on Ganeti, a new script
133 (``devel/build_chroot``) is provided, for building a chroot that contains all
134 the required development libraries and tools for compiling Ganeti on a Debian
136 - A new tool, ``harep``, for performing self-repair and recreation of instances
137 in Ganeti has been added.
138 - Split queries are enabled for tags, network, exports, cluster info, groups,
140 - New command ``show-ispecs-cmd`` for ``gnt-cluster`` and ``gnt-group``.
141 It prints the command line to set the current policies, to ease
143 - Add the ``vnet_hdr`` HV parameter for KVM, to control whether the tap
144 devices for KVM virtio-net interfaces will get created with VNET_HDR
145 (IFF_VNET_HDR) support. If set to false, it disables offloading on the
146 virtio-net interfaces, which prevents host kernel tainting and log
147 flooding, when dealing with broken or malicious virtio-net drivers.
148 It's set to true by default.
149 - Instance failover now supports a ``--cleanup`` parameter for fixing previous
151 - Support 'viridian' parameter in Xen HVM
152 - Support DSA SSH keys in bootstrap
153 - To simplify the work of packaging frameworks that want to add the needed users
154 and groups in a split-user setup themselves, at build time three files in
155 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
156 the groups to be generated, the ``users`` file contains, one per line, the
157 users to be generated, optionally followed by their primary group, where
158 important. The ``groupmemberships`` file contains, one per line, additional
159 user-group membership relations that need to be established. The syntax of
160 these files will remain stable in all future versions.
165 The following new dependencies have been added:
168 - The ``curl`` library is not optional anymore for compiling the Haskell code.
169 - ``snap-server`` library (if monitoring is enabled).
172 - The minimum Python version needed to run Ganeti is now 2.6.
173 - ``yaml`` library (only for running the QA).
178 - To simplify the work of packaging frameworks that want to add the needed users
179 and groups in a split-user setup themselves, at build time three files in
180 ``doc/users`` will be generated. The ``groups`` files contains, one per line,
181 the groups to be generated, the ``users`` file contains, one per line, the
182 users to be generated, optionally followed by their primary group, where
183 important. The ``groupmemberships`` file contains, one per line, additional
184 user-group membership relations that need to be established. The syntax of
185 these files will remain stable in all future versions.
186 - Add a default to file-driver when unspecified over RAPI (Issue 571)
187 - Mark the DSA host pubkey as optional, and remove it during config downgrade
189 - Some documentation fixes
195 *(Released Tue, 27 Aug 2013)*
197 The second release candidate of the 2.8 series. Since 2.8.0. rc1:
199 - Support 'viridian' parameter in Xen HVM (Issue 233)
200 - Include VCS version in ``gnt-cluster version``
201 - Support DSA SSH keys in bootstrap (Issue 338)
202 - Fix batch creation of instances
203 - Use FQDN to check master node status (Issue 551)
204 - Make the DRBD collector more failure-resilient
210 *(Released Fri, 2 Aug 2013)*
212 The first release candidate of the 2.8 series. Since 2.8.0 beta1:
214 - Fix upgrading/downgrading from 2.7
215 - Increase maximum RAPI message size
216 - Documentation updates
217 - Split ``confd`` between ``luxid`` and ``confd``
218 - Merge 2.7 series up to the 2.7.1 release
219 - Allow the ``modify_etc_hosts`` option to be changed
220 - Add better debugging for ``luxid`` queries
221 - Expose bulk parameter for GetJobs in RAPI client
222 - Expose missing ``network`` fields in RAPI
223 - Add some ``cluster verify`` tests
224 - Some unittest fixes
225 - Fix a malfunction in ``hspace``'s tiered allocation
226 - Fix query compatibility between haskell and python implementations
227 - Add the ``vnet_hdr`` HV parameter for KVM
228 - Add ``--cleanup`` to instance failover
229 - Change the connected groups format in ``gnt-network info`` output; it
230 was previously displayed as a raw list by mistake. (Merged from 2.7)
236 *(Released Mon, 24 Jun 2013)*
238 This was the first beta release of the 2.8 series. All important changes
239 are listed in the latest 2.8 entry.
245 *(Released Thu, 25 Jul 2013)*
247 - Add logrotate functionality in daemon-util
248 - Add logrotate example file
249 - Add missing fields to network queries over rapi
250 - Fix network object timestamps
251 - Add support for querying network timestamps
252 - Fix a typo in the example crontab
253 - Fix a documentation typo
259 *(Released Thu, 04 Jul 2013)*
261 Incompatible/important changes
262 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
264 - Instance policies for disk size were documented to be on a per-disk
265 basis, but hail applied them to the sum of all disks. This has been
267 - ``hbal`` will now exit with status 0 if, during job execution over
268 LUXI, early exit has been requested and all jobs are successful;
269 before, exit status 1 was used, which cannot be differentiated from
271 - Compatibility with newer versions of rbd has been fixed
272 - ``gnt-instance batch-create`` has been changed to use the bulk create
273 opcode from Ganeti. This lead to incompatible changes in the format of
274 the JSON file. It's now not a custom dict anymore but a dict
275 compatible with the ``OpInstanceCreate`` opcode.
276 - Parent directories for file storage need to be listed in
277 ``$sysconfdir/ganeti/file-storage-paths`` now. ``cfgupgrade`` will
278 write the file automatically based on old configuration values, but it
279 can not distribute it across all nodes and the file contents should be
280 verified. Use ``gnt-cluster copyfile
281 $sysconfdir/ganeti/file-storage-paths`` once the cluster has been
282 upgraded. The reason for requiring this list of paths now is that
283 before it would have been possible to inject new paths via RPC,
284 allowing files to be created in arbitrary locations. The RPC protocol
285 is protected using SSL/X.509 certificates, but as a design principle
286 Ganeti does not permit arbitrary paths to be passed.
287 - The parsing of the variants file for OSes (see
288 :manpage:`ganeti-os-interface(7)`) has been slightly changed: now empty
289 lines and comment lines (starting with ``#``) are ignored for better
291 - The ``setup-ssh`` tool added in Ganeti 2.2 has been replaced and is no
292 longer available. ``gnt-node add`` now invokes a new tool on the
293 destination node, named ``prepare-node-join``, to configure the SSH
294 daemon. Paramiko is no longer necessary to configure nodes' SSH
295 daemons via ``gnt-node add``.
296 - Draining (``gnt-cluster queue drain``) and un-draining the job queue
297 (``gnt-cluster queue undrain``) now affects all nodes in a cluster and
298 the flag is not reset after a master failover.
299 - Python 2.4 has *not* been tested with this release. Using 2.6 or above
300 is recommended. 2.6 will be mandatory from the 2.8 series.
306 - New network management functionality to support automatic allocation
307 of IP addresses and managing of network parameters. See
308 :manpage:`gnt-network(8)` for more details.
309 - New external storage backend, to allow managing arbitrary storage
310 systems external to the cluster. See
311 :manpage:`ganeti-extstorage-interface(7)`.
312 - New ``exclusive-storage`` node parameter added, restricted to
313 nodegroup level. When it's set to true, physical disks are assigned in
314 an exclusive fashion to instances, as documented in :doc:`Partitioned
315 Ganeti <design-partitioned>`. Currently, only instances using the
316 ``plain`` disk template are supported.
317 - The KVM hypervisor has been updated with many new hypervisor
318 parameters, including a generic one for passing arbitrary command line
319 values. See a complete list in :manpage:`gnt-instance(8)`. It is now
320 compatible up to qemu 1.4.
321 - A new tool, called ``mon-collector``, is the stand-alone executor of
322 the data collectors for a monitoring system. As of this version, it
323 just includes the DRBD data collector, that can be executed by calling
324 ``mon-collector`` using the ``drbd`` parameter. See
325 :manpage:`mon-collector(7)`.
326 - A new user option, :pyeval:`rapi.RAPI_ACCESS_READ`, has been added
327 for RAPI users. It allows granting permissions to query for
328 information to a specific user without giving
329 :pyeval:`rapi.RAPI_ACCESS_WRITE` permissions.
330 - A new tool named ``node-cleanup`` has been added. It cleans remains of
331 a cluster from a machine by stopping all daemons, removing
332 certificates and ssconf files. Unless the ``--no-backup`` option is
333 given, copies of the certificates are made.
334 - Instance creations now support the use of opportunistic locking,
335 potentially speeding up the (parallel) creation of multiple instances.
336 This feature is currently only available via the :doc:`RAPI
337 <rapi>` interface and when an instance allocator is used. If the
338 ``opportunistic_locking`` parameter is set the opcode will try to
339 acquire as many locks as possible, but will not wait for any locks
340 held by other opcodes. If not enough resources can be found to
341 allocate the instance, the temporary error code
342 :pyeval:`errors.ECODE_TEMP_NORES` is returned. The operation can be
343 retried thereafter, with or without opportunistic locking.
344 - New experimental linux-ha resource scripts.
345 - Restricted-commands support: ganeti can now be asked (via command line
346 or rapi) to perform commands on a node. These are passed via ganeti
347 RPC rather than ssh. This functionality is restricted to commands
348 specified on the ``$sysconfdir/ganeti/restricted-commands`` for security
349 reasons. The file is not copied automatically.
355 - Diskless instances are now externally mirrored (Issue 237). This for
356 now has only been tested in conjunction with explicit target nodes for
358 - Queries not needing locks or RPC access to the node can now be
359 performed by the confd daemon, making them independent from jobs, and
360 thus faster to execute. This is selectable at configure time.
361 - The functionality for allocating multiple instances at once has been
362 overhauled and is now also available through :doc:`RAPI <rapi>`.
364 There are no significant changes from version 2.7.0~rc3.
370 *(Released Tue, 25 Jun 2013)*
372 - Fix permissions on the confd query socket (Issue 477)
373 - Fix permissions on the job archive dir (Issue 498)
374 - Fix handling of an internal exception in replace-disks (Issue 472)
375 - Fix gnt-node info handling of shortened names (Issue 497)
376 - Fix gnt-instance grow-disk when wiping is enabled
377 - Documentation improvements, and support for newer pandoc
378 - Fix hspace honoring ipolicy for disks (Issue 484)
379 - Improve handling of the ``kvm_extra`` HV parameter
385 *(Released Fri, 24 May 2013)*
387 - ``devel/upload`` now works when ``/var/run`` on the target nodes is a
389 - Disks added through ``gnt-instance modify`` or created through
390 ``gnt-instance recreate-disks`` are wiped, if the
391 ``prealloc_wipe_disks`` flag is set.
392 - If wiping newly created disks fails, the disks are removed. Also,
393 partial failures in creating disks through ``gnt-instance modify``
394 triggers a cleanup of the partially-created disks.
395 - Removing the master IP address doesn't fail if the address has been
397 - Fix ownership of the OS log dir
398 - Workaround missing SO_PEERCRED constant (Issue 191)
404 *(Released Fri, 3 May 2013)*
406 This was the first release candidate of the 2.7 series. Since beta3:
408 - Fix kvm compatibility with qemu 1.4 (Issue 389)
409 - Documentation updates (admin guide, upgrade notes, install
410 instructions) (Issue 372)
411 - Fix gnt-group list nodes and instances count (Issue 436)
412 - Fix compilation without non-mandatory libraries (Issue 441)
413 - Fix xen-hvm hypervisor forcing nics to type 'ioemu' (Issue 247)
414 - Make confd logging more verbose at INFO level (Issue 435)
415 - Improve "networks" documentation in :manpage:`gnt-instance(8)`
416 - Fix failure path for instance storage type conversion (Issue 229)
417 - Update htools text backend documentation
418 - Improve the renew-crypto section of :manpage:`gnt-cluster(8)`
419 - Disable inter-cluster instance move for file-based instances, because
420 it is dependant on instance export, which is not supported for
421 file-based instances. (Issue 414)
422 - Fix gnt-job crashes on non-ascii characters (Issue 427)
423 - Fix volume group checks on non-vm-capable nodes (Issue 432)
429 *(Released Mon, 22 Apr 2013)*
431 This was the third beta release of the 2.7 series. Since beta2:
433 - Fix hail to verify disk instance policies on a per-disk basis (Issue 418).
434 - Fix data loss on wrong usage of ``gnt-instance move``
435 - Properly export errors in confd-based job queries
436 - Add ``users-setup`` tool
437 - Fix iallocator protocol to report 0 as a disk size for diskless
438 instances. This avoids hail breaking when a diskless instance is
440 - Fix job queue directory permission problem that made confd job queries
441 fail. This requires running an ``ensure-dirs --full-run`` on upgrade
442 for access to archived jobs (Issue 406).
443 - Limit the sizes of networks supported by ``gnt-network`` to something
444 between a ``/16`` and a ``/30`` to prevent memory bloat and crashes.
445 - Fix bugs in instance disk template conversion
446 - Fix GHC 7 compatibility
447 - Fix ``burnin`` install path (Issue 426).
448 - Allow very small disk grows (Issue 347).
449 - Fix a ``ganeti-noded`` memory bloat introduced in 2.5, by making sure
450 that noded doesn't import masterd code (Issue 419).
451 - Make sure the default metavg at cluster init is the same as the vg, if
452 unspecified (Issue 358).
453 - Fix cleanup of partially created disks (part of Issue 416)
459 *(Released Tue, 2 Apr 2013)*
461 This was the second beta release of the 2.7 series. Since beta1:
463 - Networks no longer have a "type" slot, since this information was
464 unused in Ganeti: instead of it tags should be used.
465 - The rapi client now has a ``target_node`` option to MigrateInstance.
466 - Fix early exit return code for hbal (Issue 386).
467 - Fix ``gnt-instance migrate/failover -n`` (Issue 396).
468 - Fix ``rbd showmapped`` output parsing (Issue 312).
469 - Networks are now referenced indexed by UUID, rather than name. This
470 will require running cfgupgrade, from 2.7.0beta1, if networks are in
472 - The OS environment now includes network information.
473 - Deleting of a network is now disallowed if any instance nic is using
474 it, to prevent dangling references.
475 - External storage is now documented in man pages.
476 - The exclusive_storage flag can now only be set at nodegroup level.
477 - Hbal can now submit an explicit priority with its jobs.
478 - Many network related locking fixes.
479 - Bump up the required pylint version to 0.25.1.
480 - Fix the ``no_remember`` option in RAPI client.
481 - Many ipolicy related tests, qa, and fixes.
482 - Many documentation improvements and fixes.
483 - Fix building with ``--disable-file-storage``.
484 - Fix ``-q`` option in htools, which was broken if passed more than
486 - Some haskell/python interaction improvements and fixes.
487 - Fix iallocator in case of missing LVM storage.
488 - Fix confd config load in case of ``--no-lvm-storage``.
489 - The confd/query functionality is now mentioned in the security
496 *(Released Wed, 6 Feb 2013)*
498 This was the first beta release of the 2.7 series. All important changes
499 are listed in the latest 2.7 entry.
505 *(Released Fri, 21 Dec 2012)*
507 Important behaviour change: hbal won't rebalance anymore instances which
508 have the ``auto_balance`` attribute set to false. This was the intention
509 all along, but until now it only skipped those from the N+1 memory
510 reservation (DRBD-specific).
512 A significant number of bug fixes in this release:
514 - Fixed disk adoption interaction with ipolicy checks.
515 - Fixed networking issues when instances are started, stopped or
516 migrated, by forcing the tap device's MAC prefix to "fe" (issue 217).
517 - Fixed the warning in cluster verify for shared storage instances not
519 - Fixed removal of storage directory on shared file storage (issue 262).
520 - Fixed validation of LVM volume group name in OpClusterSetParams
521 (``gnt-cluster modify``) (issue 285).
522 - Fixed runtime memory increases (``gnt-instance modify -m``).
523 - Fixed live migration under Xen's ``xl`` mode.
524 - Fixed ``gnt-instance console`` with ``xl``.
525 - Fixed building with newer Haskell compiler/libraries.
526 - Fixed PID file writing in Haskell daemons (confd); this prevents
527 restart issues if confd was launched manually (outside of
528 ``daemon-util``) while another copy of it was running
529 - Fixed a type error when doing live migrations with KVM (issue 297) and
530 the error messages for failing migrations have been improved.
531 - Fixed opcode validation for the out-of-band commands (``gnt-node
533 - Fixed a type error when unsetting OS hypervisor parameters (issue
534 311); now it's possible to unset all OS-specific hypervisor
536 - Fixed the ``dry-run`` mode for many operations: verification of
537 results was over-zealous but didn't take into account the ``dry-run``
538 operation, resulting in "wrong" failures.
539 - Fixed bash completion in ``gnt-job list`` when the job queue has
540 hundreds of entries; especially with older ``bash`` versions, this
541 results in significant CPU usage.
543 And lastly, a few other improvements have been made:
545 - Added option to force master-failover without voting (issue 282).
546 - Clarified error message on lock conflict (issue 287).
547 - Logging of newly submitted jobs has been improved (issue 290).
548 - Hostname checks have been made uniform between instance rename and
550 - The ``--submit`` option is now supported by ``gnt-debug delay``.
551 - Shutting down the master daemon by sending SIGTERM now stops it from
552 processing jobs waiting for locks; instead, those jobs will be started
553 once again after the master daemon is started the next time (issue
555 - Support for Xen's ``xl`` program has been improved (besides the fixes
557 - Reduced logging noise in the Haskell confd daemon (only show one log
558 entry for each config reload, instead of two).
559 - Several man page updates and typo fixes.
565 *(Released Fri, 12 Oct 2012)*
567 A small bugfix release. Among the bugs fixed:
569 - Fixed double use of ``PRIORITY_OPT`` in ``gnt-node migrate``, that
570 made the command unusable.
571 - Commands that issue many jobs don't fail anymore just because some jobs
572 take so long that other jobs are archived.
573 - Failures during ``gnt-instance reinstall`` are reflected by the exit
575 - Issue 190 fixed. Check for DRBD in cluster verify is enabled only when
577 - When ``always_failover`` is set, ``--allow-failover`` is not required
578 in migrate commands anymore.
579 - ``bash_completion`` works even if extglob is disabled.
580 - Fixed bug with locks that made failover for RDB-based instances fail.
581 - Fixed bug in non-mirrored instance allocation that made Ganeti choose
582 a random node instead of one based on the allocator metric.
583 - Support for newer versions of pylint and pep8.
584 - Hail doesn't fail anymore when trying to add an instance of type
585 ``file``, ``sharedfile`` or ``rbd``.
586 - Added new Makefile target to rebuild the whole distribution, so that
587 all files are included.
593 *(Released Fri, 27 Jul 2012)*
596 .. attention:: The ``LUXI`` protocol has been made more consistent
597 regarding its handling of command arguments. This, however, leads to
598 incompatibility issues with previous versions. Please ensure that you
599 restart Ganeti daemons soon after the upgrade, otherwise most
600 ``LUXI`` calls (job submission, setting/resetting the drain flag,
601 pausing/resuming the watcher, cancelling and archiving jobs, querying
602 the cluster configuration) will fail.
611 The current ``admin_up`` field, which used to denote whether an instance
612 should be running or not, has been removed. Instead, ``admin_state`` is
613 introduced, with 3 possible values -- ``up``, ``down`` and ``offline``.
615 The rational behind this is that an instance being “down” can have
618 - it could be down during a reboot
619 - it could be temporarily be down for a reinstall
620 - or it could be down because it is deprecated and kept just for its
623 The previous Boolean state was making it difficult to do capacity
624 calculations: should Ganeti reserve memory for a down instance? Now, the
625 tri-state field makes it clear:
627 - in ``up`` and ``down`` state, all resources are reserved for the
628 instance, and it can be at any time brought up if it is down
629 - in ``offline`` state, only disk space is reserved for it, but not
632 The field can have an extra use: since the transition between ``up`` and
633 ``down`` and vice-versus is done via ``gnt-instance start/stop``, but
634 transition between ``offline`` and ``down`` is done via ``gnt-instance
635 modify``, it is possible to given different rights to users. For
636 example, owners of an instance could be allowed to start/stop it, but
637 not transition it out of the offline state.
639 Instance policies and specs
640 +++++++++++++++++++++++++++
642 In previous Ganeti versions, an instance creation request was not
643 limited on the minimum size and on the maximum size just by the cluster
644 resources. As such, any policy could be implemented only in third-party
645 clients (RAPI clients, or shell wrappers over ``gnt-*``
646 tools). Furthermore, calculating cluster capacity via ``hspace`` again
647 required external input with regards to instance sizes.
649 In order to improve these workflows and to allow for example better
650 per-node group differentiation, we introduced instance specs, which
653 - minimum instance disk size, disk count, memory size, cpu count
654 - maximum values for the above metrics
655 - and “standard” values (used in ``hspace`` to calculate the standard
658 The minimum/maximum values can be also customised at node-group level,
659 for example allowing more powerful hardware to support bigger instance
662 Beside the instance specs, there are a few other settings belonging to
663 the instance policy framework. It is possible now to customise, per
664 cluster and node-group:
666 - the list of allowed disk templates
667 - the maximum ratio of VCPUs per PCPUs (to control CPU oversubscription)
668 - the maximum ratio of instance to spindles (see below for more
669 information) for local storage
671 All these together should allow all tools that talk to Ganeti to know
672 what are the ranges of allowed values for instances and the
673 over-subscription that is allowed.
675 For the VCPU/PCPU ratio, we already have the VCPU configuration from the
676 instance configuration, and the physical CPU configuration from the
677 node. For the spindle ratios however, we didn't track before these
678 values, so new parameters have been added:
680 - a new node parameter ``spindle_count``, defaults to 1, customisable at
681 node group or node level
682 - at new backend parameter (for instances), ``spindle_use`` defaults to 1
684 Note that spindles in this context doesn't need to mean actual
685 mechanical hard-drives; it's just a relative number for both the node
686 I/O capacity and instance I/O consumption.
688 Instance migration behaviour
689 ++++++++++++++++++++++++++++
691 While live-migration is in general desirable over failover, it is
692 possible that for some workloads it is actually worse, due to the
693 variable time of the “suspend” phase during live migration.
695 To allow the tools to work consistently over such instances (without
696 having to hard-code instance names), a new backend parameter
697 ``always_failover`` has been added to control the migration/failover
698 behaviour. When set to True, all migration requests for an instance will
699 instead fall-back to failover.
701 Instance memory ballooning
702 ++++++++++++++++++++++++++
704 Initial support for memory ballooning has been added. The memory for an
705 instance is no longer fixed (backend parameter ``memory``), but instead
706 can vary between minimum and maximum values (backend parameters
707 ``minmem`` and ``maxmem``). Currently we only change an instance's
710 - live migrating or failing over and instance and the target node
711 doesn't have enough memory
712 - user requests changing the memory via ``gnt-instance modify
718 In order to control the use of specific CPUs by instance, support for
719 controlling CPU pinning has been added for the Xen, HVM and LXC
720 hypervisors. This is controlled by a new hypervisor parameter
721 ``cpu_mask``; details about possible values for this are in the
722 :manpage:`gnt-instance(8)`. Note that use of the most specific (precise
723 VCPU-to-CPU mapping) form will work well only when all nodes in your
724 cluster have the same amount of CPUs.
729 Another area in which Ganeti was not customisable were the parameters
730 used for storage configuration, e.g. how many stripes to use for LVM,
731 DRBD resync configuration, etc.
733 To improve this area, we've added disks parameters, which are
734 customisable at cluster and node group level, and which allow to
735 specify various parameters for disks (DRBD has the most parameters
736 currently), for example:
738 - DRBD resync algorithm and parameters (e.g. speed)
739 - the default VG for meta-data volumes for DRBD
740 - number of stripes for LVM (plain disk template)
743 These parameters can be modified via ``gnt-cluster modify -D …`` and
744 ``gnt-group modify -D …``, and are used at either instance creation (in
745 case of LVM stripes, for example) or at disk “activation” time
748 Rados block device support
749 ++++++++++++++++++++++++++
751 A Rados (http://ceph.com/wiki/Rbd) storage backend has been added,
752 denoted by the ``rbd`` disk template type. This is considered
753 experimental, feedback is welcome. For details on configuring it, see
754 the :doc:`install` document and the :manpage:`gnt-cluster(8)` man page.
759 The existing master IP functionality works well only in simple setups (a
760 single network shared by all nodes); however, if nodes belong to
761 different networks, then the ``/32`` setup and lack of routing
762 information is not enough.
764 To allow the master IP to function well in more complex cases, the
765 system was reworked as follows:
767 - a master IP netmask setting has been added
768 - the master IP activation/turn-down code was moved from the node daemon
770 - whether to run the Ganeti-supplied master IP script or a user-supplied
771 on is a ``gnt-cluster init`` setting
773 Details about the location of the standard and custom setup scripts are
774 in the man page :manpage:`gnt-cluster(8)`; for information about the
775 setup script protocol, look at the Ganeti-supplied script.
780 The `SPICE <http://www.linux-kvm.org/page/SPICE>`_ support has been
783 It is now possible to use TLS-protected connections, and when renewing
784 or changing the cluster certificates (via ``gnt-cluster renew-crypto``,
785 it is now possible to specify spice or spice CA certificates. Also, it
786 is possible to configure a password for SPICE sessions via the
787 hypervisor parameter ``spice_password_file``.
789 There are also new parameters to control the compression and streaming
790 options (e.g. ``spice_image_compression``, ``spice_streaming_video``,
791 etc.). For details, see the man page :manpage:`gnt-instance(8)` and look
792 for the spice parameters.
794 Lastly, it is now possible to see the SPICE connection information via
795 ``gnt-instance console``.
800 A new tool (``tools/ovfconverter``) has been added that supports
801 conversion between Ganeti and the `Open Virtualization Format
802 <http://en.wikipedia.org/wiki/Open_Virtualization_Format>`_ (both to and
805 This relies on the ``qemu-img`` tool to convert the disk formats, so the
806 actual compatibility with other virtualization solutions depends on it.
811 The configuration query daemon (``ganeti-confd``) is now optional, and
812 has been rewritten in Haskell; whether to use the daemon at all, use the
813 Python (default) or the Haskell version is selectable at configure time
814 via the ``--enable-confd`` parameter, which can take one of the
815 ``haskell``, ``python`` or ``no`` values. If not used, disabling the
816 daemon will result in a smaller footprint; for larger systems, we
817 welcome feedback on the Haskell version which might become the default
820 If you want to use ``gnt-node list-drbd`` you need to have the Haskell
821 daemon running. The Python version doesn't implement the new call.
824 User interface changes
825 ~~~~~~~~~~~~~~~~~~~~~~
827 We have replaced the ``--disks`` option of ``gnt-instance
828 replace-disks`` with a more flexible ``--disk`` option, which allows
829 adding and removing disks at arbitrary indices (Issue 188). Furthermore,
830 disk size and mode can be changed upon recreation (via ``gnt-instance
831 recreate-disks``, which accepts the same ``--disk`` option).
833 As many people are used to a ``show`` command, we have added that as an
834 alias to ``info`` on all ``gnt-*`` commands.
836 The ``gnt-instance grow-disk`` command has a new mode in which it can
837 accept the target size of the disk, instead of the delta; this can be
838 more safe since two runs in absolute mode will be idempotent, and
839 sometimes it's also easier to specify the desired size directly.
841 Also the handling of instances with regard to offline secondaries has
842 been improved. Instance operations should not fail because one of it's
843 secondary nodes is offline, even though it's safe to proceed.
845 A new command ``list-drbd`` has been added to the ``gnt-node`` script to
846 support debugging of DRBD issues on nodes. It provides a mapping of DRBD
847 minors to instance name.
852 RAPI coverage has improved, with (for example) new resources for
853 recreate-disks, node power-cycle, etc.
858 There is partial support for ``xl`` in the Xen hypervisor; feedback is
861 Python 2.7 is better supported, and after Ganeti 2.6 we will investigate
862 whether to still support Python 2.4 or move to Python 2.6 as minimum
865 Support for Fedora has been slightly improved; the provided example
866 init.d script should work better on it and the INSTALL file should
867 document the needed dependencies.
872 The deprecated ``QueryLocks`` LUXI request has been removed. Use
873 ``Query(what=QR_LOCK, ...)`` instead.
875 The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
876 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
877 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
878 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
879 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
881 RAPI client: ``CertificateError`` now derives from
882 ``GanetiApiError``. This should make it more easy to handle Ganeti
885 Deprecation warnings due to PyCrypto/paramiko import in
886 ``tools/setup-ssh`` have been silenced, as usually they are safe; please
887 make sure to run an up-to-date paramiko version, if you use this tool.
889 The QA scripts now depend on Python 2.5 or above (the main code base
890 still works with Python 2.4).
892 The configuration file (``config.data``) is now written without
893 indentation for performance reasons; if you want to edit it, it can be
894 re-formatted via ``tools/fmtjson``.
896 A number of bugs has been fixed in the cluster merge tool.
898 ``x509`` certification verification (used in import-export) has been
899 changed to allow the same clock skew as permitted by the cluster
900 verification. This will remove some rare but hard to diagnose errors in
907 *(Released Thu, 19 Jul 2012)*
909 Very few changes from rc4 to the final release, only bugfixes:
911 - integrated fixes from release 2.5.2 (fix general boot flag for KVM
912 instance, fix CDROM booting for KVM instances)
913 - fixed node group modification of node parameters
914 - fixed issue in LUClusterVerifyGroup with multi-group clusters
915 - fixed generation of bash completion to ensure a stable ordering
922 *(Released Fri, 13 Jul 2012)*
924 Third release candidate for 2.6. The following changes were done from
927 - Fixed ``UpgradeConfig`` w.r.t. to disk parameters on disk objects.
928 - Fixed an inconsistency in the LUXI protocol with the provided
929 arguments (NOT backwards compatible)
930 - Fixed a bug with node groups ipolicy where ``min`` was greater than
931 the cluster ``std`` value
932 - Implemented a new ``gnt-node list-drbd`` call to list DRBD minors for
933 easier instance debugging on nodes (requires ``hconfd`` to work)
939 *(Released Tue, 03 Jul 2012)*
941 Second release candidate for 2.6. The following changes were done from
944 - Fixed ``gnt-cluster verify`` regarding ``master-ip-script`` on non
946 - Fixed a RAPI regression on missing beparams/memory
947 - Fixed redistribution of files on offline nodes
948 - Added possibility to run activate-disks even though secondaries are
949 offline. With this change it relaxes also the strictness on some other
950 commands which use activate disks internally:
951 * ``gnt-instance start|reboot|rename|backup|export``
952 - Made it possible to remove safely an instance if its secondaries are
954 - Made it possible to reinstall even though secondaries are offline
960 *(Released Mon, 25 Jun 2012)*
962 First release candidate for 2.6. The following changes were done from
965 - Fixed bugs with disk parameters and ``rbd`` templates as well as
967 - Made ``gnt-instance modify`` more consistent regarding new NIC/Disk
968 behaviour. It supports now the modify operation
969 - ``hcheck`` implemented to analyze cluster health and possibility of
970 improving health by rebalance
971 - ``hbal`` has been improved in dealing with split instances
977 *(Released Mon, 11 Jun 2012)*
979 Second beta release of 2.6. The following changes were done from beta2
982 - Fixed ``daemon-util`` with non-root user models
983 - Fixed creation of plain instances with ``--no-wait-for-sync``
984 - Fix wrong iv_names when running ``cfgupgrade``
985 - Export more information in RAPI group queries
986 - Fixed bug when changing instance network interfaces
987 - Extended burnin to do NIC changes
988 - query: Added ``<``, ``>``, ``<=``, ``>=`` comparison operators
989 - Changed default for DRBD barriers
990 - Fixed DRBD error reporting for syncer rate
991 - Verify the options on disk parameters
993 And of course various fixes to documentation and improved unittests and
1000 *(Released Wed, 23 May 2012)*
1002 First beta release of 2.6. The following changes were done from beta1 to
1005 - integrated patch for distributions without ``start-stop-daemon``
1006 - adapted example init.d script to work on Fedora
1007 - fixed log handling in Haskell daemons
1008 - adapted checks in the watcher for pycurl linked against libnss
1009 - add partial support for ``xl`` instead of ``xm`` for Xen
1010 - fixed a type issue in cluster verification
1011 - fixed ssconf handling in the Haskell code (was breaking confd in IPv6
1014 Plus integrated fixes from the 2.5 branch:
1016 - fixed ``kvm-ifup`` to use ``/bin/bash``
1017 - fixed parallel build failures
1018 - KVM live migration when using a custom keymap
1024 *(Released Tue, 24 Jul 2012)*
1026 A small bugfix release, with no new features:
1028 - fixed bash-isms in kvm-ifup, for compatibility with systems which use a
1029 different default shell (e.g. Debian, Ubuntu)
1030 - fixed KVM startup and live migration with a custom keymap (fixes Issue
1031 243 and Debian bug #650664)
1032 - fixed compatibility with KVM versions that don't support multiple boot
1033 devices (fixes Issue 230 and Debian bug #624256)
1035 Additionally, a few fixes were done to the build system (fixed parallel
1036 build failures) and to the unittests (fixed race condition in test for
1037 FileID functions, and the default enable/disable mode for QA test is now
1044 *(Released Fri, 11 May 2012)*
1046 A small bugfix release.
1048 The main issues solved are on the topic of compatibility with newer LVM
1051 - fixed parsing of ``lv_attr`` field
1052 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
1053 the ``--force`` flag is needed
1055 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
1056 to accept kernel 3.x too (was hardcoded to 2.6.*).
1058 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
1059 of hook results and that also made display of other errors suboptimal
1060 was fixed; the code behaves now like 2.4 and earlier.
1062 Another change in 2.5, the cleanup of the OS scripts environment, is too
1063 aggressive: it removed even the ``PATH`` variable, which requires the OS
1064 scripts to *always* need to export it. Since this is a bit too strict,
1065 we now export a minimal PATH, the same that we export for hooks.
1067 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
1068 integrated into this release.
1070 Finally, a few other miscellaneous changes were done (no new features,
1071 just small improvements):
1073 - Fix ``gnt-group --help`` display
1074 - Fix hardcoded Xen kernel path
1075 - Fix grow-disk handling of invalid units
1076 - Update synopsis for ``gnt-cluster repair-disk-sizes``
1077 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
1083 *(Released Thu, 12 Apr 2012)*
1085 Incompatible/important changes and bugfixes
1086 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1088 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
1089 resource's ``ip_check`` parameter changed from ``True`` to ``False``
1090 to match the underlying LUXI interface.
1091 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
1092 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
1093 not maintain backwards-compatibility as the underlying operation
1094 changed in an incompatible way. The RAPI client can talk to old
1095 servers, but it needs to be told so as the return value changed.
1096 - When creating file-based instances via RAPI, the ``file_driver``
1097 parameter no longer defaults to ``loop`` and must be specified.
1098 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
1100 - Support for the undocumented and deprecated RAPI instance creation
1101 request format version 0 has been dropped. Use version 1, supported
1102 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
1103 - Pyparsing 1.4.6 or above is required, see :doc:`installation
1104 documentation <install>`.
1105 - The "cluster-verify" hooks are now executed per group by the
1106 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
1107 if you just run ``gnt-cluster verify``, which generates one opcode per
1109 - The environment as passed to the OS scripts is cleared, and thus no
1110 environment variables defined in the node daemon's environment will be
1111 inherited by the scripts.
1112 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
1114 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
1115 support operations involving multiple node groups.
1116 - Offline nodes are ignored when failing over an instance.
1117 - Support for KVM version 1.0, which changed the version reporting format
1119 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
1121 - ``Makefile`` is now compatible with Automake 1.11.2
1122 - Includes all bugfixes made in the 2.4 series
1127 - The ganeti-htools project has been merged into the ganeti-core source
1128 tree and will be built as part of Ganeti (see :doc:`install-quick`).
1129 - Implemented support for :doc:`shared storage <design-shared-storage>`.
1130 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
1131 GPT-style partition tables (requires `parted
1132 <http://www.gnu.org/s/parted/>`_).
1133 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
1134 - Allowed adding tags on instance creation.
1135 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
1136 - Allow instances to be started in a paused state, enabling the user to
1137 see the complete console output on boot using the console.
1138 - Added new hypervisor flag to control default reboot behaviour
1139 (``reboot_behavior``).
1140 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
1141 - Improved out-of-band management support:
1143 - Added ``gnt-node health`` command reporting the health status of
1145 - Added ``gnt-node power`` command to manage power status of nodes.
1146 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
1148 - Instance migration can fall back to failover if instance is not
1150 - Filters can be used when listing nodes, instances, groups and locks;
1151 see :manpage:`ganeti(7)` manpage.
1152 - Added post-execution status as variables to :doc:`hooks <hooks>`
1154 - Instance tags are exported/imported together with the instance.
1155 - When given an explicit job ID, ``gnt-job info`` will work for archived
1157 - Jobs can define dependencies on other jobs (not yet supported via
1158 RAPI or command line, but used by internal commands and usable via
1161 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
1164 - Instance failover is now available as a RAPI resource
1165 (``/2/instances/[instance_name]/failover``).
1166 - ``gnt-instance info`` defaults to static information if primary node
1168 - Opcodes have a new ``comment`` attribute.
1169 - Added basic SPICE support to KVM hypervisor.
1170 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
1172 Node group improvements
1173 ~~~~~~~~~~~~~~~~~~~~~~~
1175 - ``gnt-cluster verify`` has been modified to check groups separately,
1176 thereby improving performance.
1177 - Node group support has been added to ``gnt-cluster verify-disks``,
1178 which now operates per node group.
1179 - Watcher has been changed to work better with node groups.
1181 - One process and state file per node group.
1182 - Slow watcher in one group doesn't block other group's watcher.
1184 - Added new command, ``gnt-group evacuate``, to move all instances in a
1185 node group to other groups.
1186 - Added ``gnt-instance change-group`` to move an instance to another
1188 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
1189 per-group operations.
1190 - Node groups can be tagged.
1191 - Some operations switch from an exclusive to a shared lock as soon as
1193 - Instance's primary and secondary nodes' groups are now available as
1194 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
1195 and ``snodes.group.uuid``).
1200 - Numerous updates to documentation and manpages.
1202 - :doc:`RAPI <rapi>` documentation now has detailed parameter
1204 - Some opcode/job results are now also documented, see :doc:`RAPI
1207 - A lockset's internal lock is now also visible in lock monitor.
1208 - Log messages from job queue workers now contain information about the
1209 opcode they're processing.
1210 - ``gnt-instance console`` no longer requires the instance lock.
1211 - A short delay when waiting for job changes reduces the number of LUXI
1212 requests significantly.
1213 - DRBD metadata volumes are overwritten with zeros during disk creation.
1214 - Out-of-band commands no longer acquire the cluster lock in exclusive
1216 - ``devel/upload`` now uses correct permissions for directories.
1222 *(Released Fri, 23 Mar 2012)*
1224 This was the sixth release candidate of the 2.5 series.
1230 *(Released Mon, 9 Jan 2012)*
1232 This was the fifth release candidate of the 2.5 series.
1238 *(Released Thu, 27 Oct 2011)*
1240 This was the fourth release candidate of the 2.5 series.
1246 *(Released Wed, 26 Oct 2011)*
1248 This was the third release candidate of the 2.5 series.
1254 *(Released Tue, 18 Oct 2011)*
1256 This was the second release candidate of the 2.5 series.
1262 *(Released Tue, 4 Oct 2011)*
1264 This was the first release candidate of the 2.5 series.
1270 *(Released Wed, 31 Aug 2011)*
1272 This was the third beta release of the 2.5 series.
1278 *(Released Mon, 22 Aug 2011)*
1280 This was the second beta release of the 2.5 series.
1286 *(Released Fri, 12 Aug 2011)*
1288 This was the first beta release of the 2.5 series.
1294 *(Released Thu, 27 Oct 2011)*
1296 - Fixed bug when parsing command line parameter values ending in
1298 - Fixed assertion error after unclean master shutdown
1299 - Disable HTTP client pool for RPC, significantly reducing memory usage
1301 - Fixed queue archive creation with wrong permissions
1307 *(Released Tue, 23 Aug 2011)*
1311 - Fixed documentation for importing with ``--src-dir`` option
1312 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
1313 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
1319 *(Released Fri, 5 Aug 2011)*
1321 Many bug-fixes and a few small features:
1323 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
1324 issues when you tried to add an instance with two MAC addresses in one
1326 - KVM: fixed per-instance stored UID value
1327 - KVM: configure bridged NICs at migration start
1328 - KVM: Fix a bug where instance will not start with never KVM versions
1330 - Added OS search path to ``gnt-cluster info``
1331 - Fixed an issue with ``file_storage_dir`` where you were forced to
1332 provide an absolute path, but the documentation states it is a
1333 relative path, the documentation was right
1334 - Added a new parameter to instance stop/start called ``--no-remember``
1335 that will make the state change to not be remembered
1336 - Implemented ``no_remember`` at RAPI level
1337 - Improved the documentation
1338 - Node evacuation: don't call IAllocator if node is already empty
1339 - Fixed bug in DRBD8 replace disks on current nodes
1340 - Fixed bug in recreate-disks for DRBD instances
1341 - Moved assertion checking locks in ``gnt-instance replace-disks``
1342 causing it to abort with not owning the right locks for some situation
1343 - Job queue: Fixed potential race condition when cancelling queued jobs
1344 - Fixed off-by-one bug in job serial generation
1345 - ``gnt-node volumes``: Fix instance names
1346 - Fixed aliases in bash completion
1347 - Fixed a bug in reopening log files after being sent a SIGHUP
1348 - Added a flag to burnin to allow specifying VCPU count
1349 - Bugfixes to non-root Ganeti configuration
1355 *(Released Thu, 12 May 2011)*
1357 Many bug-fixes and a few new small features:
1359 - Fixed a bug related to log opening failures
1360 - Fixed a bug in instance listing with orphan instances
1361 - Fixed a bug which prevented resetting the cluster-level node parameter
1362 ``oob_program`` to the default
1363 - Many fixes related to the ``cluster-merge`` tool
1364 - Fixed a race condition in the lock monitor, which caused failures
1365 during (at least) creation of many instances in parallel
1366 - Improved output for gnt-job info
1367 - Removed the quiet flag on some ssh calls which prevented debugging
1369 - Improved the N+1 failure messages in cluster verify by actually
1370 showing the memory values (needed and available)
1371 - Increased lock attempt timeouts so that when executing long operations
1372 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
1373 too early and thus prevent the use of the 'fair' mechanism
1374 - Changed instance query data (``gnt-instance info``) to not acquire
1375 locks unless needed, thus allowing its use on locked instance if only
1376 static information is asked for
1377 - Improved behaviour with filesystems that do not support rename on an
1379 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
1380 kept locks on all nodes during the wipe, which is unneeded
1381 - Fixed ``gnt-watcher`` handling of errors during hooks execution
1382 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
1383 10GiB) which caused the wipe to fail right at the end in some cases
1384 - Fixed master IP activation when doing master failover with no-voting
1385 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
1386 the master node itself
1387 - Fixed potential data-loss in under disk full conditions, where Ganeti
1388 wouldn't check correctly the return code and would consider
1389 partially-written files 'correct'
1390 - Fixed bug related to multiple VGs and DRBD disk replacing
1391 - Added new disk parameter ``metavg`` that allows placement of the meta
1392 device for DRBD in a different volume group
1393 - Fixed error handling in the node daemon when the system libc doesn't
1394 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
1395 - Fixed lock release during replace-disks, which kept cluster-wide locks
1396 when doing disk replaces with an iallocator script
1397 - Added check for missing bridges in cluster verify
1398 - Handle EPIPE errors while writing to the terminal better, so that
1399 piping the output to e.g. ``less`` doesn't cause a backtrace
1400 - Fixed rare case where a ^C during Luxi calls could have been
1401 interpreted as server errors, instead of simply terminating
1402 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
1404 - Added a few more parameters to the KVM hypervisor, allowing a second
1405 CDROM, custom disk type for CDROMs and a floppy image
1406 - Removed redundant message in instance rename when the name is given
1408 - Added option to ``gnt-instance recreate-disks`` to allow creating the
1409 disks on new nodes, allowing recreation when the original instance
1410 nodes are completely gone
1411 - Added option when converting disk templates to DRBD to skip waiting
1412 for the resync, in order to make the instance available sooner
1413 - Added two new variables to the OS scripts environment (containing the
1415 - Made the root_path and optional parameter for the xen-pvm hypervisor,
1416 to allow use of ``pvgrub`` as bootloader
1417 - Changed the instance memory modifications to only check out-of-memory
1418 conditions on memory increases, and turned the secondary node warnings
1419 into errors (they can still be overridden via ``--force``)
1420 - Fixed the handling of a corner case when the Python installation gets
1421 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
1422 to execute a command that doesn't exist
1423 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
1424 node of the instance returned failures during instance shutdown; this
1425 adds the option ``--ignore-consistency`` to gnt-instance move
1427 And as usual, various improvements to the error messages, documentation
1434 *(Released Wed, 09 Mar 2011)*
1436 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
1437 the RAPI users file if run twice (even with ``--dry-run``).
1439 The release fixes that bug (nothing else changed).
1445 *(Released Mon, 07 Mar 2011)*
1447 Final 2.4.0 release. Just a few small fixes:
1449 - Fixed RAPI node evacuate
1450 - Fixed the kvm-ifup script
1451 - Fixed internal error handling for special job cases
1452 - Updated man page to specify the escaping feature for options
1458 *(Released Mon, 28 Feb 2011)*
1460 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
1461 that this feature wiped the disks of the wrong instance, leading to loss
1466 - Fixed title of query field containing instance name
1467 - Expanded the glossary in the documentation
1468 - Fixed one unittest (internal issue)
1474 *(Released Mon, 21 Feb 2011)*
1476 A number of bug fixes plus just a couple functionality changes.
1478 On the user-visible side, the ``gnt-* list`` command output has changed
1479 with respect to "special" field states. The current rc1 style of display
1480 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
1481 the default output mode special fields states are displayed as follows:
1483 - Offline resource: ``*``
1484 - Unavailable/not applicable: ``-``
1485 - Data missing (RPC failure): ``?``
1486 - Unknown field: ``??``
1488 Another user-visible change is the addition of ``--force-join`` to
1493 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
1494 - Fixed regression in RAPI/instance reinstall where all parameters were
1495 required (instead of optional)
1496 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
1497 - Fixed iallocator usage (offline nodes were not considered offline)
1498 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
1499 - Fixed hypervisor and OS parameter validation with respect to
1500 non-vm_capable nodes
1501 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
1503 - Fixed ``tools/listrunner`` with respect to agent-based usage
1509 *(Released Fri, 4 Feb 2011)*
1511 Many changes and fixes since the beta1 release. While there were some
1512 internal changes, the code has been mostly stabilised for the RC
1515 Note: the dumb allocator was removed in this release, as it was not kept
1516 up-to-date with the IAllocator protocol changes. It is recommended to
1517 use the ``hail`` command from the ganeti-htools package.
1519 Note: the 2.4 and up versions of Ganeti are not compatible with the
1520 0.2.x branch of ganeti-htools. You need to upgrade to
1521 ganeti-htools-0.3.0 (or later).
1523 Regressions fixed from 2.3
1524 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1526 - Fixed the ``gnt-cluster verify-disks`` command
1527 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
1529 - Fixed disk adoption breakage
1530 - Fixed wrong headers in instance listing for field aliases
1535 - Fixed corner case in KVM handling of NICs
1536 - Fixed many cases of wrong handling of non-vm_capable nodes
1537 - Fixed a bug where a missing instance symlink was not possible to
1538 recreate with any ``gnt-*`` command (now ``gnt-instance
1539 activate-disks`` does it)
1540 - Fixed the volume group name as reported by ``gnt-cluster
1542 - Increased timeouts for the import-export code, hopefully leading to
1543 fewer aborts due network or instance timeouts
1544 - Fixed bug in ``gnt-node list-storage``
1545 - Fixed bug where not all daemons were started on cluster
1546 initialisation, but only at the first watcher run
1547 - Fixed many bugs in the OOB implementation
1548 - Fixed watcher behaviour in presence of instances with offline
1550 - Fixed instance list output for instances running on the wrong node
1551 - a few fixes to the cluster-merge tool, but it still cannot merge
1552 multi-node groups (currently it is not recommended to use this tool)
1558 - Improved network configuration for the KVM hypervisor
1559 - Added e1000 as a supported NIC for Xen-HVM
1560 - Improved the lvmstrap tool to also be able to use partitions, as
1561 opposed to full disks
1562 - Improved speed of disk wiping (the cluster parameter
1563 ``prealloc_wipe_disks``, so that it has a low impact on the total time
1564 of instance creations
1565 - Added documentation for the OS parameters
1566 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
1567 hypervisor is not responding
1568 - Added display of blacklisted and hidden OS information in
1569 ``gnt-cluster info``
1570 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
1571 NIC and node parameters, which might create problems with currently
1572 invalid (but undetected) configuration files, but prevents validation
1573 failures when unrelated parameters are modified
1574 - Changed cluster initialisation to wait for the master daemon to become
1576 - Expanded the RAPI interface:
1578 - Added config redistribution resource
1579 - Added activation/deactivation of instance disks
1580 - Added export of console information
1582 - Implemented log file reopening on SIGHUP, which allows using
1583 logrotate(8) for the Ganeti log files
1584 - Added a basic OOB helper script as an example
1590 *(Released Fri, 14 Jan 2011)*
1595 - Fixed timezone issues when formatting timestamps
1596 - Added support for node groups, available via ``gnt-group`` and other
1598 - Added out-of-band framework and management, see :doc:`design
1599 document <design-oob>`
1600 - Removed support for roman numbers from ``gnt-node list`` and
1601 ``gnt-instance list``.
1602 - Allowed modification of master network interface via ``gnt-cluster
1603 modify --master-netdev``
1604 - Accept offline secondaries while shutting down instance disks
1605 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
1606 - Added support for multiple LVM volume groups
1607 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
1609 - Added commands to list available fields:
1611 - ``gnt-node list-fields``
1612 - ``gnt-group list-fields``
1613 - ``gnt-instance list-fields``
1615 - Updated documentation and man pages
1620 - Moved ``rapi_users`` file into separate directory, now named
1621 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
1623 - Added new tool for running commands on many machines,
1624 ``tools/ganeti-listrunner``
1625 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
1626 improving the ``gnt-instance console`` output
1627 - Allowed customisation of disk index separator at ``configure`` time
1628 - Export node group allocation policy to :doc:`iallocator <iallocator>`
1629 - Added support for non-partitioned md disks in ``lvmstrap``
1630 - Added script to gracefully power off KVM instances
1631 - Split ``utils`` module into smaller parts
1632 - Changed query operations to return more detailed information, e.g.
1633 whether an information is unavailable due to an offline node. To use
1634 this new functionality, the LUXI call ``Query`` must be used. Field
1635 information is now stored by the master daemon and can be retrieved
1636 using ``QueryFields``. Instances, nodes and groups can also be queried
1637 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
1638 exposed via RAPI). The following commands make use of this
1639 infrastructure change:
1641 - ``gnt-group list``
1642 - ``gnt-group list-fields``
1644 - ``gnt-node list-fields``
1645 - ``gnt-instance list``
1646 - ``gnt-instance list-fields``
1647 - ``gnt-debug locks``
1652 - New RAPI resources (see :doc:`rapi`):
1656 - ``/2/groups/[group_name]``
1657 - ``/2/groups/[group_name]/assign-nodes``
1658 - ``/2/groups/[group_name]/modify``
1659 - ``/2/groups/[group_name]/rename``
1660 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
1664 - Implemented ``no_install`` for instance creation
1665 - Implemented OS parameters for instance reinstallation, allowing
1666 use of special settings on reinstallation (e.g. for preserving data)
1671 - Added IPv6 support in import/export
1672 - Pause DRBD synchronization while wiping disks on instance creation
1673 - Updated unittests and QA scripts
1674 - Improved network parameters passed to KVM
1675 - Converted man pages from docbook to reStructuredText
1681 *(Released Mon, 20 Dec 2010)*
1683 Released version 2.3.1~rc1 without any changes.
1689 *(Released Wed, 1 Dec 2010)*
1691 - impexpd: Disable OpenSSL compression in socat if possible (backport
1692 from master, commit e90739d625b, see :doc:`installation guide
1693 <install-quick>` for details)
1694 - Changed unittest coverage report to exclude test scripts
1695 - Added script to check version format
1701 *(Released Wed, 1 Dec 2010)*
1703 Released version 2.3.0~rc1 without any changes.
1709 *(Released Fri, 19 Nov 2010)*
1711 A number of bugfixes and documentation updates:
1713 - Update ganeti-os-interface documentation
1714 - Fixed a bug related to duplicate MACs or similar items which should be
1716 - Fix breakage in OS state modify
1717 - Reinstall instance: disallow offline secondaries (fixes bug related to
1718 OS changing but reinstall failing)
1719 - plus all the other fixes between 2.2.1 and 2.2.2
1725 *(Released Tue, 2 Nov 2010)*
1727 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
1728 - Fixed master failover race with watcher
1729 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
1731 - Accept previously stopped instance for export with instance removal
1732 - Simplify and extend the environment variables for instance OS scripts
1733 - Added new node flags, ``master_capable`` and ``vm_capable``
1734 - Added optional instance disk wiping prior during allocation. This is a
1735 cluster-wide option and can be set/modified using
1736 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
1737 - Added IPv6 support, see :doc:`design document <design-2.3>` and
1738 :doc:`install-quick`
1739 - Added a new watcher option (``--ignore-pause``)
1740 - Added option to ignore offline node on instance start/stop
1741 (``--ignore-offline``)
1742 - Allow overriding OS parameters with ``gnt-instance reinstall``
1743 - Added ability to change node's secondary IP address using ``gnt-node
1745 - Implemented privilege separation for all daemons except
1746 ``ganeti-noded``, see ``configure`` options
1747 - Complain if an instance's disk is marked faulty in ``gnt-cluster
1749 - Implemented job priorities (see ``ganeti(7)`` manpage)
1750 - Ignore failures while shutting down instances during failover from
1752 - Exit daemon's bootstrap process only once daemon is ready
1753 - Export more information via ``LUInstanceQuery``/remote API
1754 - Improved documentation, QA and unittests
1755 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
1756 restart if the file was created or changed
1757 - Added LUXI protocol version sent with each request and response,
1758 allowing detection of server/client mismatches
1759 - Moved the Python scripts among gnt-* and ganeti-* into modules
1760 - Moved all code related to setting up SSH to an external script,
1762 - Infrastructure changes for node group support in future versions
1768 *(Released Fri, 19 Nov 2010)*
1770 A few small bugs fixed, and some improvements to the build system:
1772 - Fix documentation regarding conversion to drbd
1773 - Fix validation of parameters in cluster modify (``gnt-cluster modify
1775 - Fix error handling in node modify with multiple changes
1776 - Allow remote imports without checked names
1782 *(Released Tue, 19 Oct 2010)*
1784 - Disable SSL session ID cache in RPC client
1790 *(Released Thu, 14 Oct 2010)*
1792 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
1793 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
1795 - Fix problems with interaction between Curl and Python's HTTP server,
1796 resulting in increased speed in many RPC calls
1797 - Improve our release script to prevent breakage with older aclocal and
1804 *(Released Thu, 7 Oct 2010)*
1806 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
1807 value retrieved from master
1808 - Added support for blacklisted or hidden OS definitions
1809 - Added simple lock monitor (accessible via (``gnt-debug locks``)
1810 - Added support for -mem-path in KVM hypervisor abstraction layer
1811 - Allow overriding instance parameters in tool for inter-cluster
1812 instance moves (``tools/move-instance``)
1813 - Improved opcode summaries (e.g. in ``gnt-job list``)
1814 - Improve consistency of OS listing by sorting it
1815 - Documentation updates
1821 *(Released Fri, 8 Oct 2010)*
1823 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
1829 *(Released Mon, 4 Oct 2010)*
1831 - Fixed regression in ``gnt-instance rename``
1837 *(Released Wed, 22 Sep 2010)*
1839 - Fixed OS_VARIANT variable for OS scripts
1840 - Fixed cluster tag operations via RAPI
1841 - Made ``setup-ssh`` exit with non-zero code if an error occurred
1842 - Disabled RAPI CA checks in watcher
1848 *(Released Mon, 23 Aug 2010)*
1850 - Support DRBD versions of the format "a.b.c.d"
1852 - Re-introduce support for usage from multiple threads in RAPI client
1853 - Instance renames and modify via RAPI
1854 - Work around race condition between processing and archival in job
1856 - Mark opcodes following failed one as failed, too
1857 - Job field ``lock_status`` was removed due to difficulties making it
1858 work with the changed job queue in Ganeti 2.2; a better way to monitor
1859 locks is expected for a later 2.2.x release
1860 - Fixed dry-run behaviour with many commands
1861 - Support ``ssh-agent`` again when adding nodes
1862 - Many additional bugfixes
1868 *(Released Fri, 30 Jul 2010)*
1870 Important change: the internal RPC mechanism between Ganeti nodes has
1871 changed from using a home-grown http library (based on the Python base
1872 libraries) to use the PycURL library. This requires that PycURL is
1873 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
1874 against GnuTLS by default. cURL's support for GnuTLS had known issues
1875 before cURL 7.21.0 and we recommend using the latest cURL release or
1876 linking against OpenSSL. Most other distributions already link PycURL
1877 and cURL against OpenSSL. The command::
1879 python -c 'import pycurl; print pycurl.version'
1881 can be used to determine the libraries PycURL and cURL are linked
1884 Other significant changes:
1886 - Rewrote much of the internals of the job queue, in order to achieve
1887 better parallelism; this decouples job query operations from the job
1888 processing, and it should allow much nicer behaviour of the master
1889 daemon under load, and it also has uncovered some long-standing bugs
1890 related to the job serialisation (now fixed)
1891 - Added a default iallocator setting to the cluster parameters,
1892 eliminating the need to always pass nodes or an iallocator for
1893 operations that require selection of new node(s)
1894 - Added experimental support for the LXC virtualization method
1895 - Added support for OS parameters, which allows the installation of
1896 instances to pass parameter to OS scripts in order to customise the
1898 - Added a hypervisor parameter controlling the migration type (live or
1899 non-live), since hypervisors have various levels of reliability; this
1900 has renamed the 'live' parameter to 'mode'
1901 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
1902 logical volumes, meaning that cluster verify will ignore them and not
1903 flag their presence as errors
1904 - The watcher will now reset the error count for failed instances after
1905 8 hours, thus allowing self-healing if the problem that caused the
1906 instances to be down/fail to start has cleared in the meantime
1907 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
1908 check for, and warn, if the drbd module parameter ``usermode_helper``
1909 is not consistent with the cluster-wide setting; this is needed to
1910 make diagnose easier of failed drbd creations
1911 - Started adding base IPv6 support, but this is not yet
1912 enabled/available for use
1913 - Rename operations (cluster, instance) will now return the new name,
1914 which is especially useful if a short name was passed in
1915 - Added support for instance migration in RAPI
1916 - Added a tool to pre-configure nodes for the SSH setup, before joining
1917 them to the cluster; this will allow in the future a simplified model
1918 for node joining (but not yet fully enabled in 2.2); this needs the
1919 paramiko python library
1920 - Fixed handling of name-resolving errors
1921 - Fixed consistency of job results on the error path
1922 - Fixed master-failover race condition when executed multiple times in
1924 - Fixed many bugs related to the job queue (mostly introduced during the
1925 2.2 development cycle, so not all are impacting 2.1)
1926 - Fixed instance migration with missing disk symlinks
1927 - Fixed handling of unknown jobs in ``gnt-job archive``
1928 - And many other small fixes/improvements
1932 - Enhanced both the unittest and the QA coverage
1933 - Switched the opcode validation to a generic model, and extended the
1934 validation to all opcode parameters
1935 - Changed more parts of the code that write shell scripts to use the
1937 - Switched the master daemon to use the asyncore library for the Luxi
1944 *(Released Thu, 17 Jun 2010)*
1946 - Added tool (``move-instance``) and infrastructure to move instances
1947 between separate clusters (see :doc:`separate documentation
1948 <move-instance>` and :doc:`design document <design-2.2>`)
1949 - Added per-request RPC timeout
1950 - RAPI now requires a Content-Type header for requests with a body (e.g.
1951 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
1952 :rfc:`2616` (HTTP/1.1), section 7.2.1)
1953 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
1955 - Implemented initial support for running Ganeti daemons as separate
1956 users, see configure-time flags ``--with-user-prefix`` and
1957 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
1959 - Instances can be removed after export (``gnt-backup export
1960 --remove-instance``)
1961 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
1962 key (instead of 1024 bit)
1963 - Added new cluster configuration file for cluster domain secret
1964 - Import/export now use SSL instead of SSH
1965 - Added support for showing estimated time when exporting an instance,
1966 see the ``ganeti-os-interface(7)`` manpage and look for
1973 *(Released Tue, 16 Nov 2010)*
1975 Some more bugfixes. Unless critical bugs occur, this will be the last
1978 - Fix case of MAC special-values
1979 - Fix mac checker regex
1980 - backend: Fix typo causing "out of range" error
1981 - Add missing --units in gnt-instance list man page
1987 *(Released Tue, 24 Aug 2010)*
1990 - Don't ignore secondary node silently on non-mirrored disk templates
1992 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
1993 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
1994 - Properly document the usb_mouse parameter
1995 - Fix path in ganeti-rapi(8) (issue 116)
1996 - Adjust error message when the ganeti user's .ssh directory is
1998 - Add same-node-check when changing the disk template to drbd
2004 *(Released Fri, 16 Jul 2010)*
2007 - Add an option to only select some reboot types during qa/burnin.
2008 (on some hypervisors consequent reboots are not supported)
2009 - Fix infrequent race condition in master failover. Sometimes the old
2010 master ip address would be still detected as up for a short time
2011 after it was removed, causing failover to fail.
2012 - Decrease mlockall warnings when the ctypes module is missing. On
2013 Python 2.4 we support running even if no ctypes module is installed,
2014 but we were too verbose about this issue.
2015 - Fix building on old distributions, on which man doesn't have a
2017 - Fix RAPI not to ignore the MAC address on instance creation
2018 - Implement the old instance creation format in the RAPI client.
2024 *(Released Thu, 01 Jul 2010)*
2026 A small bugfix release:
2027 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
2028 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
2029 a non-existing option
2030 - Fix instance create: the --force-variant option was ignored
2031 - Improve pylint 0.21 compatibility and warnings with Python 2.6
2032 - Fix modify node storage with non-FQDN arguments
2033 - Fix RAPI client to authenticate under Python 2.6 when used
2034 for more than 5 requests needing authentication
2035 - Fix gnt-instance modify -t (storage) giving a wrong error message
2036 when converting a non-shutdown drbd instance to plain
2042 *(Released Fri, 18 Jun 2010)*
2044 A small bugfix release:
2046 - Fix live migration of KVM instances started with older Ganeti
2047 versions which had fewer hypervisor parameters
2048 - Fix gnt-instance grow-disk on down instances
2049 - Fix an error-reporting bug during instance migration
2050 - Better checking of the ``--net`` and ``--disk`` values, to avoid
2051 silently ignoring broken ones
2052 - Fix an RPC error reporting bug affecting, for example, RAPI client
2054 - Fix bug triggered by different API version os-es on different nodes
2055 - Fix a bug in instance startup with custom hvparams: OS level
2056 parameters would fail to be applied.
2057 - Fix the RAPI client under Python 2.6 (but more work is needed to
2058 make it work completely well with OpenSSL)
2059 - Fix handling of errors when resolving names from DNS
2065 *(Released Thu, 3 Jun 2010)*
2067 A medium sized development cycle. Some new features, and some
2068 fixes/small improvements/cleanups.
2070 Significant features
2071 ~~~~~~~~~~~~~~~~~~~~
2073 The node deamon now tries to mlock itself into memory, unless the
2074 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
2075 its logs, and falls back to console logging. This allows emergency
2076 features such as ``gnt-node powercycle`` to work even in the event of a
2077 broken node disk (tested offlining the disk hosting the node's
2078 filesystem and dropping its memory caches; don't try this at home)
2080 KVM: add vhost-net acceleration support. It can be tested with a new
2081 enough version of the kernel and of qemu-kvm.
2083 KVM: Add instance chrooting feature. If you use privilege dropping for
2084 your VMs you can also now force them to chroot to an empty directory,
2085 before starting the emulated guest.
2087 KVM: Add maximum migration bandwith and maximum downtime tweaking
2088 support (requires a new-enough version of qemu-kvm).
2090 Cluster verify will now warn if the master node doesn't have the master
2091 ip configured on it.
2093 Add a new (incompatible) instance creation request format to RAPI which
2094 supports all parameters (previously only a subset was supported, and it
2095 wasn't possible to extend the old format to accomodate all the new
2096 features. The old format is still supported, and a client can check for
2097 this feature, before using it, by checking for its presence in the
2098 ``features`` RAPI resource.
2100 Now with ancient latin support. Try it passing the ``--roman`` option to
2101 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
2102 (requires the python-roman module to be installed, in order to work).
2107 As usual many internal code refactorings, documentation updates, and
2110 - Lots of improvements and cleanups to the experimental Remote API
2111 (RAPI) client library.
2112 - A new unit test suite for the core daemon libraries.
2113 - A fix to creating missing directories makes sure the umask is not
2114 applied anymore. This enforces the same directory permissions
2116 - Better handling terminating daemons with ctrl+c (used when running
2117 them in debugging mode).
2118 - Fix a race condition in live migrating a KVM instance, when stat()
2119 on the old proc status file returned EINVAL, which is an unexpected
2121 - Fixed manpage checking with newer man and utf-8 charachters. But now
2122 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
2128 *(Released Fri, 7 May 2010)*
2130 Fix a bug which prevented untagged KVM instances from starting.
2136 *(Released Fri, 7 May 2010)*
2138 Another release with a long development cycle, during which many
2139 different features were added.
2141 Significant features
2142 ~~~~~~~~~~~~~~~~~~~~
2144 The KVM hypervisor now can run the individual instances as non-root, to
2145 reduce the impact of a VM being hijacked due to bugs in the
2146 hypervisor. It is possible to run all instances as a single (non-root)
2147 user, to manually specify a user for each instance, or to dynamically
2148 allocate a user out of a cluster-wide pool to each instance, with the
2149 guarantee that no two instances will run under the same user ID on any
2152 An experimental RAPI client library, that can be used standalone
2153 (without the other Ganeti libraries), is provided in the source tree as
2154 ``lib/rapi/client.py``. Note this client might change its interface in
2155 the future, as we iterate on its capabilities.
2157 A new command, ``gnt-cluster renew-crypto`` has been added to easily
2158 replace the cluster's certificates and crypto keys. This might help in
2159 case they have been compromised, or have simply expired.
2161 A new disk option for instance creation has been added that allows one
2162 to "adopt" currently existing logical volumes, with data
2163 preservation. This should allow easier migration to Ganeti from
2164 unmanaged (or managed via other software) instances.
2166 Another disk improvement is the possibility to convert between redundant
2167 (DRBD) and plain (LVM) disk configuration for an instance. This should
2168 allow better scalability (starting with one node and growing the
2169 cluster, or shrinking a two-node cluster to one node).
2171 A new feature that could help with automated node failovers has been
2172 implemented: if a node sees itself as offline (by querying the master
2173 candidates), it will try to shutdown (hard) all instances and any active
2174 DRBD devices. This reduces the risk of duplicate instances if an
2175 external script automatically failovers the instances on such nodes. To
2176 enable this, the cluster parameter ``maintain_node_health`` should be
2177 enabled; in the future this option (per the name) will enable other
2178 automatic maintenance features.
2180 Instance export/import now will reuse the original instance
2181 specifications for all parameters; that means exporting an instance,
2182 deleting it and the importing it back should give an almost identical
2183 instance. Note that the default import behaviour has changed from
2184 before, where it created only one NIC; now it recreates the original
2187 Cluster verify has added a few new checks: SSL certificates validity,
2188 /etc/hosts consistency across the cluster, etc.
2193 As usual, many internal changes were done, documentation fixes,
2196 - Fixed cluster initialization with disabled cluster storage (regression
2197 introduced in 2.1.1)
2198 - File-based storage supports growing the disks
2199 - Fixed behaviour of node role changes
2200 - Fixed cluster verify for some corner cases, plus a general rewrite of
2201 cluster verify to allow future extension with more checks
2202 - Fixed log spamming by watcher and node daemon (regression introduced
2204 - Fixed possible validation issues when changing the list of enabled
2206 - Fixed cleanup of /etc/hosts during node removal
2207 - Fixed RAPI response for invalid methods
2208 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
2209 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
2210 from ide disks, etc.)
2211 - Allow OS changes without re-installation (to record a changed OS
2212 outside of Ganeti, or to allow OS renames)
2213 - Allow instance creation without OS installation (useful for example if
2214 the OS will be installed manually, or restored from a backup not in
2216 - Implemented option to make cluster ``copyfile`` use the replication
2218 - Added list of enabled hypervisors to ssconf (possibly useful for
2220 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
2222 - A partial form of node re-IP is possible via node readd, which now
2223 allows changed node primary IP
2224 - Command line utilities now show an informational message if the job is
2226 - The logs of the master daemon now show the PID/UID/GID of the
2233 *(Released Fri, 12 Mar 2010)*
2235 During the 2.1.0 long release candidate cycle, a lot of improvements and
2236 changes have accumulated with were released later as 2.1.1.
2241 The node evacuate command (``gnt-node evacuate``) was significantly
2242 rewritten, and as such the IAllocator protocol was changed - a new
2243 request type has been added. This unfortunate change during a stable
2244 series is designed to improve performance of node evacuations; on
2245 clusters with more than about five nodes and which are well-balanced,
2246 evacuation should proceed in parallel for all instances of the node
2247 being evacuated. As such, any existing IAllocator scripts need to be
2248 updated, otherwise the above command will fail due to the unknown
2249 request. The provided "dumb" allocator has not been updated; but the
2250 ganeti-htools package supports the new protocol since version 0.2.4.
2252 Another important change is increased validation of node and instance
2253 names. This might create problems in special cases, if invalid host
2254 names are being used.
2256 Also, a new layer of hypervisor parameters has been added, that sits at
2257 OS level between the cluster defaults and the instance ones. This allows
2258 customisation of virtualization parameters depending on the installed
2259 OS. For example instances with OS 'X' may have a different KVM kernel
2260 (or any other parameter) than the cluster defaults. This is intended to
2261 help managing a multiple OSes on the same cluster, without manual
2262 modification of each instance's parameters.
2264 A tool for merging clusters, ``cluster-merge``, has been added in the
2265 tools sub-directory.
2270 - Improved the int/float conversions that should make the code more
2271 robust in face of errors from the node daemons
2272 - Fixed the remove node code in case of internal configuration errors
2273 - Fixed the node daemon behaviour in face of inconsistent queue
2274 directory (e.g. read-only file-system where we can't open the files
2276 - Fixed the behaviour of gnt-node modify for master candidate demotion;
2277 now it either aborts cleanly or, if given the new "auto_promote"
2278 parameter, will automatically promote other nodes as needed
2279 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
2280 completely prevent Ganeti from working
2281 - Fixed bug for instance export when not all disks were successfully
2283 - Fixed behaviour of node add when the new node is slow in starting up
2285 - Fixed handling of signals in the LUXI client, which should improve
2286 behaviour of command-line scripts
2287 - Added checks for invalid node/instance names in the configuration (now
2288 flagged during cluster verify)
2289 - Fixed watcher behaviour for disk activation errors
2290 - Fixed two potentially endless loops in http library, which led to the
2291 RAPI daemon hanging and consuming 100% CPU in some cases
2292 - Fixed bug in RAPI daemon related to hashed passwords
2293 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
2295 - Enhanced compatibility with non-Debian OSes, but not using absolute
2296 path in some commands and allowing customisation of the ssh
2297 configuration directory
2298 - Fixed possible future issue with new Python versions by abiding to the
2299 proper use of ``__slots__`` attribute on classes
2300 - Added checks that should prevent directory traversal attacks
2301 - Many documentation fixes based on feedback from users
2306 - Added an "early_release" more for instance replace disks and node
2307 evacuate, where we release locks earlier and thus allow higher
2308 parallelism within the cluster
2309 - Added watcher hooks, intended to allow the watcher to restart other
2310 daemons (e.g. from the ganeti-nbma project), but they can be used of
2311 course for any other purpose
2312 - Added a compile-time disable for DRBD barriers, to increase
2313 performance if the administrator trusts the power supply or the
2314 storage system to not lose writes
2315 - Added the option of using syslog for logging instead of, or in
2316 addition to, Ganeti's own log files
2317 - Removed boot restriction for paravirtual NICs for KVM, recent versions
2318 can indeed boot from a paravirtual NIC
2319 - Added a generic debug level for many operations; while this is not
2320 used widely yet, it allows one to pass the debug value all the way to
2322 - Enhanced the hooks environment for instance moves (failovers,
2323 migrations) where the primary/secondary nodes changed during the
2324 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
2325 - Enhanced data validations for many user-supplied values; one important
2326 item is the restrictions imposed on instance and node names, which
2327 might reject some (invalid) host names
2328 - Add a configure-time option to disable file-based storage, if it's not
2329 needed; this allows greater security separation between the master
2330 node and the other nodes from the point of view of the inter-node RPC
2332 - Added user notification in interactive tools if job is waiting in the
2333 job queue or trying to acquire locks
2334 - Added log messages when a job is waiting for locks
2335 - Added filtering by node tags in instance operations which admit
2336 multiple instances (start, stop, reboot, reinstall)
2337 - Added a new tool for cluster mergers, ``cluster-merge``
2338 - Parameters from command line which are of the form ``a=b,c=d`` can now
2339 use backslash escapes to pass in values which contain commas,
2340 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
2342 - For KVM, the instance name is the first parameter passed to KVM, so
2343 that it's more visible in the process list
2349 *(Released Tue, 2 Mar 2010)*
2351 Ganeti 2.1 brings many improvements with it. Major changes:
2353 - Added infrastructure to ease automated disk repairs
2354 - Added new daemon to export configuration data in a cheaper way than
2355 using the remote API
2356 - Instance NICs can now be routed instead of being associated with a
2358 - Improved job locking logic to reduce impact of jobs acquiring multiple
2359 locks waiting for other long-running jobs
2361 In-depth implementation details can be found in the Ganeti 2.1 design
2367 - Added chroot hypervisor
2368 - Added more options to xen-hvm hypervisor (``kernel_path`` and
2370 - Added more options to xen-pvm hypervisor (``use_bootloader``,
2371 ``bootloader_path`` and ``bootloader_args``)
2372 - Added the ``use_localtime`` option for the xen-hvm and kvm
2373 hypervisors, and the default value for this has changed to false (in
2374 2.0 xen-hvm always enabled it)
2375 - Added luxi call to submit multiple jobs in one go
2376 - Added cluster initialization option to not modify ``/etc/hosts``
2378 - Added network interface parameters
2379 - Added dry run mode to some LUs
2380 - Added RAPI resources:
2382 - ``/2/instances/[instance_name]/info``
2383 - ``/2/instances/[instance_name]/replace-disks``
2384 - ``/2/nodes/[node_name]/evacuate``
2385 - ``/2/nodes/[node_name]/migrate``
2386 - ``/2/nodes/[node_name]/role``
2387 - ``/2/nodes/[node_name]/storage``
2388 - ``/2/nodes/[node_name]/storage/modify``
2389 - ``/2/nodes/[node_name]/storage/repair``
2391 - Added OpCodes to evacuate or migrate all instances on a node
2392 - Added new command to list storage elements on nodes (``gnt-node
2393 list-storage``) and modify them (``gnt-node modify-storage``)
2394 - Added new ssconf files with master candidate IP address
2395 (``ssconf_master_candidates_ips``), node primary IP address
2396 (``ssconf_node_primary_ips``) and node secondary IP address
2397 (``ssconf_node_secondary_ips``)
2398 - Added ``ganeti-confd`` and a client library to query the Ganeti
2399 configuration via UDP
2400 - Added ability to run hooks after cluster initialization and before
2402 - Added automatic mode for disk replace (``gnt-instance replace-disks
2404 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
2405 after catastrophic data-loss
2406 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
2408 - Added ``gnt-instance move`` command to move instances
2409 - Added ``gnt-cluster watcher`` command to control watcher
2410 - Added ``gnt-node powercycle`` command to powercycle nodes
2411 - Added new job status field ``lock_status``
2412 - Added parseable error codes to cluster verification (``gnt-cluster
2413 verify --error-codes``) and made output less verbose (use
2414 ``--verbose`` to restore previous behaviour)
2415 - Added UUIDs to the main config entities (cluster, nodes, instances)
2416 - Added support for OS variants
2417 - Added support for hashed passwords in the Ganeti remote API users file
2419 - Added option to specify maximum timeout on instance shutdown
2420 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
2421 - Added new helper script to start and stop Ganeti daemons
2422 (``daemon-util``), with the intent to reduce the work necessary to
2423 adjust Ganeti for non-Debian distributions and to start/stop daemons
2425 - Added more unittests
2426 - Fixed critical bug in ganeti-masterd startup
2427 - Removed the configure-time ``kvm-migration-port`` parameter, this is
2428 now customisable at the cluster level for both the KVM and Xen
2429 hypervisors using the new ``migration_port`` parameter
2430 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
2431 reinstalling an instance
2432 - Allowed ``@`` in tag names
2433 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
2434 - Many documentation updates
2435 - Distribute hypervisor files on ``gnt-cluster redist-conf``
2436 - ``gnt-instance reinstall`` can now reinstall multiple instances
2437 - Updated many command line parameters
2438 - Introduced new OS API version 15
2439 - No longer support a default hypervisor
2440 - Treat virtual LVs as inexistent
2441 - Improved job locking logic to reduce lock contention
2442 - Match instance and node names case insensitively
2443 - Reimplemented bash completion script to be more complete
2450 *(Released Thu, 4 Feb 2010)*
2452 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
2453 - Fix a string formatting bug
2454 - Improve safety of the code in some error paths
2455 - Improve data validation in the master of values returned from nodes
2461 *(Released Thu, 17 Dec 2009)*
2463 - Fix security issue due to missing validation of iallocator names; this
2464 allows local and remote execution of arbitrary executables
2465 - Fix failure of gnt-node list during instance removal
2466 - Ship the RAPI documentation in the archive
2472 *(Released Wed, 30 Sep 2009)*
2474 - Fixed many wrong messages
2475 - Fixed a few bugs related to the locking library
2476 - Fixed MAC checking at instance creation time
2477 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
2478 - Fixed a few issues related to signal handling in both daemons and
2480 - Fixed the example startup script provided
2481 - Fixed insserv dependencies in the example startup script (patch from
2483 - Fixed handling of drained nodes in the iallocator framework
2484 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
2486 - Fixed error related to invalid job IDs in job polling
2487 - Fixed job/opcode persistence on unclean master shutdown
2488 - Fixed handling of partial job processing after unclean master
2490 - Fixed error reporting from LUs, previously all errors were converted
2491 into execution errors
2492 - Fixed error reporting from burnin
2493 - Decreased significantly the memory usage of the job queue
2494 - Optimised slightly multi-job submission
2495 - Optimised slightly opcode loading
2496 - Backported the multi-job submit framework from the development
2497 branch; multi-instance start and stop should be faster
2498 - Added script to clean archived jobs after 21 days; this will reduce
2499 the size of the queue directory
2500 - Added some extra checks in disk size tracking
2501 - Added an example ethers hook script
2502 - Added a cluster parameter that prevents Ganeti from modifying of
2504 - Added more node information to RAPI responses
2505 - Added a ``gnt-job watch`` command that allows following the ouput of a
2507 - Added a bind-address option to ganeti-rapi
2508 - Added more checks to the configuration verify
2509 - Enhanced the burnin script such that some operations can be retried
2511 - Converted instance reinstall to multi-instance model
2517 *(Released Fri, 7 Aug 2009)*
2519 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
2520 to allow using the pre-2.0.2 behaviour in activation, if any existing
2521 instances have mismatched disk sizes in the configuration
2522 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
2523 any configuration mismatches for disk sizes
2524 - Added ``gnt-master cluste-failover --no-voting`` to allow master
2525 failover to work on two-node clusters
2526 - Fixed the ``--net`` option of ``gnt-backup import``, which was
2528 - Fixed detection of OS script errors in ``gnt-backup export``
2529 - Fixed exit code of ``gnt-backup export``
2535 *(Released Fri, 17 Jul 2009)*
2537 - Added experimental support for stripped logical volumes; this should
2538 enhance performance but comes with a higher complexity in the block
2539 device handling; stripping is only enabled when passing
2540 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
2541 affected even in the non-stripped mode
2542 - Improved resiliency against transient failures at the end of DRBD
2543 resyncs, and in general of DRBD resync checks
2544 - Fixed a couple of issues with exports and snapshot errors
2545 - Fixed a couple of issues in instance listing
2546 - Added display of the disk size in ``gnt-instance info``
2547 - Fixed checking for valid OSes in instance creation
2548 - Fixed handling of the "vcpus" parameter in instance listing and in
2549 general of invalid parameters
2550 - Fixed http server library, and thus RAPI, to handle invalid
2551 username/password combinations correctly; this means that now they
2552 report unauthorized for queries too, not only for modifications,
2553 allowing earlier detect of configuration problems
2554 - Added a new "role" node list field, equivalent to the master/master
2555 candidate/drained/offline flags combinations
2556 - Fixed cluster modify and changes of candidate pool size
2557 - Fixed cluster verify error messages for wrong files on regular nodes
2558 - Fixed a couple of issues with node demotion from master candidate role
2559 - Fixed node readd issues
2560 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
2561 - Added a new ``--no-voting`` option for masterfailover to fix failover
2562 on two-nodes clusters when the former master node is unreachable
2563 - Added instance reinstall over RAPI
2569 *(Released Tue, 16 Jun 2009)*
2571 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
2572 allow re-adding the start in single-user option (regression from 1.2)
2573 - the watcher writes the instance status to a file, to allow monitoring
2574 to report the instance status (from the master) based on cached
2575 results of the watcher's queries; while this can get stale if the
2576 watcher is being locked due to other work on the cluster, this is
2577 still an improvement
2578 - the watcher now also restarts the node daemon and the rapi daemon if
2580 - fixed the watcher to handle full and drained queue cases
2581 - hooks export more instance data in the environment, which helps if
2582 hook scripts need to take action based on the instance's properties
2583 (no longer need to query back into ganeti)
2584 - instance failovers when the instance is stopped do not check for free
2585 RAM, so that failing over a stopped instance is possible in low memory
2587 - rapi uses queries for tags instead of jobs (for less job traffic), and
2588 for cluster tags it won't talk to masterd at all but read them from
2590 - a couple of error handling fixes in RAPI
2591 - drbd handling: improved the error handling of inconsistent disks after
2592 resync to reduce the frequency of "there are some degraded disks for
2593 this instance" messages
2594 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
2595 error handling path called a wrong function name)
2601 *(Released Wed, 27 May 2009)*
2603 - no changes from rc5
2609 *(Released Wed, 20 May 2009)*
2611 - fix a couple of bugs (validation, argument checks)
2612 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
2613 - some small improvements to RAPI and IAllocator
2614 - make watcher automatically start the master daemon if down
2620 *(Released Mon, 27 Apr 2009)*
2622 - change the OS list to not require locks; this helps with big clusters
2623 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
2624 volume group is broken
2625 - ``gnt-instance info``, without any arguments, doesn't run for all
2626 instances anymore; either pass ``--all`` or pass the desired
2627 instances; this helps against mistakes on big clusters where listing
2628 the information for all instances takes a long time
2629 - miscellaneous doc and man pages fixes
2635 *(Released Wed, 8 Apr 2009)*
2637 - Change the internal locking model of some ``gnt-node`` commands, in
2638 order to reduce contention (and blocking of master daemon) when
2639 batching many creation/reinstall jobs
2640 - Fixes to Xen soft reboot
2641 - No longer build documentation at build time, instead distribute it in
2642 the archive, in order to reduce the need for the whole docbook/rst
2649 *(Released Fri, 27 Mar 2009)*
2651 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
2652 - Fix watcher startup sequence, improves the behaviour of busy clusters
2653 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
2654 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
2655 bind address and other places
2656 - Some documentation fixes and updates
2662 *(Released Mon, 2 Mar 2009)*
2664 - More documentation updates, now all docs should be more-or-less
2666 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
2668 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
2670 - Fix an issue related to $libdir/run/ganeti and cluster creation
2676 *(Released Thu, 19 Feb 2009)*
2678 - Xen PVM and KVM have switched the default value for the instance root
2679 disk to the first partition on the first drive, instead of the whole
2680 drive; this means that the OS installation scripts must be changed
2682 - Man pages have been updated
2683 - RAPI has been switched by default to HTTPS, and the exported functions
2684 should all work correctly
2685 - RAPI v1 has been removed
2686 - Many improvements to the KVM hypervisor
2687 - Block device errors are now better reported
2688 - Many other bugfixes and small improvements
2694 *(Released Mon, 26 Jan 2009)*
2696 - Version 2 is a general rewrite of the code and therefore the
2697 differences are too many to list, see the design document for 2.0 in
2698 the ``doc/`` subdirectory for more details
2699 - In this beta version there is not yet a migration path from 1.2 (there
2700 will be one in the final 2.0 release)
2701 - A few significant changes are:
2703 - all commands are executed by a daemon (``ganeti-masterd``) and the
2704 various ``gnt-*`` commands are just front-ends to it
2705 - all the commands are entered into, and executed from a job queue,
2706 see the ``gnt-job(8)`` manpage
2707 - the RAPI daemon supports read-write operations, secured by basic
2708 HTTP authentication on top of HTTPS
2709 - DRBD version 0.7 support has been removed, DRBD 8 is the only
2710 supported version (when migrating from Ganeti 1.2 to 2.0, you need
2711 to migrate to DRBD 8 first while still running Ganeti 1.2)
2712 - DRBD devices are using statically allocated minor numbers, which
2713 will be assigned to existing instances during the migration process
2714 - there is support for both Xen PVM and Xen HVM instances running on
2716 - KVM virtualization is supported too
2717 - file-based storage has been implemented, which means that it is
2718 possible to run the cluster without LVM and DRBD storage, for
2719 example using a shared filesystem exported from shared storage (and
2720 still have live migration)
2726 *(Released Tue, 13 Jan 2009)*
2728 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
2729 - Reuse the old instance mac address by default on instance import, if
2730 the instance name is the same.
2731 - Handle situations in which the node info rpc returns incomplete
2733 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
2734 - Improved version of batcher:
2736 - state file support
2737 - instance mac address support
2738 - support for HVM clusters/instances
2740 - Add an option to show the number of cpu sockets and nodes in
2742 - Support OSes that handle more than one version of the OS api (but do
2743 not change the current API in any other way)
2744 - Fix ``gnt-node migrate``
2745 - ``gnt-debug`` man page
2746 - Fixes various more typos and small issues
2747 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
2753 *(Released Wed, 24 Sep 2008)*
2755 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
2756 type of disk exported to fully virtualized instances.
2757 - provide access to the serial console of HVM instances
2758 - instance auto_balance flag, set by default. If turned off it will
2759 avoid warnings on cluster verify if there is not enough memory to fail
2760 over an instance. in the future it will prevent automatically failing
2761 it over when we will support that.
2762 - batcher tool for instance creation, see ``tools/README.batcher``
2763 - ``gnt-instance reinstall --select-os`` to interactively select a new
2764 operating system when reinstalling an instance.
2765 - when changing the memory amount on instance modify a check has been
2766 added that the instance will be able to start. also warnings are
2767 emitted if the instance will not be able to fail over, if auto_balance
2769 - documentation fixes
2770 - sync fields between ``gnt-instance list/modify/add/import``
2771 - fix a race condition in drbd when the sync speed was set after giving
2772 the device a remote peer.
2778 *(Released Tue, 22 Jul 2008)*
2780 - note: the allowed size and number of tags per object were reduced
2781 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
2782 - fixed twisted 8.x compatibility
2783 - fixed ``gnt-instance replace-disks`` with iallocator
2784 - add TCP keepalives on twisted connections to detect restarted nodes
2785 - disk increase support, see ``gnt-instance grow-disk``
2786 - implement bulk node/instance query for RAPI
2787 - add tags in node/instance listing (optional)
2788 - experimental migration (and live migration) support, read the man page
2789 for ``gnt-instance migrate``
2790 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
2791 has some small improvements in handling its state file
2797 *(Released Fri, 13 Jun 2008)*
2799 - Experimental readonly, REST-based remote API implementation;
2800 automatically started on master node, TCP port 5080, if enabled by
2801 ``--enable-rapi`` parameter to configure script.
2802 - Instance allocator support. Add and import instance accept a
2803 ``--iallocator`` parameter, and call that instance allocator to decide
2804 which node to use for the instance. The iallocator document describes
2805 what's expected from an allocator script.
2806 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
2807 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
2808 node is lost there is still enough memory to fail over the instances
2810 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
2811 ``gnt-cluster verify``, to check for site-specific compliance. All the
2812 hooks will run, and their output, if any, will be displayed. Any
2813 failing hook will make the verification return an error value.
2814 - ``gnt-cluster verify`` now checks that its peers are reachable on the
2815 primary and secondary interfaces
2816 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
2817 that is still declared as part of the cluster and has failed.
2818 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
2819 specifying output fields, that just adds the chosen fields to the
2821 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
2822 export from the filesystem.
2823 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
2824 have been added. Using them you can enable/disable acpi and pae
2825 support, and specify a path for a cd image to be exported to the
2826 instance. These parameters as the name suggest only work on HVM
2828 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
2829 PAE support will be set to the previously hardcoded values, but the
2830 (previously hardcoded) path to the CDROM ISO image will be unset and
2831 if required, needs to be set manually with ``gnt-instance modify``
2833 - The address to which an instance's VNC console is bound is now
2834 selectable per-instance, rather than being cluster wide. Of course
2835 this only applies to instances controlled via VNC, so currently just
2836 applies to HVM clusters.
2842 *(Released Mon, 18 Feb 2008)*
2844 - more tweaks to the disk activation code (especially helpful for DRBD)
2845 - change the default ``gnt-instance list`` output format, now there is
2846 one combined status field (see the manpage for the exact values this
2848 - some more fixes for the mac export to hooks change
2849 - make Ganeti not break with DRBD 8.2.x (which changed the version
2850 format in ``/proc/drbd``) (issue 24)
2851 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
2852 template, allowing migration from DRBD0.7+MD to DRBD8
2858 *(Released Wed, 30 Jan 2008)*
2860 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
2862 - add command aliases infrastructure and a few aliases
2863 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
2864 man pages and the ``--help`` option of ``gnt-node
2865 list``/``gnt-instance list``
2866 - fix ``gnt-backup list`` with down nodes (issue 21)
2867 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
2868 - fix the dist archive and add a check for including svn/git files in
2870 - some developer-related changes: improve the burnin and the QA suite,
2871 add an upload script for testing during development
2877 *(Released Wed, 16 Jan 2008)*
2879 - experimental HVM support, read the install document, section
2880 "Initializing the cluster"
2881 - allow for the PVM hypervisor per-instance kernel and initrd paths
2882 - add a new command ``gnt-cluster verify-disks`` which uses a new
2883 algorithm to improve the reconnection of the DRBD pairs if the device
2884 on the secondary node has gone away
2885 - make logical volume code auto-activate LVs at disk activation time
2886 - slightly improve the speed of activating disks
2887 - allow specification of the MAC address at instance creation time, and
2888 changing it later via ``gnt-instance modify``
2889 - fix handling of external commands that generate lots of output on
2891 - update documentation with regard to minimum version of DRBD8 supported
2897 *(Released Tue, 4 Dec 2007)*
2899 - Log the ``xm create`` output to the node daemon log on failure (to
2900 help diagnosing the error)
2901 - In debug mode, log all external commands output if failed to the logs
2902 - Change parsing of lvm commands to ignore stderr
2908 *(Released Wed, 28 Nov 2007)*
2910 - Another round of updates to the DRBD 8 code to deal with more failures
2911 in the replace secondary node operation
2912 - Some more logging of failures in disk operations (lvm, drbd)
2913 - A few documentation updates
2920 *(Released Tue, 13 Nov 2007)*
2922 - Change configuration file format from Python's Pickle to JSON.
2923 Upgrading is possible using the cfgupgrade utility.
2924 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
2925 faster replace disks and is more stable (DRBD 8 has many improvements
2926 compared to DRBD 0.7)
2927 - Added command line tags support (see man pages for ``gnt-instance``,
2928 ``gnt-node``, ``gnt-cluster``)
2929 - Added instance rename support
2930 - Added multi-instance startup/shutdown
2931 - Added cluster rename support
2932 - Added ``gnt-node evacuate`` to simplify some node operations
2933 - Added instance reboot operation that can speedup reboot as compared to
2935 - Soften the requirement that hostnames are in FQDN format
2936 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
2938 - Removed dependency on debian's patched fping that uses the
2939 non-standard ``-S`` option
2940 - Now the OS definitions are searched for in multiple, configurable
2941 paths (easier for distros to package)
2942 - Some changes to the hooks infrastructure (especially the new
2943 post-configuration update hook)
2944 - Other small bugfixes
2946 .. vim: set textwidth=72 syntax=rst :